Concepts
- $(\varepsilon (3)
- $(l_0 (3)
- $\ell_1$ norm (3)
- $\epsilon$-stationary point (3)
- $f$-fairness (1)
- $p$-matchoids (1)
- 3d consistency (6)
- 3d environments (7)
- 3d gaussian splatting (44)
- 3d generation (3)
- 3d geometry (9)
- 3d modeling (3)
- 3d object detection (11)
- 3d perception (4)
- 3d point clouds (6)
- 3d pose estimation (3)
- 3d reasoning (3)
- 3d reconstruction (24)
- 3d representations (3)
- 3d scene reconstruction (7)
- 3d scene understanding (6)
- 3d scenes (4)
- 3d spatial reasoning (4)
- 3d vision (3)
- 3d visual grounding (4)
- 4d gaussian splatting (3)
- 4d reconstruction (5)
- ablation experiments (4)
- ablation studies (49)
- ablation study (7)
- ablations (4)
- abstention (3)
- abstract syntax tree (3)
- accelerated convergence (4)
- acceleration (5)
- acceptance rates (4)
- accountability (4)
- accuracy assessment (4)
- accuracy benchmarks (3)
- accuracy degradation (6)
- accuracy enhancement (3)
- accuracy evaluation (8)
- accuracy gains (4)
- accuracy gap (4)
- accuracy improvement (46)
- accuracy improvements (6)
- accuracy metrics (4)
- accuracy preservation (6)
- accuracy retention (3)
- accuracy trade-off (4)
- action distribution (3)
- action prediction (4)
- action recognition (6)
- action selection (6)
- action sequences (5)
- action space (11)
- action trajectories (3)
- actionable insights (7)
- activation function (3)
- activation functions (8)
- activation patterns (4)
- activation space (6)
- activation sparsity (3)
- activation steering (3)
- active exploration (3)
- active learning (13)
- active learning algorithm (3)
- adam (5)
- adamw (3)
- adaptability (20)
- adaptation (9)
- adaptive adjustment (3)
- adaptive algorithm (3)
- adaptive computation (3)
- adaptive framework (3)
- adaptive integration (3)
- adaptive learning (6)
- adaptive methods (3)
- adaptive optimization (5)
- adaptive optimizers (4)
- adaptive policies (3)
- adaptive reasoning (7)
- adaptive regularization (3)
- adaptive sampling (6)
- adaptive selection (6)
- adaptive stepsizes (4)
- adaptivity (4)
- additive error (3)
- adjacency matrix (3)
- adjoint matching (3)
- advanced reasoning (3)
- advantage estimation (4)
- advantage function (4)
- adversarial arrival (1)
- adversarial attacks (34)
- adversarial examples (20)
- adversarial imitation learning (3)
- adversarial inputs (3)
- adversarial learning (8)
- adversarial manipulation (6)
- adversarial noise (3)
- adversarial perturbations (14)
- adversarial prompts (4)
- adversarial robustness (17)
- adversarial training (14)
- aesthetic quality (3)
- agent behavior (4)
- agent interactions (6)
- agent performance (5)
- agentic ai (3)
- agentic systems (4)
- agentic workflow (3)
- aggregation (4)
- agnostic case (3)
- agnostic learning (6)
- agnostic setting (3)
- ai agents (4)
- ai alignment (7)
- ai assistants (3)
- ai governance (4)
- ai models (5)
- ai safety (6)
- ai systems (3)
- ai-generated images (3)
- ai-generated videos (3)
- aleatoric uncertainty (12)
- algebraic structure (3)
- algorithm acceleration (4)
- algorithm adaptation (4)
- algorithm analysis (4)
- algorithm design (24)
- algorithm development (6)
- algorithm effectiveness (5)
- algorithm efficiency (9)
- algorithm implementation (3)
- algorithm performance (12)
- algorithm scalability (3)
- algorithm validation (4)
- algorithmic approaches (3)
- algorithmic choices (3)
- algorithmic design (6)
- algorithmic efficiency (14)
- algorithmic fairness (5)
- algorithmic framework (7)
- algorithmic improvement (4)
- algorithmic performance (9)
- algorithmic reasoning (5)
- algorithmic stability (5)
- algorithms (6)
- alignment (31)
- alignment algorithms (3)
- alignment approaches (3)
- alignment framework (4)
- alignment methods (8)
- alignment performance (5)
- alignment techniques (3)
- alpacaeval (3)
- alphafold (3)
- alphazero (3)
- alternating optimization (4)
- ambiguity (3)
- amortized variational inference (3)
- analog in-memory computing (3)
- analytical framework (4)
- analytical methods (3)
- anatomical structures (3)
- anisotropy (3)
- annotated datasets (4)
- annotation (4)
- annotation costs (4)
- annotation pipeline (4)
- annotations (3)
- anomaly detection (17)
- anomaly localisation (2)
- anomaly localization (4)
- answer generation (4)
- application domains (3)
- application scenarios (3)
- approximate message passing (6)
- approximation (8)
- approximation algorithm (6)
- approximation algorithms (7)
- approximation error (13)
- approximation errors (8)
- approximation guarantee (7)
- approximation guarantees (5)
- approximation methods (4)
- approximation quality (6)
- approximation theory (5)
- arboricity (1)
- architectural changes (4)
- architectural choices (4)
- architectural complexity (3)
- architectural constraints (3)
- architectural design (7)
- architectural differences (3)
- architectural innovations (4)
- architectural modifications (6)
- architecture (6)
- architecture design (4)
- architecture-agnostic (6)
- architectures (5)
- arithmetic reasoning (5)
- articulated objects (4)
- artifacts (3)
- artificial general intelligence (4)
- artificial neural networks (16)
- associative memories (3)
- asymmetric error costs (1)
- asymptotic behavior (6)
- asymptotic consistency (4)
- asymptotic convergence (5)
- asymptotic normality (5)
- asymptotic optimality (5)
- asymptotic variance (7)
- atari games (4)
- attack effectiveness (5)
- attack performance (3)
- attack strategies (3)
- attack success rate (16)
- attack success rates (13)
- attention computation (3)
- attention heads (21)
- attention layers (6)
- attention maps (9)
- attention mechanism (31)
- attention mechanisms (28)
- attention modules (3)
- attention patterns (5)
- attention scores (10)
- attention sink (3)
- attention weights (5)
- attractor dynamics (3)
- audio language models (3)
- audio-language models (3)
- audio-visual large language models (3)
- augmented reality (8)
- augmented views (4)
- auroc (6)
- autoencoder (4)
- autoencoders (4)
- autoformalization (4)
- automated framework (11)
- automated pipeline (6)
- automated theorem proving (3)
- automatic differentiation (4)
- automatic metrics (3)
- automatic speech recognition (6)
- automation (3)
- automl (3)
- autonomous agents (10)
- autonomous driving (40)
- autonomous navigation (4)
- autonomous skill acquisition (3)
- autonomous systems (8)
- autonomous vehicles (4)
- autonomy (3)
- autoregressive framework (6)
- autoregressive generation (9)
- autoregressive image generation (3)
- autoregressive language models (6)
- autoregressive model (6)
- autoregressive modeling (6)
- autoregressive models (42)
- autoregressive prediction (5)
- autoregressive reasoning (3)
- auxiliary inputs (3)
- average accuracy (3)
- average precision (3)
- average reward (3)
- average treatment effect (4)
- average-reward (3)
- back-projection (3)
- back-propagation (4)
- backdoor attacks (15)
- backdoor detection (3)
- backpropagation (19)
- backpropagation through time (4)
- backtracking (5)
- bandit algorithms (3)
- bandit feedback (8)
- bandit problems (4)
- baseline comparison (15)
- baseline method (3)
- baseline methods (24)
- baseline models (7)
- baseline performance (3)
- baselines (6)
- batch effects (3)
- batch normalization (3)
- batch size (9)
- bayes risk (3)
- bayesian estimation (3)
- bayesian flow network (3)
- bayesian inference (19)
- bayesian methods (5)
- bayesian optimization (24)
- bayesian perspective (3)
- bayesian statistics (3)
- beam search (6)
- behavior cloning (12)
- behavioral cloning (6)
- benchmark construction (5)
- benchmark dataset (34)
- benchmark datasets (91)
- benchmark development (3)
- benchmark evaluation (61)
- benchmark experiments (5)
- benchmark framework (6)
- benchmark performance (13)
- benchmark problems (3)
- benchmark results (5)
- benchmark scenarios (3)
- benchmark suite (7)
- benchmark tasks (10)
- benchmark. (3)
- benchmarking (30)
- benchmarking datasets (3)
- benchmarking framework (4)
- benchmarking suite (3)
- benchmarks (64)
- benign overfitting (4)
- best-of-n (5)
- best-of-n sampling (3)
- bi-level optimization (14)
- bias (7)
- bias amplification (3)
- bias correction (4)
- bias mitigation (10)
- bias reduction (4)
- biased counterfactuals (1)
- biases (4)
- bilevel optimization (19)
- billion-parameter models (3)
- binary classification (16)
- binding affinity (8)
- binding problem (2)
- biological constraints (3)
- biological mechanisms (3)
- biological plausibility (4)
- biological systems (3)
- bipartite matching (3)
- black-box access (5)
- black-box attacks (3)
- black-box models (11)
- black-box optimization (7)
- block coordinate descent (4)
- boltzmann distribution (9)
- boosting (3)
- bootstrapping (5)
- bottleneck (4)
- boundary conditions (3)
- bradley-terry model (6)
- brain connectivity (3)
- brain-computer interfaces (9)
- budget allocation (3)
- bundle adjustment (3)
- byte pair encoding (3)
- calibration (20)
- calibration data (4)
- calibration error (3)
- calibration performance (4)
- calibration set (4)
- camera pose estimation (3)
- camera poses (9)
- camera trajectories (3)
- candidate responses (5)
- candidate selection (4)
- candidate solutions (4)
- canonical correlation analysis (3)
- capabilities (3)
- capacitated vehicle routing problem (4)
- caption generation (3)
- case studies (6)
- catastrophic forgetting (46)
- causal analysis (3)
- causal attention (4)
- causal dependencies (4)
- causal discovery (16)
- causal effect (6)
- causal effect estimation (6)
- causal effects (6)
- causal framework (3)
- causal graph (5)
- causal graphs (6)
- causal inference (32)
- causal influence (3)
- causal intervention (3)
- causal language models (3)
- causal models (3)
- causal perspective (3)
- causal reasoning (18)
- causal relationship (7)
- causal relationships (10)
- causal representation learning (11)
- causal structure (8)
- causal structures (4)
- causality (3)
- central limit theorem (5)
- chain of thought (4)
- chain-of-thought (43)
- chain-of-thought prompting (8)
- chain-of-thought reasoning (41)
- challenging datasets (3)
- chamfer distance (4)
- chemical datasets (3)
- cifar-10 (17)
- cifar-100 (6)
- circuit dynamics (4)
- class imbalance (15)
- class prevalences (1)
- class prototypes (3)
- class-conditional generation (8)
- class-incremental learning (3)
- classification (19)
- classification accuracy (8)
- classification benchmarks (5)
- classification performance (8)
- classification tasks (28)
- classifier (4)
- classifier guidance (6)
- classifier-free guidance (21)
- clean accuracy (3)
- climate change (3)
- clinical applications (3)
- clinical decision-making (5)
- clinical deployment conditions (1)
- clinical narratives (2)
- clinical score prediction (1)
- clinical workflows (3)
- clinically curated data (1)
- clip (13)
- closed-form expression (6)
- closed-form expressions (3)
- closed-form solution (8)
- closed-form solutions (8)
- closed-set problem (2)
- closed-source models (7)
- clustering (24)
- clustering algorithms (7)
- clustering performance (6)
- cnns (4)
- co-evolution (3)
- coarse-to-fine pipeline (3)
- coco (3)
- code availability (5)
- code generation (24)
- code reasoning (3)
- coding benchmarks (3)
- coding tasks (3)
- cognitive biases (4)
- cognitive modeling (4)
- cognitive neuroscience (6)
- cognitive processes (6)
- cognitive science (6)
- cognitive states (5)
- cognitive tasks (3)
- cogvideox (3)
- coherence (6)
- coherent reasoning (4)
- cohesion (3)
- collaboration (4)
- collaborative filtering (4)
- collaborative optimization (4)
- collaborative reasoning (3)
- combinatorial complexity (5)
- combinatorial optimization (25)
- combinatorial search space (3)
- combinatorial space (3)
- commonsense reasoning (9)
- communication cost (5)
- communication costs (6)
- communication efficiency (9)
- communication overhead (16)
- community detection (5)
- compact domain (3)
- compact representation (4)
- competitive performance (25)
- competitive programming (4)
- competitive ratio (3)
- competitive results (3)
- compiler optimization (3)
- complementary information (8)
- complementary strengths (4)
- complex data distributions (3)
- complex environments (8)
- complex reasoning (14)
- complex reasoning tasks (5)
- complex scenarios (3)
- complex scenes (3)
- complex tasks (8)
- complexity (7)
- complexity analysis (10)
- complexity bounds (3)
- complexity measure (4)
- complexity reduction (4)
- composition (3)
- compositional control (3)
- compositional generalization (11)
- compositional properties (3)
- compositional reasoning (9)
- compositional structure (4)
- compositionality (7)
- compounding errors (4)
- comprehensive benchmark (4)
- comprehensive dataset (3)
- comprehensive evaluation (13)
- comprehensive evaluations (6)
- comprehensive experiments (13)
- compressed sensing (3)
- compressibility (3)
- compression (8)
- compression methods (4)
- compression ratios (3)
- computation cost (3)
- computation efficiency (3)
- computation reduction (6)
- computational biology (3)
- computational bottleneck (8)
- computational budget (6)
- computational burden (4)
- computational challenges (4)
- computational complexity (57)
- computational constraints (6)
- computational cost (62)
- computational costs (48)
- computational demands (5)
- computational efficiency (202)
- computational feasibility (6)
- computational fluid dynamics (4)
- computational hardness (5)
- computational intractability (4)
- computational model (4)
- computational models (4)
- computational neuroscience (4)
- computational overhead (85)
- computational pathology (3)
- computational redundancy (4)
- computational resources (20)
- computational tractability (4)
- compute budget (6)
- compute budgets (3)
- compute cost (3)
- compute efficiency (8)
- compute scaling (6)
- compute-optimal scaling (4)
- computer vision (27)
- computer-aided design (5)
- concentration bounds (3)
- concept bottleneck models (6)
- concept class (4)
- concept classes (3)
- concept drift (5)
- concept erasure (3)
- concept learning (3)
- concept-based models (3)
- condition number (10)
- conditional diffusion models (4)
- conditional distribution (3)
- conditional distributions (3)
- conditional entropy (3)
- conditional flow matching (7)
- conditional generation (10)
- conditional gradient (3)
- conditional independence (8)
- conditional mutual information (5)
- conditional sampling (6)
- conditional variance (3)
- conditioning (6)
- confidence calibration (5)
- confidence intervals (14)
- confidence score (3)
- confidence scores (4)
- conflicting objectives (4)
- conformal inference (3)
- conformal prediction (32)
- conformal risk control (3)
- conformational diversity (3)
- conformational flexibility (3)
- confounders (5)
- connectivity patterns (4)
- consistency (15)
- consistency constraint (3)
- consistency models (4)
- constrained markov decision process (4)
- constrained optimization (22)
- constraint satisfaction (8)
- constraint violations (4)
- constraints (3)
- contamination (3)
- content authenticity (3)
- content provenance (3)
- context length (13)
- context window (5)
- context windows (3)
- context-aware reasoning (3)
- context-dependent meanings (1)
- contextual bandit (3)
- contextual bandits (10)
- contextual coherence (3)
- contextual dynamic pricing (3)
- contextual information (11)
- contextual integrity (3)
- contextual knowledge (6)
- contextual reasoning (4)
- contextual understanding (3)
- continual adaptation (4)
- continual learning (38)
- continual test-time adaptation (3)
- continuity (3)
- continuous control (5)
- continuous control tasks (3)
- continuous embeddings (3)
- continuous functions (3)
- continuous latent space (4)
- continuous time (4)
- contrastive alignment (5)
- contrastive decoding (4)
- contrastive language-image pre-training (3)
- contrastive language-image pretraining (3)
- contrastive learning (53)
- contrastive loss (10)
- contrastive loss function (3)
- contrastive losses (3)
- contrastive objective (3)
- contrastive objectives (3)
- contrastive pretraining (3)
- contrastive self-supervised learning (4)
- contrastive vision-language training (3)
- control (4)
- control policies (5)
- control signals (4)
- control theory (5)
- controllability (12)
- controllable generation (7)
- controllable image generation (3)
- controlled experiments (8)
- controlnet (4)
- convergence (55)
- convergence acceleration (10)
- convergence analysis (23)
- convergence behavior (7)
- convergence guarantee (5)
- convergence guarantees (31)
- convergence proof (4)
- convergence properties (9)
- convergence rate (29)
- convergence rates (39)
- convergence speed (18)
- conversational systems (3)
- convex combination (4)
- convex functions (3)
- convex geometry (4)
- convex losses (4)
- convex optimization (16)
- convex set (3)
- convexity assumptions (3)
- convolutional filters (3)
- convolutional layers (4)
- convolutional neural networks (18)
- cooperative game theory (4)
- coordination (3)
- copyright infringement (4)
- coreset selection (3)
- correctness (3)
- correlated equilibria (3)
- correlation (4)
- correlation clustering (4)
- correspondence (3)
- cortical regions (4)
- cosine similarity (6)
- cost asymmetries (1)
- cost constraints (2)
- cost efficiency (3)
- cost function (4)
- cost functions (4)
- cost reduction (3)
- cost-effectiveness (5)
- cost-efficiency (3)
- cost-performance trade-off (1)
- cost-weighted performance (1)
- counterexamples (3)
- counterfactual explanations (7)
- counterfactual inference (5)
- counterfactual reasoning (6)
- counterfactual text generation (1)
- counting tasks (1)
- covariance matrices (3)
- covariance matrix (5)
- covariate shift (8)
- covariate shifts (4)
- covariates (4)
- coverage (4)
- coverage guarantees (4)
- creativity (5)
- credit assignment (11)
- critical points (3)
- cross-attention (16)
- cross-dataset generalization (6)
- cross-domain adaptation (2)
- cross-domain generalization (12)
- cross-domain knowledge transfer (3)
- cross-domain scenarios (3)
- cross-entropy (7)
- cross-entropy loss (7)
- cross-modal alignment (13)
- cross-modal attention (3)
- cross-modal fusion (4)
- cross-modal generation (3)
- cross-modal interaction (5)
- cross-modal interactions (5)
- cross-modal learning (4)
- cross-modal misalignment (3)
- cross-modal reasoning (4)
- cross-modal representation learning (3)
- cross-modal retrieval (7)
- cross-task generalization (7)
- cross-task synergy (3)
- cross-view consistency (4)
- cross-view fusion (3)
- cumulative errors (3)
- cumulative regret (3)
- cumulative return (3)
- curated dataset (4)
- curriculum learning (19)
- curse of dimensionality (7)
- curvature (3)
- data acquisition (3)
- data analysis (5)
- data assimilation (5)
- data attribution (9)
- data augmentation (42)
- data augmentation strategy (3)
- data augmentations (5)
- data cleaning (3)
- data collection (9)
- data complexity (3)
- data contamination (9)
- data corruption (4)
- data curation (8)
- data distribution (27)
- data distributions (8)
- data diversity (11)
- data efficiency (20)
- data engine (3)
- data fidelity (3)
- data filtering (3)
- data fusion (3)
- data generation (6)
- data generation framework (3)
- data heterogeneity (23)
- data imbalance (5)
- data integration (4)
- data leakage (7)
- data manifold (5)
- data memorization (2)
- data parallelism (3)
- data poisoning (3)
- data privacy (14)
- data protection (3)
- data quality (14)
- data replay (3)
- data representation (3)
- data representations (5)
- data sampling (3)
- data scaling (3)
- data scarcity (18)
- data selection (10)
- data selection strategy (3)
- data shapley (3)
- data synthesis (10)
- data valuation (8)
- data-driven approach (4)
- data-driven approaches (7)
- data-driven methods (3)
- data-driven models (4)
- data-efficient (5)
- data-generating process (5)
- data-scarce settings (6)
- dataset biases (3)
- dataset collection (4)
- dataset construction (6)
- dataset curation (18)
- dataset distillation (11)
- dataset diversity (3)
- dataset evaluation (9)
- dataset generation (4)
- dataset limitations (3)
- dataset quality (4)
- dataset release (3)
- dataset size (10)
- debiasing (4)
- deblurring (3)
- decentralized agents (3)
- decentralized settings (3)
- decentralized training (5)
- decision boundaries (13)
- decision boundary (7)
- decision quality (4)
- decision support systems (2)
- decision transformer (3)
- decision trees (6)
- decision-making (35)
- decision-making performance (4)
- decision-making process (6)
- decision-making processes (4)
- decision-making tasks (4)
- decoder-only transformer (6)
- decoder-only transformers (4)
- decoding accuracy (4)
- decoding approaches (1)
- decoding performance (3)
- decoding process (4)
- decoding strategy (4)
- decoupling (3)
- deductive reasoning (3)
- deep generative models (10)
- deep learning architectures (11)
- deep learning methods (6)
- deep learning model (3)
- deep learning models (13)
- deep networks (3)
- deep neural network (4)
- deep neural networks (50)
- deep reasoning (5)
- deep reinforcement learning (21)
- deepfake detection (9)
- deepseek-r1 (6)
- defense mechanisms (7)
- defenses (3)
- degeneracy ordering (1)
- degrees of freedom (3)
- demographic parity (3)
- demonstrations (3)
- denoising (20)
- denoising diffusion probabilistic models (3)
- denoising process (13)
- denoising processes (3)
- denoising steps (12)
- denoising trajectory (5)
- dense prediction (3)
- density estimation (8)
- density functional theory (3)
- deployment efficiency (3)
- depth (3)
- depth estimation (10)
- depth information (3)
- depth maps (5)
- depth priors (3)
- design choices (8)
- design principles (5)
- design space (4)
- detail preservation (3)
- detectability (7)
- detection (5)
- detection accuracy (5)
- detection method (4)
- detection methods (5)
- detection performance (8)
- dexterous manipulation (5)
- diagnostic benchmark (4)
- diagnostic reasoning (3)
- diagnostic tool (3)
- diagonal linear networks (3)
- differentiable optimization (7)
- differentiable rendering (8)
- differential equations (4)
- differential geometry (3)
- differential privacy (45)
- differentially private (6)
- differentially private algorithm (3)
- differentially private algorithms (3)
- difficulty levels (1)
- diffusion (4)
- diffusion generative models (3)
- diffusion language models (7)
- diffusion model (26)
- diffusion models (195)
- diffusion policies (9)
- diffusion policy (4)
- diffusion process (7)
- diffusion samplers (4)
- diffusion sampling (6)
- diffusion transformer (19)
- diffusion transformers (21)
- diffusion-based framework (7)
- diffusion-based generative models (5)
- diffusion-based methods (5)
- diffusion-based models (12)
- diffusion-based samplers (3)
- digital twin (5)
- digital twins (6)
- dimensionality (6)
- dimensionality reduction (13)
- diminishing returns (3)
- dino (3)
- dinov2 (6)
- direct preference optimization (43)
- directed acyclic graph (7)
- directed acyclic graphs (8)
- directed edges (3)
- directed graph (3)
- directional convergence (3)
- discrepancies (4)
- discrete data (5)
- discrete diffusion (4)
- discrete diffusion models (15)
- discrete distributions (3)
- discrete representations (3)
- discrete state space (3)
- discrete tokens (7)
- discrete-time (3)
- discretization (6)
- discretization error (3)
- discriminability (5)
- discrimination (4)
- discriminative features (5)
- discriminative models (6)
- discriminative power (5)
- discriminative representations (9)
- discriminative tasks (3)
- disentangled representation learning (4)
- disentangled representations (8)
- disentanglement (4)
- distance-based local structure (1)
- distillation (20)
- distortion (4)
- distributed learning (6)
- distributed optimization (5)
- distributed training (8)
- distribution estimation (3)
- distribution gap (3)
- distribution matching (5)
- distribution shift (23)
- distribution shifts (41)
- distributional mismatch (3)
- distributional shift (5)
- distributional shifts (22)
- distributionally robust optimization (5)
- divergence (5)
- diverse datasets (3)
- diverse domains (4)
- diverse scenarios (3)
- diverse solutions (3)
- diversity (13)
- diversity guidance (1)
- diversity measures (2)
- do-calculus (3)
- document ranking (3)
- domain adaptation (16)
- domain discrepancy (3)
- domain expertise (5)
- domain experts (3)
- domain gap (8)
- domain gaps (5)
- domain generalization (15)
- domain invariant representation learning (1)
- domain knowledge (4)
- domain shift (7)
- domain shifts (11)
- domain-customized solutions (1)
- domain-invariant features (6)
- domain-specific benchmarks (3)
- domain-specific expertise (5)
- domain-specific knowledge (7)
- domain-specific models (4)
- domain-specific tasks (3)
- doubly robust estimator (3)
- downstream analysis (3)
- downstream applications (4)
- downstream benchmarks (3)
- downstream performance (17)
- downstream task performance (4)
- downstream tasks (64)
- dp-sgd (3)
- dpo (7)
- draft model (7)
- dropout (3)
- drug design (4)
- drug discovery (16)
- dual formulation (4)
- dual-branch architecture (4)
- dust3r (3)
- dynamic adaptation (7)
- dynamic adjustment (3)
- dynamic constraints (3)
- dynamic environments (20)
- dynamic evaluation (3)
- dynamic graphs (5)
- dynamic pricing (3)
- dynamic programming (10)
- dynamic scenarios (3)
- dynamic scene reconstruction (8)
- dynamic scenes (15)
- dynamic setting (3)
- dynamic settings (3)
- dynamic updates (3)
- dynamic view synthesis (3)
- dynamical systems (18)
- dynamical systems theory (3)
- dynamics (3)
- dynamics prediction (3)
- early sampling consistency (1)
- early stopping (4)
- earth observation (3)
- edge computing (3)
- edge devices (9)
- edge of stability (3)
- edge weights (5)
- edit distance error (1)
- effective dimension (3)
- effective dimensionality (4)
- effective rank (3)
- efficacy (3)
- efficiency critiques (1)
- efficiency enhancement (3)
- efficiency evaluation (3)
- efficiency gains (6)
- efficiency improvement (7)
- efficiency improvements (3)
- efficient algorithm (7)
- efficient algorithms (13)
- efficient deployment (3)
- efficient inference (5)
- efficient learning (4)
- efficient sampling (3)
- efficient training (6)
- eigenfunctions (3)
- eigengap (3)
- eigenvalues (4)
- eigenvectors (5)
- electroencephalography (10)
- electronic design automation (5)
- electronic health records (7)
- embedded efficiency understanding (1)
- embedding dimension (5)
- embedding distribution (3)
- embedding models (3)
- embedding space (19)
- embedding vectors (3)
- embeddings (14)
- embodied agents (20)
- embodied ai (20)
- embodied intelligence (11)
- embodied reasoning (4)
- emergence (3)
- emergent properties (4)
- emotion recognition (5)
- empirical analyses (8)
- empirical analysis (27)
- empirical comparison (4)
- empirical demonstration (18)
- empirical distribution (3)
- empirical evaluation (60)
- empirical evaluations (59)
- empirical evidence (21)
- empirical experiments (16)
- empirical findings (5)
- empirical insights (6)
- empirical investigation (3)
- empirical observations (6)
- empirical performance (46)
- empirical performance improvements (2)
- empirical results (95)
- empirical risk minimization (10)
- empirical studies (19)
- empirical study (17)
- empirical success (7)
- empirical validation (63)
- empirical verification (3)
- encoder representations (3)
- encoder-decoder architecture (6)
- encoding scheme (3)
- end-to-end evaluation (4)
- end-to-end framework (7)
- end-to-end learning (8)
- end-to-end optimization (8)
- end-to-end training (12)
- energy efficiency (5)
- energy function (6)
- energy functions (5)
- energy landscape (4)
- energy-based model (4)
- energy-based models (8)
- enhancement techniques (3)
- ensemble learning (5)
- ensemble methods (9)
- entropy (4)
- entropy computation (1)
- entropy minimization (8)
- entropy regularization (4)
- entropy-regularized (3)
- environment dynamics (5)
- environmental dynamics (3)
- environmental sustainability (3)
- environmental uncertainty (3)
- episodic memory (5)
- epistemic uncertainty (18)
- equivariance (6)
- equivariant neural networks (7)
- error accumulation (17)
- error amplification (3)
- error analysis (10)
- error bound (4)
- error bounds (10)
- error correction (5)
- error patterns (5)
- error propagation (8)
- error rates (3)
- estimation (6)
- estimation accuracy (6)
- estimation error (14)
- estimator (3)
- estimators (5)
- ethical concerns (4)
- ethical implications (3)
- euclidean space (4)
- evaluation benchmark (4)
- evaluation benchmarks (5)
- evaluation criteria (4)
- evaluation dataset (3)
- evaluation framework (26)
- evaluation frameworks (5)
- evaluation methodologies (3)
- evaluation methodology (3)
- evaluation methods (4)
- evaluation metrics (54)
- evaluation performance (3)
- evaluation pipeline (5)
- evaluation practices (3)
- evaluation protocol (9)
- evaluation protocols (8)
- evaluation settings (3)
- evaluation suite (7)
- evaluation tasks (3)
- evaluation-only benchmark (2)
- evaluators (3)
- event cameras (12)
- evidence lower bound (8)
- evolutionary algorithm (3)
- evolutionary algorithms (6)
- exact field matching (1)
- excess error (3)
- excess risk (4)
- excess risk bounds (4)
- executable code (4)
- execution time (4)
- execution time reduction (1)
- exhaustive search (3)
- expected improvement (4)
- expected utility (3)
- experience replay (5)
- experimental benchmarks (4)
- experimental data (3)
- experimental design (7)
- experimental evaluation (17)
- experimental evaluations (10)
- experimental evidence (3)
- experimental results (66)
- experimental study (4)
- experimental validation (21)
- expert annotations (3)
- expert demonstrations (9)
- expert specialization (4)
- expert trajectories (4)
- explainability (17)
- explainable ai (8)
- exploration (19)
- exploration and exploitation (4)
- exploration efficiency (3)
- exploration strategies (5)
- exploration-exploitation (3)
- exploration-exploitation balance (9)
- exploration-exploitation trade-off (9)
- exploratory behaviors (3)
- exponential convergence (3)
- exponential dependence (4)
- exponential mechanism (3)
- exponential moving average (5)
- expressive capacity (4)
- expressive power (20)
- expressiveness (9)
- expressivity (17)
- extensive experiments (25)
- extensive-form games (4)
- extrapolation (12)
- extrapolation errors (3)
- eye gaze (3)
- f1 score (8)
- f1-score (3)
- factorization (4)
- factual accuracy (6)
- failure cases (3)
- failure detection (4)
- failure modes (16)
- fairness (12)
- fairness assessment (3)
- fairness constraints (3)
- fairness factor (1)
- fairness guarantees (3)
- faithfulness (7)
- false discovery rate (4)
- false positive rate (4)
- false positive rates (4)
- feasibility (3)
- feature alignment (6)
- feature attribution (5)
- feature consistency (3)
- feature dimension (4)
- feature discrimination (4)
- feature distillation (5)
- feature distributions (5)
- feature embeddings (3)
- feature engineering (4)
- feature expressiveness (3)
- feature extraction (15)
- feature extractor (7)
- feature extractors (3)
- feature fusion (3)
- feature learning (21)
- feature manipulation (4)
- feature maps (3)
- feature matching (6)
- feature mixing (3)
- feature projection (3)
- feature quality evaluation (1)
- feature reliance (3)
- feature representation (6)
- feature representations (9)
- feature selection (8)
- feature similarity (4)
- feature space (7)
- feature spaces (4)
- feature unlearning (3)
- feature vectors (4)
- fedavg (4)
- federated continual learning (4)
- federated graph learning (5)
- federated learning (54)
- federated training (2)
- feed-forward architecture (3)
- feed-forward inference (3)
- feed-forward network (6)
- feed-forward networks (3)
- feedback loop (4)
- feedback mechanisms (3)
- feedforward networks (3)
- few-shot adaptation (4)
- few-shot classification (3)
- few-shot learning (31)
- few-shot settings (3)
- few-shot transfer (4)
- fid (10)
- fid score (4)
- fid scores (7)
- fidelity (14)
- fidelity metrics (3)
- fine-grained analysis (3)
- fine-grained captions (3)
- fine-grained control (10)
- fine-grained distillation (1)
- fine-grained evaluation (6)
- fine-grained optimization (3)
- fine-grained perception (7)
- fine-grained reasoning (3)
- fine-grained retrieval (3)
- fine-tuned models (6)
- fine-tuning (198)
- fine-tuning framework (4)
- fine-tuning method (3)
- fine-tuning strategies (6)
- fine-tuning strategy (3)
- finetuning (19)
- finite samples (3)
- finite-horizon (7)
- finite-sample guarantees (9)
- finite-sample performance (4)
- finite-time analysis (3)
- finite-time convergence (3)
- first-order algorithms (3)
- first-order logic (3)
- first-order methods (6)
- first-order optimization (3)
- fisher information matrix (4)
- flashattention (8)
- flatness (3)
- flexibility (3)
- flops (7)
- flops reduction (8)
- flow matching (29)
- flow matching model (3)
- flow matching models (3)
- flow models (6)
- flow-based generative model (6)
- flow-based generative models (3)
- flow-based models (5)
- fluency (3)
- fluorescence microscopy (2)
- follow-the-regularized-leader (3)
- forecasting (5)
- forecasting accuracy (4)
- forecasting performance (5)
- forgetting (4)
- formal guarantees (4)
- formal language querying (1)
- formal verification (4)
- forward pass (3)
- forward propagation (7)
- forward transfer (3)
- foundation model (12)
- foundation models (86)
- foundational models (3)
- fourier neural operator (3)
- fourier transform (4)
- fp32 (3)
- framework development (3)
- frequency domain (7)
- frobenius norm (5)
- front-door adjustment (3)
- frontier models (4)
- full-information feedback (3)
- function approximation (9)
- function evaluations (4)
- function spaces (3)
- functional correctness (6)
- functional landscape (3)
- functional magnetic resonance imaging (5)
- functionality improvement (1)
- fundamental limitations (5)
- future prediction (3)
- fuzzy semantic requirements (1)
- gaia benchmark (3)
- gait assessment (1)
- gait-feature baseline (1)
- game theory (4)
- game-theoretic framework (3)
- game-theoretic model (3)
- gans (5)
- gating mechanism (6)
- gating mechanisms (6)
- gaussian approximation (3)
- gaussian distribution (7)
- gaussian distributions (5)
- gaussian mechanism (3)
- gaussian mixture model (7)
- gaussian mixture models (4)
- gaussian noise (10)
- gaussian primitives (8)
- gaussian prior (4)
- gaussian process (14)
- gaussian processes (15)
- gaussian splatting (19)
- general capabilities (3)
- general-purpose models (3)
- generalisation (7)
- generalizability (42)
- generalizable representations (4)
- generalization (242)
- generalization abilities (5)
- generalization ability (33)
- generalization behavior (6)
- generalization bound (6)
- generalization bounds (12)
- generalization capabilities (24)
- generalization capability (13)
- generalization error (11)
- generalization error bound (4)
- generalization error bounds (3)
- generalization gap (3)
- generalization guarantees (10)
- generalization improvement (3)
- generalization performance (27)
- generalization properties (3)
- generalization protocols (1)
- generalized category discovery (4)
- generalized linear models (5)
- generalized smoothness (3)
- generation (4)
- generation capabilities (3)
- generation diversity (4)
- generation fidelity (3)
- generation quality (29)
- generation tasks (4)
- generative adversarial networks (4)
- generative ai (15)
- generative approaches (4)
- generative capabilities (9)
- generative capacity (4)
- generative diffusion models (3)
- generative framework (5)
- generative methods (6)
- generative model (25)
- generative modeling (57)
- generative models (119)
- generative performance (3)
- generative priors (11)
- generative process (10)
- generative quality (8)
- generative tasks (11)
- generative trajectory (5)
- generative video models (3)
- genetic algorithm (4)
- geo-localization (3)
- geodesics (4)
- geometric accuracy (9)
- geometric consistency (11)
- geometric constraints (4)
- geometric cues (4)
- geometric information (5)
- geometric insights (3)
- geometric interpretation (3)
- geometric primitives (4)
- geometric priors (9)
- geometric properties (3)
- geometric reasoning (3)
- geometric structure (14)
- geometric structures (5)
- geometric transformations (3)
- geometric understanding (3)
- geometry (3)
- global coherence (3)
- global consistency (3)
- global context (4)
- global convergence (13)
- global interactions (3)
- global minima (5)
- global model (6)
- global optima (4)
- global optimality (3)
- global optimization (5)
- global optimum (3)
- global semantics (5)
- global valuation problem (1)
- gnns (4)
- goal-conditioned reinforcement learning (5)
- gpqa (4)
- gpt-4o (8)
- gpu memory (7)
- gpu memory usage (4)
- gpu utilization (6)
- gradient alignment (3)
- gradient analysis (3)
- gradient approximation (3)
- gradient ascent (3)
- gradient clipping (8)
- gradient compression (3)
- gradient conflicts (5)
- gradient descent (51)
- gradient descent dynamics (3)
- gradient estimates (4)
- gradient estimation (6)
- gradient estimators (3)
- gradient flow (14)
- gradient flows (3)
- gradient guidance (3)
- gradient information (3)
- gradient magnitude (4)
- gradient oracle (3)
- gradient projection (5)
- gradient propagation (3)
- gradient similarity (4)
- gradient steps (3)
- gradient updates (5)
- gradient vanishing (3)
- gradient variance (8)
- gradient-based algorithms (3)
- gradient-based learning (3)
- gradient-based methods (13)
- gradient-based optimization (14)
- gradient-based training (5)
- gradients (6)
- granularity (3)
- graph anomaly detection (3)
- graph classification (14)
- graph condensation (3)
- graph contrastive learning (4)
- graph convolutional networks (4)
- graph databases (3)
- graph datasets (5)
- graph diffusion models (3)
- graph embedding (4)
- graph foundation models (12)
- graph generalization (3)
- graph generation (3)
- graph learning (4)
- graph machine learning (4)
- graph matching (4)
- graph neural network (16)
- graph neural networks (80)
- graph reasoning (3)
- graph representation (3)
- graph representation learning (5)
- graph representations (5)
- graph rewiring (4)
- graph structure (6)
- graph topology (3)
- graph transformer (4)
- graph transformers (8)
- graph-based methods (3)
- graph-structured data (5)
- graphic matroids (1)
- graphons (3)
- greedy algorithm (5)
- grokking (3)
- grounded reasoning (4)
- grounding (5)
- group fairness (4)
- group relative policy optimization (39)
- grpo (12)
- grpo algorithm (5)
- gsm8k (7)
- guardrail models (3)
- gui agents (5)
- guided generation (3)
- hallucinated content (3)
- hallucination (12)
- hallucination benchmarks (3)
- hallucination detection (4)
- hallucination mitigation (9)
- hallucinations (26)
- hamiltonian dynamics (3)
- hamiltonian monte carlo (4)
- hand-object interaction (3)
- handcrafted heuristics (3)
- hard constraints (4)
- harmful content (3)
- harmful prompts (4)
- harmful queries (3)
- harmful requests (3)
- healthcare applications (5)
- heavy-tailed distributions (6)
- heavy-tailed noise (4)
- hebbian learning (4)
- helpfulness (4)
- hermite expansion (3)
- hessian (3)
- hessian matrix (3)
- heterogeneity (7)
- heterogeneous data (8)
- heterogeneous environments (3)
- heterogeneous features (3)
- heterogeneous graphs (3)
- heterogeneous treatment effects (3)
- heterophilic graphs (4)
- heterophily (3)
- heuristic methods (5)
- heuristic rules (3)
- heuristics (4)
- hidden activations (3)
- hidden confounders (5)
- hidden layers (4)
- hidden representations (4)
- hidden state (3)
- hidden states (10)
- hierarchical architecture (3)
- hierarchical features (4)
- hierarchical framework (6)
- hierarchical relationships (4)
- hierarchical structure (9)
- hierarchical structures (4)
- high dimensionality (5)
- high probability (8)
- high-dimensional (4)
- high-dimensional asymptotics (3)
- high-dimensional control (4)
- high-dimensional data (13)
- high-dimensional datasets (4)
- high-dimensional distributions (3)
- high-dimensional embeddings (3)
- high-dimensional limit (4)
- high-dimensional problems (3)
- high-dimensional regime (3)
- high-dimensional settings (9)
- high-dimensional spaces (5)
- high-dimensional state space (3)
- high-fidelity generation (5)
- high-fidelity reconstruction (3)
- high-fidelity synthesis (7)
- high-fidelity textures (3)
- high-frequency components (5)
- high-frequency details (7)
- high-level characteristics (1)
- high-level semantics (3)
- high-order interactions (3)
- high-quality dataset (6)
- high-quality datasets (3)
- high-quality samples (5)
- high-quality translation (1)
- high-resolution images (10)
- higher-dimensional space (3)
- higher-order interactions (5)
- hilbert-schmidt independence criterion (3)
- hippocampus (3)
- historical data (3)
- homophily (4)
- hopfield networks (3)
- human activity recognition (3)
- human annotation (3)
- human annotations (8)
- human annotators (6)
- human capabilities (3)
- human cognition (6)
- human demonstrations (4)
- human evaluation (3)
- human evaluations (3)
- human expectations (3)
- human feedback (28)
- human intent (3)
- human interpretability (3)
- human pose estimation (4)
- human preferences (14)
- human-ai interaction (3)
- human-in-the-loop (4)
- human-like reasoning (3)
- human-object interaction (4)
- humanoid robots (4)
- hybrid approach (3)
- hybrid architecture (4)
- hybrid architectures (3)
- hybrid framework (4)
- hybrid language models (3)
- hybrid optimization (4)
- hyper-parameters (4)
- hyperbolic geometry (3)
- hyperbolic space (6)
- hypergraph (3)
- hypergraph neural network (3)
- hypergraph neural networks (4)
- hypergraphs (5)
- hyperparameter (3)
- hyperparameter optimization (9)
- hyperparameter settings (3)
- hyperparameter tuning (33)
- hyperparameters (22)
- hypothesis class (8)
- hypothesis generation (6)
- hypothesis space (5)
- hypothesis testing (7)
- i.i.d. samples (4)
- identifiability (11)
- identity consistency (4)
- identity fidelity (3)
- identity preservation (6)
- image captioning (3)
- image classification (39)
- image compression (4)
- image data (3)
- image datasets (7)
- image denoising (3)
- image editing (3)
- image embeddings (6)
- image encoder (3)
- image fusion (5)
- image generation (38)
- image generation benchmarks (3)
- image generation models (3)
- image generation process (3)
- image inpainting (5)
- image latents (4)
- image quality (14)
- image quality assessment (4)
- image reconstruction (10)
- image restoration (11)
- image retrieval (6)
- image segmentation (5)
- image super-resolution (5)
- image synthesis (12)
- image tokens (4)
- image transformations (3)
- image understanding (3)
- image-text pairs (8)
- image-to-image translation (4)
- image-to-video generation (5)
- imagenet (18)
- imagenet-1k (11)
- imaging inverse problems (3)
- imaging modalities (4)
- imitation learning (33)
- implicit actions (1)
- implicit bias (8)
- implicit constraints (3)
- implicit differentiation (3)
- implicit neural representations (8)
- implicit reasoning (2)
- implicit regularization (8)
- implicit representation (3)
- implicit temporal dynamics (1)
- importance sampling (8)
- in-context examples (3)
- in-context learning (66)
- in-distribution (4)
- in-distribution data (3)
- in-distribution samples (3)
- in-domain performance (4)
- independence testing (3)
- independent and identically distributed (4)
- independent set (1)
- individual fairness (3)
- indoor environments (3)
- induction heads (4)
- inductive bias (28)
- inductive biases (36)
- inductive reasoning (3)
- industrial anomaly detection (3)
- industrial applications (4)
- inequality constraints (3)
- inference (21)
- inference acceleration (19)
- inference capacity (1)
- inference complexity (4)
- inference cost (10)
- inference costs (18)
- inference dynamics (3)
- inference efficiency (33)
- inference latency (17)
- inference optimization (6)
- inference overhead (9)
- inference performance (4)
- inference pipeline (3)
- inference process (5)
- inference scaling (4)
- inference speed (18)
- inference speedup (6)
- inference strategy (4)
- inference throughput (3)
- inference time (16)
- inference-time alignment (3)
- inference-time approach (3)
- inference-time computation (5)
- inference-time scaling (8)
- infinite-horizon (5)
- infinite-width limit (4)
- influence functions (12)
- information aggregation (4)
- information bottleneck (5)
- information bottleneck principle (4)
- information bottleneck theory (3)
- information exchange (3)
- information exponent (3)
- information extraction (5)
- information flow (7)
- information gain (10)
- information geometry (3)
- information leakage (5)
- information loss (10)
- information preservation (3)
- information propagation (7)
- information retrieval (7)
- information theory (9)
- information-theoretic framework (5)
- information-theoretic lower bounds (4)
- information-theoretic perspective (3)
- information-theoretic quantity (3)
- informativeness (3)
- initialization (6)
- innovation (3)
- inpainting (6)
- input modalities (3)
- input perturbations (6)
- input representation (3)
- input-output consistency (1)
- instability (3)
- instance segmentation (5)
- instruction fine-tuning (5)
- instruction following (10)
- instruction tuning (10)
- instruction-following (7)
- instruction-following capabilities (4)
- instruction-tuned models (4)
- instruction-tuning (4)
- instruction-tuning dataset (4)
- integration (3)
- intellectual property protection (3)
- intelligent agents (5)
- interaction dynamics (3)
- interaction mechanisms (3)
- interaction modeling (3)
- interaction trajectories (3)
- interactivity (4)
- intermediate layers (5)
- intermediate reasoning steps (4)
- intermediate representations (7)
- intermediate states (3)
- internal mechanisms (3)
- internal representations (8)
- interpolation (13)
- interpretability (112)
- interpretability assessment (3)
- interpretable features (3)
- interpretable models (13)
- interpretable predictions (5)
- interpretable representations (3)
- interventional data (4)
- intra-class compactness (5)
- intra-class diversity (6)
- intracranial recordings (3)
- intrinsic dimension (4)
- intrinsic geometry (6)
- intrinsic motivation (3)
- invariance (3)
- inverse design (4)
- inverse problem (4)
- inverse problems (12)
- inverse reinforcement learning (5)
- inverse rendering (3)
- irregular sampling (3)
- irrelevant information (3)
- iteration complexity (7)
- iterative denoising (6)
- iterative improvement (3)
- iterative method (3)
- iterative optimization (8)
- iterative process (5)
- iterative reasoning (3)
- iterative refinement (16)
- iterative sampling (6)
- iterative training (5)
- jacobian (4)
- jacobian matrix (4)
- jacobians (3)
- jailbreak attacks (15)
- jailbreaking (4)
- joint distribution (12)
- joint optimization (11)
- joint training (6)
- k-means clustering (3)
- kernel ridge regression (3)
- key-value cache (19)
- key-value caching (3)
- keypoint lifting (1)
- kl divergence (27)
- kl regularization (5)
- kl-divergence (5)
- knowledge acquisition (5)
- knowledge adaptation (3)
- knowledge conflicts (4)
- knowledge distillation (42)
- knowledge fusion (3)
- knowledge graph (6)
- knowledge graphs (14)
- knowledge injection (3)
- knowledge integration (6)
- knowledge preservation (4)
- knowledge retention (6)
- knowledge retrieval (3)
- knowledge sharing (3)
- knowledge tracing (3)
- knowledge transfer (32)
- knowledge-intensive tasks (5)
- koopman operator (4)
- kullback-leibler divergence (10)
- kullback–leibler divergence (6)
- kv cache (8)
- kv cache compression (6)
- kv cache eviction (3)
- kv-cache (3)
- l_1)$-smoothness (3)
- label distribution (3)
- label noise (7)
- labeled data (6)
- laminar matroids (1)
- langevin dynamics (10)
- language diversity (1)
- language generation (3)
- language grounding (3)
- language model (12)
- language model agents (1)
- language model pretraining (5)
- language model training (3)
- language models (47)
- language tasks (6)
- large batch sizes (3)
- large datasets (5)
- large language model (62)
- large multi-modality models (3)
- large multimodal models (9)
- large reasoning models (33)
- large reconstruction model (3)
- large vision language models (4)
- large vision models (3)
- large vision-language models (28)
- large-scale applications (3)
- large-scale benchmark (7)
- large-scale benchmarks (3)
- large-scale data (3)
- large-scale dataset (19)
- large-scale datasets (16)
- large-scale evaluation (3)
- large-scale experiments (3)
- large-scale generation (1)
- large-scale models (9)
- large-scale pre-training (4)
- large-scale pretraining (5)
- large-scale retrieval (3)
- large-scale training (5)
- last-iterate convergence (6)
- late fusion (3)
- latency (3)
- latency overhead (3)
- latency reduction (9)
- latent continuous-time dynamics (1)
- latent diffusion model (6)
- latent diffusion models (11)
- latent directions (3)
- latent embeddings (3)
- latent representation (11)
- latent representations (20)
- latent space (50)
- latent spaces (8)
- latent variable model (3)
- latent variable models (8)
- latent variables (14)
- latent-space interventions (1)
- layer normalization (4)
- layout-to-image generation (3)
- learnability (11)
- learnable parameters (3)
- learned representations (13)
- learning algorithm (7)
- learning algorithms (15)
- learning capacity (3)
- learning complexity (3)
- learning curves (3)
- learning difficulty (3)
- learning dynamics (25)
- learning efficiency (14)
- learning objective (3)
- learning objectives (4)
- learning paradigm (3)
- learning process (5)
- learning rate (9)
- learning rates (11)
- learning signals (3)
- learning tasks (4)
- learning theory (7)
- learning-augmented algorithms (6)
- learning-based approaches (3)
- learning-based methods (9)
- length generalization (6)
- lidar (6)
- lifelong learning (3)
- lightweight adapters (3)
- lightweight framework (3)
- lightweight model (4)
- lightweight module (3)
- likelihood estimation (6)
- likelihood maximization (6)
- limitations (3)
- linear approximations (4)
- linear attention (11)
- linear bandits (5)
- linear classifier (3)
- linear combination (3)
- linear complexity (9)
- linear computational complexity (4)
- linear constraints (3)
- linear convergence rate (5)
- linear function approximation (4)
- linear layers (4)
- linear mode connectivity (7)
- linear models (4)
- linear optimization (3)
- linear probing (8)
- linear programming (6)
- linear regression (10)
- linear rnns (3)
- linear systems (3)
- linear temporal logic (4)
- linear transformation (3)
- linear transformations (5)
- linear transformers (4)
- linguistic instructions (3)
- link prediction (13)
- lipschitz continuity (9)
- littlestone dimension (5)
- llama (3)
- llama 3 (3)
- llama models (4)
- llava-1.5 (3)
- llm (8)
- llm agents (11)
- llm alignment (4)
- llm backbones (3)
- llm inference (3)
- llm pretraining (3)
- llm reasoning (4)
- llm unlearning (3)
- llm-as-a-judge (13)
- llm-based agents (10)
- local attention (4)
- local consistency (4)
- local convergence (3)
- local dependencies (3)
- local differential privacy (4)
- local features (3)
- local fidelity (4)
- local minima (3)
- local optima (12)
- local sgd (3)
- locality (6)
- localization (8)
- localization accuracy (3)
- localized computations (1)
- logarithmic factors (6)
- logarithmic increase (1)
- logarithmic regret (3)
- logical consistency (4)
- logical inference (4)
- logical reasoning (14)
- logistic regression (7)
- logit adjustment (3)
- logits (4)
- long chain-of-thought (3)
- long video understanding (3)
- long-context reasoning (4)
- long-context scenarios (3)
- long-context tasks (3)
- long-horizon planning (10)
- long-horizon prediction (3)
- long-horizon reasoning (5)
- long-horizon tasks (8)
- long-range dependencies (28)
- long-range dependency modeling (3)
- long-tailed distributions (3)
- long-tailed recognition (3)
- long-term dependencies (4)
- long-term forecasting (3)
- long-term memory (6)
- long-term time series forecasting (3)
- long-video benchmarks (3)
- longbench (3)
- lora (13)
- lora adapters (4)
- loss function (10)
- loss functions (12)
- loss gradients (3)
- loss landscape (20)
- loss landscapes (3)
- loss minimization (5)
- lossless compression (5)
- lossless inference (3)
- lossy compression (5)
- low-bit quantization (3)
- low-data regimes (4)
- low-data settings (3)
- low-dimensional manifold (4)
- low-dimensional representation (4)
- low-dimensional subspace (6)
- low-frequency components (7)
- low-light image enhancement (3)
- low-rank adaptation (39)
- low-rank adapter (3)
- low-rank adapters (3)
- low-rank approximation (5)
- low-rank matrices (3)
- low-rank optimization (3)
- low-rank structure (8)
- low-rank tensor decompositions (3)
- low-resource languages (7)
- low-resource settings (4)
- lower bound (24)
- lower bounds (21)
- lvbench (3)
- machine learning algorithms (6)
- machine learning applications (4)
- machine learning methodologies (3)
- machine learning methods (4)
- machine learning model (4)
- machine learning models (18)
- machine translation (3)
- machine unlearning (22)
- majority voting (9)
- mamba (11)
- mamba architecture (8)
- manipulation tasks (8)
- manual annotation (4)
- mapping (4)
- marginal distributions (3)
- marginal likelihood (5)
- marginal probability distribution (3)
- markov chain monte carlo (6)
- markov chains (6)
- markov decision process (12)
- markov decision processes (15)
- markov equivalence class (4)
- markov games (3)
- markov processes (3)
- marl benchmarks (3)
- masked autoencoder (3)
- masked autoencoding (3)
- masked diffusion models (6)
- masking (4)
- massively parallel computation (3)
- math (3)
- math benchmarks (4)
- math reasoning (5)
- math reasoning benchmarks (5)
- mathematical benchmarks (8)
- mathematical datasets (3)
- mathematical model (3)
- mathematical problem solving (4)
- mathematical reasoning (29)
- mathematical reasoning benchmarks (12)
- mathematical reasoning datasets (3)
- mathematical reasoning tasks (4)
- matrix factorization (7)
- matrix inversion (4)
- matrix multiplication (3)
- matroid (1)
- matérn kernel (4)
- maximum independent set (3)
- maximum likelihood estimation (8)
- maximum likelihood estimator (3)
- maximum mean discrepancy (3)
- maze navigation (4)
- mean absolute error (3)
- mean estimation (5)
- mean squared error (10)
- mean-field game (2)
- mean-field theory (3)
- measurement consistency (3)
- mechanism design (5)
- mechanistic analysis (3)
- mechanistic insight (3)
- mechanistic interpretability (16)
- medical image segmentation (3)
- medical imaging (9)
- medical imaging datasets (3)
- membership inference (3)
- membership inference attacks (7)
- memorization (18)
- memory (3)
- memory bank (3)
- memory complexity (5)
- memory consumption (6)
- memory cost (3)
- memory costs (6)
- memory efficiency (27)
- memory footprint (8)
- memory optimization (3)
- memory overhead (21)
- memory reduction (3)
- memory requirements (5)
- memory usage decrease (1)
- memory-efficient (3)
- memory-efficient optimization (3)
- memory-efficient training (3)
- message passing (8)
- message passing neural networks (3)
- message-passing (9)
- meta-algorithm (4)
- meta-evaluation benchmark (3)
- meta-learning (20)
- meta-reinforcement learning (3)
- meta-training (4)
- methodology (5)
- metric space (3)
- min-max optimization (8)
- minimax optimal rates (3)
- minimax optimization (4)
- minimax regret (3)
- minimax-optimal (3)
- miou (6)
- mirror descent (4)
- misalignment (10)
- miscalibration (3)
- misclassification (3)
- misinformation (6)
- missing data (3)
- missing values (4)
- mistake bound (5)
- mitigation strategies (6)
- mixed datasets (3)
- mixture of experts (6)
- mixture-of-experts (40)
- mixture-of-experts architecture (6)
- mle-bench (3)
- mllms (11)
- mlps (5)
- mmlu (5)
- mmwave radar (3)
- modality alignment (6)
- modality collapse (3)
- modality fusion (3)
- modality gap (8)
- modality imbalance (6)
- modality-agnostic (4)
- modality-specific information (3)
- mode collapse (6)
- model accuracy (9)
- model activations (3)
- model adaptation (13)
- model aggregation (4)
- model alignment (6)
- model architecture (22)
- model architectures (17)
- model behavior (15)
- model calibration (4)
- model capabilities (5)
- model capability (5)
- model capacity (15)
- model cascade (3)
- model collapse (9)
- model comparison (10)
- model complexity (7)
- model components (4)
- model compression (13)
- model confidence (4)
- model constraints (3)
- model depth (7)
- model design (7)
- model development (3)
- model dimension (3)
- model distillation (5)
- model editing (8)
- model efficiency (8)
- model evaluation (22)
- model families (8)
- model fine-tuning (4)
- model generalizability (5)
- model generalization (23)
- model improvement (3)
- model inference (3)
- model interpretability (13)
- model limitations (3)
- model merging (10)
- model misspecification (9)
- model optimization (6)
- model outputs (5)
- model overconfidence (3)
- model overfitting (4)
- model parameters (19)
- model performance (78)
- model pre-training (3)
- model predictions (5)
- model provenance (3)
- model quality (3)
- model quantization (4)
- model refinement (4)
- model reliability (8)
- model representations (5)
- model retraining (14)
- model robustness (20)
- model safety (3)
- model scalability (11)
- model scales (5)
- model scaling (6)
- model selection (12)
- model sensitivity (5)
- model similarity (5)
- model size (15)
- model size reduction (3)
- model sizes (5)
- model stability (5)
- model suite (3)
- model supervision (3)
- model training (23)
- model uncertainty (9)
- model updates (5)
- model utility (5)
- model validation (3)
- model variants (3)
- model weights (11)
- model-agnostic (23)
- model-agnostic framework (12)
- model-based methods (3)
- model-based reinforcement learning (8)
- model-free framework (3)
- model-free methods (3)
- model-free reinforcement learning (4)
- modeling (3)
- modular architecture (5)
- modular components (4)
- modular framework (6)
- modularity (4)
- molecular boltzmann distributions (3)
- molecular conformer generation (3)
- molecular dynamics (6)
- molecular dynamics simulations (3)
- molecular generation (7)
- molecular modeling (3)
- molecular property prediction (9)
- molecular representations (3)
- molecular systems (3)
- moment retrieval (3)
- momentum (5)
- monocular depth estimation (13)
- monocular video (3)
- monocular videos (6)
- monosemantic features (3)
- monosemanticity (3)
- monotonic improvement (4)
- monotonicity (5)
- monte carlo estimation (3)
- monte carlo sampling (4)
- monte carlo tree search (19)
- motion blur (3)
- motion control (3)
- motion datasets (1)
- motion dynamics (10)
- motion encoders (1)
- motion generation (5)
- motion modeling (4)
- motion patterns (4)
- motion planning (4)
- motion primitives (4)
- motion quality (4)
- motion representation (4)
- motion trajectories (5)
- mujoco (3)
- multi-agent architecture (3)
- multi-agent collaboration (11)
- multi-agent framework (7)
- multi-agent interactions (4)
- multi-agent learning (4)
- multi-agent reinforcement learning (25)
- multi-agent system (5)
- multi-agent systems (28)
- multi-armed bandit (9)
- multi-armed bandits (4)
- multi-class classification (4)
- multi-domain datasets (3)
- multi-head attention (7)
- multi-head latent attention (4)
- multi-head self-attention (5)
- multi-hop qa (3)
- multi-hop question answering (3)
- multi-hop reasoning (5)
- multi-index models (3)
- multi-layer perceptron (3)
- multi-layer perceptrons (3)
- multi-modal (5)
- multi-modal data (4)
- multi-modal dataset (5)
- multi-modal inputs (4)
- multi-modal large language model (3)
- multi-modal large language models (17)
- multi-modal learning (5)
- multi-modal models (8)
- multi-modal reasoning (4)
- multi-modality (4)
- multi-objective optimization (22)
- multi-objective reinforcement learning (3)
- multi-scale explicit domains (1)
- multi-scale features (3)
- multi-segment explicit domains (1)
- multi-stage tasks (3)
- multi-stage training strategy (4)
- multi-step denoising (3)
- multi-step reasoning (26)
- multi-step sampling (4)
- multi-subspace explicit domains (1)
- multi-task learning (14)
- multi-task settings (3)
- multi-task training (4)
- multi-token prediction (4)
- multi-turn interaction (5)
- multi-turn interactions (4)
- multi-view clustering (6)
- multi-view consistency (5)
- multi-view inconsistencies (3)
- multi-view inputs (3)
- multi-view videos (3)
- multilayer perceptrons (7)
- multilingual capabilities (3)
- multilingual datasets (3)
- multilingual reasoning (2)
- multimodal (4)
- multimodal action distributions (3)
- multimodal alignment (5)
- multimodal applications (3)
- multimodal benchmark (4)
- multimodal benchmarks (6)
- multimodal capabilities (3)
- multimodal contrastive learning (3)
- multimodal data (7)
- multimodal dataset (12)
- multimodal datasets (10)
- multimodal distributions (4)
- multimodal fine-tuning (3)
- multimodal foundation models (4)
- multimodal framework (4)
- multimodal fusion (7)
- multimodal generation (4)
- multimodal information (3)
- multimodal inputs (7)
- multimodal integration (4)
- multimodal language models (3)
- multimodal large language model (17)
- multimodal large language models (128)
- multimodal learning (12)
- multimodal llms (10)
- multimodal modeling (3)
- multimodal models (18)
- multimodal perception (3)
- multimodal pretraining (3)
- multimodal reasoning (19)
- multimodal representation learning (5)
- multimodal representations (4)
- multimodal scenarios (3)
- multimodal sentiment analysis (3)
- multimodal systems (3)
- multimodal tasks (15)
- multimodal understanding (14)
- multiple instance learning (6)
- multiple-choice questions (5)
- multiplicative approximation (4)
- multiplicative interactions (3)
- multitask learning (4)
- multivariate data (3)
- multivariate time series (7)
- multivariate time-series (3)
- mutual dependency (3)
- mutual information (25)
- nash equilibria (6)
- nash equilibrium (13)
- natural gradient (3)
- natural images (7)
- natural language generation (4)
- natural language inference (3)
- natural language instructions (6)
- natural language processing (27)
- natural language queries (4)
- natural language understanding (6)
- near-optimal performance (3)
- near-optimality (3)
- needle-in-a-haystack (4)
- negative log-likelihood (3)
- negative pairs (3)
- negative transfer (8)
- nerf (3)
- network architecture (3)
- network architectures (8)
- network depth (4)
- network topology (4)
- neural activity (8)
- neural architecture (4)
- neural architecture search (7)
- neural architectures (7)
- neural circuits (5)
- neural collapse (7)
- neural combinatorial optimization (4)
- neural decoding (6)
- neural dynamics (5)
- neural fields (3)
- neural language models (3)
- neural models (4)
- neural network (12)
- neural network architecture (5)
- neural network architectures (7)
- neural network parameters (3)
- neural network training (10)
- neural network verification (3)
- neural odes (4)
- neural operator (3)
- neural operators (10)
- neural pde solvers (3)
- neural population activity (4)
- neural processes (4)
- neural radiance field (4)
- neural radiance fields (13)
- neural rendering (5)
- neural representations (10)
- neural retrievers (1)
- neural scaling laws (4)
- neural solvers (3)
- neural tangent kernel (10)
- neuroimaging data (3)
- neuromorphic computing (4)
- neuromorphic hardware (4)
- neuroscience (3)
- next-token prediction (9)
- next-token probabilities (3)
- nlp tasks (5)
- node classification (23)
- node embeddings (5)
- node features (3)
- node representations (6)
- node-level tasks (3)
- noise (3)
- noise contrastive estimation (3)
- noise distribution (3)
- noise injection (6)
- noise levels (7)
- noise reduction (6)
- noise robustness (6)
- noise schedule (3)
- noise sensitivity (4)
- noise suppression (4)
- noisy data (5)
- noisy environments (5)
- noisy labels (4)
- non-asymptotic analysis (3)
- non-asymptotic bounds (4)
- non-asymptotic convergence (4)
- non-convex objectives (5)
- non-convex optimization (14)
- non-convexity (3)
- non-degeneracy condition (3)
- non-euclidean data (3)
- non-formal querying (1)
- non-iid data (3)
- non-linearity (3)
- non-reasoning models (4)
- non-stationarity (9)
- non-stationary data (5)
- non-stationary environments (3)
- nonconvex optimization (4)
- nonlinear dependencies (3)
- nonlinear dynamics (6)
- nonlinear systems (3)
- nonlinearities (3)
- nonparametric regression (3)
- nonsmooth optimization (3)
- nonstationarity (3)
- normalization (5)
- normalization layers (3)
- normalizing flow (4)
- normalizing flows (8)
- novel classes (5)
- novel view synthesis (28)
- novelty (6)
- np-hard (16)
- np-hard problems (3)
- np-hardness (5)
- numerical experiments (56)
- numerical instabilities (3)
- numerical instability (4)
- numerical methods (6)
- numerical optimization (3)
- numerical precision (4)
- numerical reasoning (3)
- numerical results (3)
- numerical simulations (8)
- numerical stability (6)
- numerical studies (3)
- numerical validation (3)
- nuscenes (3)
- nuscenes benchmark (3)
- object detection (19)
- object hallucination (4)
- object localization (5)
- object re-identification (3)
- object recognition (7)
- object retrieval (3)
- object-centric representations (3)
- objective evaluations (3)
- objective function (4)
- objective metrics (4)
- observational data (18)
- occlusion (4)
- occlusions (8)
- occupancy prediction (4)
- ocean forecasting (3)
- off-policy (3)
- off-policy algorithm (3)
- off-policy algorithms (3)
- off-policy evaluation (4)
- off-policy learning (4)
- off-policy reinforcement learning (3)
- offline data (5)
- offline datasets (3)
- offline learning (6)
- offline reinforcement learning (20)
- offline setting (3)
- one-shot federated learning (3)
- one-step diffusion (5)
- online adaptation (4)
- online coloring (1)
- online conformal prediction (3)
- online convex optimization (7)
- online decision-making (3)
- online gradient descent (3)
- online learning (24)
- online learning framework (3)
- online linear optimization (3)
- online mirror descent (7)
- online optimization (7)
- online reinforcement learning (3)
- online rlhf (3)
- online stochastic gradient descent (3)
- online-to-batch conversion (3)
- ood detection (3)
- opacity (3)
- open challenge (4)
- open question (3)
- open-ended action spaces (1)
- open-ended generation (3)
- open-source (6)
- open-source datasets (5)
- open-source library (3)
- open-source llms (7)
- open-source models (14)
- open-sourced llms (3)
- open-vocabulary (7)
- open-vocabulary semantic segmentation (4)
- open-weight models (4)
- open-world recognition (2)
- open-world settings (3)
- optical character recognition (4)
- optical flow (6)
- optical flow estimation (3)
- optimal actions (3)
- optimal convergence rates (4)
- optimal decision-making (3)
- optimal offline fairness (1)
- optimal performance (3)
- optimal policies (7)
- optimal policy (17)
- optimal regret (3)
- optimal solution (6)
- optimal solutions (7)
- optimal transport (33)
- optimality (13)
- optimality gap (7)
- optimality guarantees (3)
- optimization (52)
- optimization algorithm (7)
- optimization algorithms (25)
- optimization challenges (5)
- optimization dynamics (7)
- optimization framework (19)
- optimization hyperparameters (3)
- optimization landscape (4)
- optimization method (3)
- optimization methods (13)
- optimization objective (6)
- optimization performance (6)
- optimization problem (26)
- optimization problems (11)
- optimization procedure (3)
- optimization process (9)
- optimization properties (4)
- optimization steps (5)
- optimization strategies (3)
- optimization tasks (4)
- optimization techniques (21)
- optimization theory (5)
- optimization trajectory (3)
- optimization-based approaches (3)
- optimization-based methods (3)
- optimizer states (3)
- ordinary differential equation (4)
- ordinary differential equations (6)
- orthogonal matrix (3)
- orthogonal projection (4)
- orthogonal transformations (5)
- orthogonality (4)
- out of distribution generalization (3)
- out-of-distribution (15)
- out-of-distribution benchmarks (4)
- out-of-distribution data (7)
- out-of-distribution datasets (4)
- out-of-distribution detection (16)
- out-of-distribution generalisation (3)
- out-of-distribution generalization (19)
- out-of-distribution performance (3)
- out-of-distribution samples (6)
- out-of-distribution scenarios (6)
- out-of-distribution tasks (8)
- out-of-domain generalization (7)
- out-of-domain tasks (3)
- outlier removal (3)
- outlier types (2)
- output distribution (8)
- output diversity (3)
- output language control (1)
- output logits (3)
- output quality (5)
- over-smoothing (9)
- over-squashing (4)
- overconfidence (4)
- overfitting (79)
- overparameterization (9)
- overparameterized models (4)
- overparameterized networks (3)
- overparameterized neural networks (3)
- oversmoothing (7)
- overthinking (7)
- pac learning (4)
- pairwise comparisons (9)
- pan-sharpening (3)
- parallel token generation (3)
- parallel training (5)
- parallelism (4)
- parallelization (7)
- parameter complexity (3)
- parameter count (7)
- parameter efficiency (27)
- parameter estimation (7)
- parameter interference (3)
- parameter models (3)
- parameter optimization (3)
- parameter scaling (10)
- parameter selection (4)
- parameter sharing (4)
- parameter space (9)
- parameter symmetries (3)
- parameter tuning (5)
- parameter updates (9)
- parameter-efficient (7)
- parameter-efficient adaptation (5)
- parameter-efficient fine-tuning (43)
- parameter-efficient finetuning (3)
- parameter-efficient tuning (3)
- parameterization (10)
- parameterizations (5)
- parameterized greedy policy (1)
- parametric knowledge (3)
- pareto front (4)
- pareto frontier (13)
- pareto-optimal solutions (5)
- partial differential equations (20)
- partial observability (12)
- partial observations (3)
- pass@1 (3)
- pearson correlation (3)
- perception (5)
- perception models (3)
- perceptual features (1)
- perceptual metrics (3)
- perceptual quality (6)
- performance analysis (6)
- performance assessment (8)
- performance baselines (4)
- performance benchmarking (6)
- performance benchmarks (17)
- performance bottlenecks (3)
- performance bounds (3)
- performance ceiling (3)
- performance characterization (4)
- performance comparison (20)
- performance decline (3)
- performance degradation (73)
- performance differences (3)
- performance disparities (3)
- performance disparity (3)
- performance drop (6)
- performance enhancement (47)
- performance estimation (4)
- performance evaluation (89)
- performance gain (5)
- performance gains (40)
- performance gap (15)
- performance gaps (7)
- performance guarantees (19)
- performance improvement (95)
- performance improvements (28)
- performance loss (3)
- performance metrics (27)
- performance optimization (38)
- performance preservation (5)
- performance scaling (4)
- performance stability (3)
- performance validation (7)
- performance-efficiency trade-offs (3)
- performative prediction (3)
- periodicity (6)
- permutation invariance (5)
- perplexity (13)
- perplexity objective (1)
- persistent homology (4)
- personalization (6)
- personalized federated learning (3)
- perturbation (5)
- perturbation analysis (4)
- perturbation bounds (3)
- perturbations (17)
- phase transition (9)
- phase transitions (5)
- photoplethysmography (4)
- photorealistic rendering (5)
- physical consistency (3)
- physical constraints (7)
- physical fidelity (3)
- physical plausibility (9)
- physical priors (5)
- physical systems (3)
- physically based rendering (3)
- physics-informed machine learning (4)
- physics-informed neural networks (17)
- physiological signals (4)
- pixel space (3)
- pixel-level annotations (3)
- pixel-level segmentation (3)
- planning (9)
- planning algorithms (3)
- planning performance (6)
- plasticity (9)
- platonic representation hypothesis (3)
- plug-and-play (3)
- plug-and-play framework (6)
- plug-and-play methods (3)
- plug-and-play module (3)
- plug-and-play solution (5)
- plug-in module (4)
- point cloud registration (3)
- point clouds (13)
- pointmaps (3)
- poisoning attacks (3)
- policy alignment (3)
- policy class (2)
- policy distribution (3)
- policy evaluation (9)
- policy exploration (3)
- policy generation (3)
- policy gradient (3)
- policy gradient methods (3)
- policy gradients (4)
- policy improvement (6)
- policy initialization (3)
- policy learning (18)
- policy optimization (28)
- policy reuse (3)
- policy training (3)
- policy update (3)
- polylogarithmic factors (4)
- polynomial complexity (3)
- polynomial scaling (5)
- polynomial time (8)
- polynomial-time (3)
- polynomial-time algorithm (6)
- pomdps (3)
- population risk (3)
- pose estimation (5)
- positional bias (3)
- positional embeddings (4)
- positional encoding (7)
- positional encodings (9)
- positive pairs (4)
- post-hoc analysis (3)
- post-training (19)
- post-training framework (4)
- post-training quantization (12)
- posterior approximation (4)
- posterior distribution (10)
- posterior estimation (3)
- posterior inference (4)
- posterior sampling (9)
- potential outcomes (3)
- power-law distribution (5)
- power-law scaling (3)
- ppo (5)
- practical algorithm (3)
- practical applicability (6)
- practical application (4)
- practical applications (11)
- practical deployment (6)
- practical effectiveness (4)
- practical efficiency (3)
- practical insights (3)
- practical utility (4)
- pre-trained foundation models (3)
- pre-trained language models (3)
- pre-trained llms (3)
- pre-trained model (8)
- pre-trained models (46)
- pre-trained transformers (4)
- pre-trained vision-language models (3)
- pre-training (29)
- precision (12)
- preconditioners (3)
- preconditioning (5)
- predict-then-optimize (3)
- prediction (8)
- prediction accuracy (20)
- prediction error (6)
- prediction errors (4)
- prediction heads (3)
- prediction intervals (5)
- prediction performance (5)
- prediction quality (6)
- prediction sets (8)
- prediction tasks (5)
- prediction uncertainty (3)
- prediction-powered inference (3)
- predictions (4)
- predictive accuracy (23)
- predictive capabilities (3)
- predictive coding (3)
- predictive distributions (4)
- predictive modeling (9)
- predictive models (8)
- predictive performance (22)
- predictive power (8)
- predictive tasks (5)
- predictive uncertainty (4)
- preference alignment (11)
- preference dataset (5)
- preference datasets (4)
- preference learning (11)
- preference model (4)
- preference optimization (16)
- preference pairs (6)
- preference-based reinforcement learning (8)
- preprocessing (3)
- pretrained diffusion model (3)
- pretrained diffusion models (3)
- pretrained language models (7)
- pretrained model (5)
- pretrained models (26)
- pretrained video diffusion models (3)
- pretrained vision-language models (5)
- pretrained weights (5)
- pretraining (24)
- pretraining framework (3)
- pretraining objectives (3)
- primal-dual algorithm (5)
- primal-dual methods (3)
- principal component analysis (4)
- principal components (3)
- principled foundation (3)
- prior distribution (7)
- prior knowledge (9)
- privacy (3)
- privacy amplification (4)
- privacy budget (6)
- privacy concerns (10)
- privacy constraints (3)
- privacy guarantees (9)
- privacy leakage (7)
- privacy protection (4)
- privacy risk (3)
- privacy risks (8)
- privacy-preserving (9)
- privacy-utility trade-off (4)
- privacy-utility trade-offs (3)
- probabilistic inference (9)
- probabilistic model (7)
- probabilistic modeling (3)
- probabilistic models (7)
- probabilistic predictions (4)
- probabilistic scoring functions (1)
- probability distribution (4)
- probability distributions (15)
- probability estimates (3)
- probability measures (5)
- probability simplex (3)
- problem complexity (4)
- problem-solving (8)
- procedural generation (5)
- process reward model (3)
- process reward models (10)
- program synthesis (7)
- programming languages (4)
- progressive training strategy (3)
- projection (3)
- prompt alignment (5)
- prompt embeddings (3)
- prompt engineering (10)
- prompt injection attacks (3)
- prompt learning (5)
- prompt optimization (7)
- prompt tuning (5)
- prompt-aware diversity (1)
- prompt-based methods (3)
- prompting strategies (4)
- proper scoring rules (3)
- proprietary models (14)
- proprioception (3)
- protein design (5)
- protein dynamics (3)
- protein engineering (3)
- protein language models (5)
- protein structures (3)
- protein-ligand complexes (4)
- protein-protein interactions (4)
- provable guarantees (4)
- provable speedups (1)
- proximal operator (3)
- proximal policy optimization (7)
- proxy model (3)
- pruning (12)
- pruning methods (4)
- pruning strategies (4)
- pseudo labels (4)
- pseudo-labeling (5)
- pseudo-labels (15)
- psnr (8)
- public benchmarks (6)
- public datasets (6)
- pytorch (3)
- q-learning (6)
- q-values (6)
- qa pairs (5)
- quadratic assignment problem (3)
- quadratic complexity (10)
- quadratic programming (3)
- quadratic time complexity (5)
- qualitative analysis (6)
- qualitative evaluation (4)
- qualitative evaluations (9)
- qualitative experiments (4)
- qualitative results (3)
- quality assessment (3)
- quality control (3)
- quality degradation (5)
- quality scores (4)
- quantile regression (3)
- quantitative analysis (9)
- quantitative evaluation (8)
- quantitative evaluations (10)
- quantitative experiments (3)
- quantitative metrics (6)
- quantitative performance (3)
- quantization (18)
- quantization error (3)
- quantization-aware training (3)
- quantum algorithms (4)
- quantum chemistry (3)
- quantum computing (3)
- query complexity (11)
- query efficiency (3)
- question answering (12)
- question-answer pairs (11)
- question-answering (7)
- qwen2.5-vl (7)
- rademacher complexity (7)
- rag systems (3)
- random arrival order (1)
- random forests (3)
- random fourier features (3)
- random matrix theory (7)
- random sampling (4)
- randomized controlled trials (5)
- randomized smoothing (3)
- rank (2)
- rapid adaptation (4)
- rapid convergence (3)
- rare pathologies (2)
- raw images (3)
- re-identification (3)
- real data (5)
- real datasets (9)
- real-time applications (3)
- real-time decoding (4)
- real-time inference (4)
- real-time rendering (4)
- real-world applicability (5)
- real-world benchmarks (12)
- real-world challenges (3)
- real-world coding tasks (1)
- real-world data (14)
- real-world dataset (5)
- real-world datasets (71)
- real-world deployment (9)
- real-world environments (5)
- real-world experiments (4)
- real-world networks (3)
- real-world scenarios (16)
- real-world tasks (6)
- realizable case (3)
- realizable setting (3)
- reasoning (32)
- reasoning abilities (11)
- reasoning ability (7)
- reasoning accuracy (6)
- reasoning benchmarks (27)
- reasoning capabilities (49)
- reasoning capability (4)
- reasoning chain (3)
- reasoning chains (3)
- reasoning complexity (3)
- reasoning depth (8)
- reasoning efficiency (4)
- reasoning fidelity (3)
- reasoning model (3)
- reasoning models (27)
- reasoning paths (15)
- reasoning patterns (5)
- reasoning performance (18)
- reasoning process (10)
- reasoning processes (3)
- reasoning segmentation (4)
- reasoning steps (3)
- reasoning strategies (6)
- reasoning tasks (27)
- reasoning traces (13)
- reasoning trajectories (10)
- reasoning-intensive tasks (3)
- recall (6)
- receptive field (4)
- receptive fields (3)
- recognition accuracy (4)
- recommendation systems (7)
- recommender systems (12)
- reconstruction (3)
- reconstruction accuracy (6)
- reconstruction error (9)
- reconstruction errors (4)
- reconstruction fidelity (12)
- reconstruction loss (4)
- reconstruction performance (3)
- reconstruction quality (11)
- reconstruction-based methods (3)
- rectified flow (6)
- recurrent models (6)
- recurrent neural network (3)
- recurrent neural networks (22)
- recurrent state-space model (3)
- red teaming (3)
- red-teaming (7)
- redundancy (6)
- redundancy reduction (6)
- redundant computation (3)
- reference model (3)
- reference policy (4)
- refined analysis (3)
- register tokens (3)
- regression (10)
- regression benchmarks (3)
- regression models (3)
- regression task (3)
- regression tasks (14)
- regret (9)
- regret analysis (16)
- regret bound (14)
- regret bounds (13)
- regret guarantees (6)
- regret lower bound (4)
- regret minimization (17)
- regret scaling (5)
- regret upper bound (4)
- regularity assumptions (3)
- regularization (17)
- regularization strength (4)
- regularization techniques (4)
- regularization term (12)
- regularization terms (3)
- regulatory frameworks (3)
- reinforce (3)
- reinforcement fine-tuning (15)
- reinforcement learning (384)
- reinforcement learning algorithm (4)
- reinforcement learning algorithms (3)
- reinforcement learning framework (8)
- reinforcement learning from human feedback (12)
- reinforcement learning with verifiable rewards (3)
- rejection sampling (8)
- relational databases (4)
- relational learning (4)
- relational reasoning (5)
- relative positional encoding (3)
- relevance (3)
- reliability (13)
- relu activation (3)
- relu activation function (3)
- relu networks (7)
- remote sensing (17)
- removal attacks (3)
- rendering quality (10)
- reparameterization (4)
- replay buffer (5)
- report generation (3)
- representation alignment (6)
- representation collapse (3)
- representation extraction (3)
- representation learning (61)
- representation power (3)
- representation quality (5)
- representation space (7)
- representation spaces (3)
- representation superposition (3)
- representation vectors (4)
- representational alignment (6)
- representational capacity (7)
- representational collapse (4)
- representational geometry (3)
- representational power (5)
- representational quality (4)
- representational similarity (5)
- representations (7)
- representer theorem (3)
- reproducibility (19)
- reproducible evaluation (4)
- reproducible research (4)
- reproducing kernel hilbert space (9)
- reproducing kernel hilbert spaces (3)
- research community (3)
- research directions (3)
- residual connections (3)
- residual networks (4)
- residual stream (7)
- resilience (4)
- resnet (5)
- resource allocation (13)
- resource constraints (8)
- resource demands (3)
- resource efficiency (4)
- resource utilization (3)
- resource-constrained devices (5)
- resource-constrained environments (9)
- resource-intensive (3)
- response generation (5)
- response length (4)
- response quality (5)
- retinex theory (3)
- retraining (4)
- retrieval (5)
- retrieval accuracy (5)
- retrieval augmented generation (5)
- retrieval performance (5)
- retrieval precision (3)
- retrieval tasks (9)
- retrieval-augmented generation (37)
- return distribution (3)
- reverse-engineering (4)
- reward design (4)
- reward distribution (3)
- reward distributions (3)
- reward engineering (3)
- reward feedback (3)
- reward function (14)
- reward functions (14)
- reward hacking (14)
- reward maximization (7)
- reward mechanism (3)
- reward model (15)
- reward modeling (5)
- reward models (20)
- reward performance (3)
- reward shaping (8)
- reward signal (4)
- reward signals (4)
- rgb (3)
- rgb-d cameras (4)
- ridge regression (4)
- riemannian manifold (7)
- riemannian manifolds (7)
- riemannian optimization (3)
- rigorous evaluation (3)
- risk assessment (5)
- risk control (4)
- risk management (4)
- rkhs (3)
- rlhf (6)
- rmse (3)
- roberta-large (4)
- robot manipulation (6)
- robotic manipulation (24)
- robotics (10)
- robotics tasks (3)
- robust adaptation (4)
- robust agents (4)
- robust estimation (5)
- robust generalization (16)
- robust learning (7)
- robust models (6)
- robust optimization (12)
- robust performance (10)
- robust representations (8)
- robustness (170)
- robustness analysis (3)
- robustness assessment (4)
- robustness enhancement (8)
- robustness evaluation (4)
- robustness improvement (3)
- robustness to noise (4)
- rollouts (5)
- rope (3)
- rotary position embedding (4)
- rotary positional encodings (4)
- routing mechanisms (3)
- routing problems (3)
- rule-based reinforcement learning (5)
- rule-based rewards (3)
- running time (3)
- runtime analysis (3)
- runtime complexity (4)
- runtime efficiency (6)
- runtime reduction (4)
- safe reinforcement learning (5)
- safety (10)
- safety alignment (25)
- safety constraints (14)
- safety evaluation (3)
- safety guarantees (5)
- safety mechanisms (5)
- safety performance (3)
- safety risks (3)
- safety verification (3)
- safety-critical applications (4)
- safety-critical domains (5)
- safety-critical scenarios (3)
- sample complexities (3)
- sample complexity (70)
- sample complexity bounds (6)
- sample diversity (8)
- sample efficiency (62)
- sample inefficiency (4)
- sample quality (12)
- sample selection (3)
- sample size (7)
- sample-efficiency (3)
- sample-efficient learning (3)
- sampling (10)
- sampling algorithms (4)
- sampling complexity (3)
- sampling efficiency (23)
- sampling error (3)
- sampling method (3)
- sampling methods (8)
- sampling procedures (4)
- sampling process (5)
- sampling quality (3)
- sampling scheme (3)
- sampling steps (6)
- sampling strategies (4)
- sampling strategy (5)
- sampling trajectory (3)
- sampling-based scaling (1)
- satellite imagery (5)
- scalability (100)
- scalability challenges (7)
- scalability issues (3)
- scalable algorithm (4)
- scalable algorithms (7)
- scalable alternative (3)
- scalable approach (3)
- scalable framework (6)
- scalable inference (3)
- scalable learning (5)
- scalable methods (7)
- scalable oversight (3)
- scalable solution (7)
- scalable training (8)
- scalable vector graphics (3)
- scaling behavior (6)
- scaling efficiency (3)
- scaling experiments (3)
- scaling law (7)
- scaling laws (19)
- scannet (3)
- scene consistency (3)
- scene diversity (4)
- scene dynamics (3)
- scene geometry (7)
- scene reconstruction (7)
- scene representation (3)
- scene understanding (17)
- schrödinger bridge (7)
- scientific discovery (4)
- scientific machine learning (6)
- score estimation (4)
- score function (10)
- score matching (4)
- score-based diffusion models (4)
- score-based generative models (7)
- score-based methods (3)
- scoring function (3)
- scoring functions (3)
- search algorithms (3)
- search efficiency (4)
- search space (6)
- search strategies (3)
- second-order methods (5)
- security evaluation (3)
- security risks (4)
- security vulnerabilities (4)
- segment anything model (8)
- segmentation (13)
- segmentation masks (3)
- segmentation models (3)
- segmentation performance (3)
- segmentation tasks (4)
- selection bias (6)
- self-attention (27)
- self-attention layers (3)
- self-attention mechanism (6)
- self-attention mechanisms (4)
- self-consistency (5)
- self-correction (11)
- self-distillation (7)
- self-evolution (4)
- self-improvement (10)
- self-play (4)
- self-play fine-tuning (3)
- self-reflection (9)
- self-similarity (3)
- self-supervised framework (10)
- self-supervised learning (67)
- self-supervised models (4)
- self-supervised objectives (3)
- self-supervised pretraining (6)
- self-supervised representation learning (4)
- self-supervision (3)
- self-verification (4)
- semantic alignment (26)
- semantic anchors (3)
- semantic attributes (3)
- semantic coherence (10)
- semantic concepts (3)
- semantic consistency (21)
- semantic content (6)
- semantic context (5)
- semantic correctness (3)
- semantic cues (5)
- semantic differences (4)
- semantic equivalence (3)
- semantic features (7)
- semantic fidelity (12)
- semantic grounding (4)
- semantic information (18)
- semantic knowledge (3)
- semantic misalignment (5)
- semantic organization (3)
- semantic patterns (3)
- semantic reasoning (8)
- semantic relationships (9)
- semantic relevance (6)
- semantic representations (8)
- semantic segmentation (31)
- semantic shifts (3)
- semantic similarity (12)
- semantic spectrum (1)
- semantic structure (3)
- semantic understanding (14)
- semantickitti (3)
- semantics (5)
- semi-automatic data construction (3)
- semi-structured data (1)
- semi-structured retrieval benchmark (1)
- semi-supervised learning (21)
- semiparametric efficiency (3)
- sensitive information (4)
- sensitivity (6)
- sensitivity analysis (11)
- sensor fusion (3)
- separability (3)
- sequence generation (5)
- sequence length (4)
- sequence lengths (5)
- sequence modeling (11)
- sequence models (4)
- sequence parallelism (3)
- sequence processing (3)
- sequential data (5)
- sequential decision process (3)
- sequential decision-making (19)
- sequential modeling (4)
- sequential monte carlo (7)
- sequential parsing (1)
- sequential reasoning (8)
- sequential recommendation (4)
- service-level objectives (4)
- sgd (4)
- shallow architectures (4)
- shape features (3)
- shape representation (3)
- shapley values (8)
- shared latent space (5)
- shared representation (3)
- sharpness-aware minimization (6)
- short-term memory (2)
- shortcut learning (4)
- shortcut models (3)
- side information (4)
- signal processing (4)
- signal strength (3)
- signal-to-noise ratio (8)
- signed distance fields (3)
- signed distance functions (3)
- signed graphs (3)
- sim-to-real gap (4)
- sim-to-real transfer (4)
- simclr (3)
- similarity metrics (3)
- simplicial complexes (5)
- simulated data (5)
- simulated datasets (3)
- simulation (12)
- simulation accuracy (3)
- simulation environments (4)
- simulation experiments (4)
- simulation studies (4)
- simulation-based inference (4)
- simulations (13)
- single-index models (4)
- singular value decomposition (11)
- singular values (4)
- skip connections (4)
- small language models (7)
- smiles (3)
- smooth functions (4)
- smoothed analysis (3)
- smoothness (10)
- sobolev norm (3)
- social interaction (3)
- social networks (3)
- social reasoning (4)
- social welfare (12)
- soft labels (3)
- softmax attention (12)
- softmax policies (1)
- software engineering (4)
- solution quality (10)
- solution space (5)
- sota baselines (3)
- sota methods (12)
- sota performance (11)
- source-free domain adaptation (4)
- space complexity (3)
- sparse attention (10)
- sparse autoencoders (20)
- sparse coding (3)
- sparse reward (3)
- sparse rewards (5)
- sparsification (3)
- sparsity (21)
- spatial alignment (6)
- spatial attention (1)
- spatial awareness (5)
- spatial coherence (5)
- spatial consistency (7)
- spatial constraints (4)
- spatial dependencies (4)
- spatial details (3)
- spatial distribution (3)
- spatial features (4)
- spatial generalization (3)
- spatial grounding (4)
- spatial information (6)
- spatial intelligence (5)
- spatial localization (3)
- spatial perception (5)
- spatial reasoning (41)
- spatial relations (6)
- spatial relationship understanding (1)
- spatial relationships (15)
- spatial resolution (9)
- spatial structure (6)
- spatial transcriptomics (6)
- spatial understanding (7)
- spatio-temporal consistency (3)
- spatio-temporal information (3)
- spatiotemporal dynamics (6)
- spatiotemporal reasoning (4)
- specialized domains (3)
- specialized models (6)
- spectral analysis (3)
- spectral bias (7)
- spectral clustering (4)
- spectral decomposition (3)
- spectral domain (3)
- spectral filters (3)
- spectral norm (5)
- spectral properties (8)
- speculative decoding (24)
- speech recognition (4)
- speed-up (4)
- speedup (19)
- speedups (5)
- spherical harmonics (4)
- spiking activity (4)
- spiking neural networks (24)
- spiking transformers (3)
- spurious correlation (4)
- spurious correlations (23)
- spurious features (3)
- square loss (4)
- squared exponential kernel (4)
- squared loss (4)
- ssim (3)
- stability (32)
- stability analysis (3)
- stability enhancement (3)
- stability guarantees (3)
- stable convergence (3)
- stable diffusion (9)
- stable learning (3)
- standard benchmarks (7)
- standardized benchmarks (7)
- standardized datasets (5)
- standardized evaluation framework (3)
- state estimation (3)
- state of the art (15)
- state representation (3)
- state space (4)
- state space models (20)
- state transitions (5)
- state-action pairs (3)
- state-of-the-art accuracy (11)
- state-of-the-art algorithms (7)
- state-of-the-art alternatives (3)
- state-of-the-art approaches (15)
- state-of-the-art baselines (27)
- state-of-the-art benchmarks (5)
- state-of-the-art llms (3)
- state-of-the-art methods (100)
- state-of-the-art models (32)
- state-of-the-art performance (160)
- state-of-the-art results (29)
- state-of-the-art scores (3)
- state-of-the-art techniques (6)
- state-space models (14)
- stationarity (3)
- stationary point (3)
- stationary points (7)
- statistical analysis (7)
- statistical complexity (5)
- statistical dependence (3)
- statistical efficiency (11)
- statistical guarantees (8)
- statistical inference (18)
- statistical learning (6)
- statistical learning theory (7)
- statistical performance (5)
- statistical physics (4)
- statistical properties (11)
- statistical significance (5)
- statistical tests (3)
- statistical validity (3)
- stealthiness (5)
- steepest descent (3)
- steering vectors (3)
- step-by-step reasoning (5)
- step-level supervision (3)
- stereo depth estimation (3)
- stochastic approximation (5)
- stochastic block model (4)
- stochastic control (3)
- stochastic differential equation (5)
- stochastic differential equations (13)
- stochastic dynamics (5)
- stochastic environments (4)
- stochastic gradient descent (31)
- stochastic gradient noise (3)
- stochastic gradient optimization (3)
- stochastic interpolants (4)
- stochastic model (3)
- stochastic nature (3)
- stochastic optimal control (5)
- stochastic optimization (14)
- stochastic process (4)
- stochastic processes (7)
- stochasticity (15)
- storage efficiency (4)
- straight-through estimator (3)
- strategic classification (7)
- strategic prompts (1)
- streaming data (4)
- stress-test (2)
- strong baselines (3)
- strong convexity (4)
- strongly convex (5)
- structural assumptions (6)
- structural biases (3)
- structural causal model (5)
- structural causal models (3)
- structural characteristics (3)
- structural coherence (6)
- structural consistency (15)
- structural constraints (6)
- structural diversity (4)
- structural features (4)
- structural fidelity (7)
- structural guidance (4)
- structural heterogeneity (5)
- structural information (11)
- structural patterns (3)
- structural priors (4)
- structural properties (6)
- structural reasoning (2)
- structural representations (3)
- structural uncertainty (3)
- structural understanding (3)
- structure prediction (3)
- structure-based drug design (5)
- structure-from-motion (7)
- structured data (9)
- structured generation (3)
- structured perturbations (3)
- structured priors (3)
- structured pruning (6)
- structured reasoning (5)
- structured representations (4)
- student model (6)
- student models (3)
- style consistency (3)
- style transfer (4)
- sub-linear regret (3)
- sub-optimality (3)
- sub-optimality gap (4)
- sublinear regret (9)
- suboptimal decisions (3)
- suboptimal performance (3)
- suboptimality gap (3)
- subproblems (3)
- subset selection (4)
- success probability (4)
- success rate (8)
- success rates (3)
- sudoku (3)
- sufficient condition (3)
- sufficient conditions (3)
- summarization (5)
- super-resolution (10)
- superior performance (13)
- supervised classification (3)
- supervised fine-tuning (101)
- supervised finetuning (4)
- supervised learning (24)
- supervised methods (3)
- supervised training (3)
- supervision (4)
- supervision signals (4)
- supervisory signals (4)
- surface reconstruction (9)
- surrogate gradients (4)
- surrogate losses (3)
- surrogate model (8)
- surrogate modeling (3)
- surrogate models (9)
- swap regret (3)
- swe-bench (3)
- symbolic reasoning (5)
- symbolic regression (4)
- symbolic verification (3)
- symmetric matrices (3)
- symmetric positive definite (3)
- symmetries (3)
- symmetry priors (3)
- synaptic plasticity (5)
- syntactic correctness (3)
- synthetic benchmarks (15)
- synthetic biology (4)
- synthetic data (63)
- synthetic data generation (16)
- synthetic dataset (13)
- synthetic datasets (55)
- synthetic examples (3)
- synthetic experiments (4)
- synthetic graphs (4)
- synthetic images (8)
- synthetic samples (6)
- synthetic tasks (6)
- system dynamics (3)
- system identification (3)
- systematic analysis (8)
- systematic assessment (3)
- systematic evaluation (16)
- systematic investigation (6)
- systematic reasoning (5)
- systematic review (3)
- systematic study (5)
- t2i models (3)
- tabular data (16)
- tabular datasets (4)
- tabular foundation models (4)
- tactile sensing (5)
- target accuracy (3)
- target distribution (8)
- targeted interventions (3)
- task accuracy (3)
- task adaptation (4)
- task complexity (11)
- task decomposition (8)
- task difficulty (7)
- task diversity (4)
- task generalization (7)
- task instructions (3)
- task interference (3)
- task performance (15)
- task planning (6)
- task structure (3)
- task success (3)
- task success rates (3)
- task transfer (3)
- task vectors (3)
- task-relevant information (7)
- task-specific adaptation (3)
- task-specific features (4)
- task-specific fine-tuning (6)
- task-specific knowledge (8)
- task-specific models (5)
- task-specific parameters (3)
- task-specific performance (4)
- task-specific rewards (3)
- task-specific training (5)
- taxonomy (6)
- teacher model (8)
- teacher models (3)
- teacher-specific adapters (1)
- teacher-student framework (3)
- temperature scaling (5)
- temporal alignment (4)
- temporal coherence (15)
- temporal consistency (19)
- temporal context (7)
- temporal dependencies (20)
- temporal difference learning (4)
- temporal distance (4)
- temporal dynamics (21)
- temporal evolution (3)
- temporal granularity (3)
- temporal grounding (4)
- temporal information (10)
- temporal modeling (7)
- temporal patterns (5)
- temporal reasoning (11)
- temporal redundancy (5)
- temporal representations (3)
- temporal resolution (10)
- temporal scales (6)
- temporal structure (5)
- temporal understanding (5)
- temporal variations (5)
- tensor decomposition (4)
- tensor parallelism (3)
- test accuracy (7)
- test cases (4)
- test error (3)
- test performance (4)
- test-time adaptation (33)
- test-time computation (4)
- test-time compute (7)
- test-time optimization (3)
- test-time scaling (32)
- text classification (9)
- text datasets (3)
- text embeddings (10)
- text encoder (8)
- text generation (9)
- text localization (2)
- text prompts (4)
- text quality (7)
- text-attributed graphs (3)
- text-based games (1)
- text-driven image editing (3)
- text-guided image editing (8)
- text-image alignment (3)
- text-to-image (13)
- text-to-image benchmarks (3)
- text-to-image diffusion (5)
- text-to-image diffusion models (20)
- text-to-image generation (42)
- text-to-image models (20)
- text-to-image retrieval (4)
- text-to-image synthesis (4)
- text-to-motion generation (3)
- text-to-speech (3)
- text-to-sql (3)
- text-to-video (5)
- text-to-video diffusion models (4)
- text-to-video generation (6)
- textual descriptions (8)
- textual prompts (4)
- theoretical analyses (11)
- theoretical analysis (140)
- theoretical bounds (3)
- theoretical characterization (4)
- theoretical complexity (3)
- theoretical convergence (6)
- theoretical convergence guarantees (5)
- theoretical findings (8)
- theoretical foundation (7)
- theoretical foundations (12)
- theoretical framework (26)
- theoretical frameworks (3)
- theoretical guarantee (7)
- theoretical guarantees (60)
- theoretical insight (4)
- theoretical insights (18)
- theoretical justification (9)
- theoretical perspective (3)
- theoretical proof (3)
- theoretical properties (5)
- theoretical results (18)
- theoretical study (5)
- theoretical understanding (13)
- theoretical validation (3)
- theory of mind (3)
- thinking length (1)
- thompson sampling (13)
- thresholded classifiers (1)
- throughput (14)
- throughput improvement (3)
- tight bounds (4)
- tighter bounds (3)
- time complexity (13)
- time discretization (3)
- time horizon (6)
- time series (9)
- time series analysis (8)
- time series classification (3)
- time series forecasting (18)
- time series foundation models (5)
- time windows (3)
- time-series analysis (3)
- time-series forecasting (3)
- time-to-first-token (6)
- timescales (4)
- token budget (6)
- token compression (5)
- token consumption (4)
- token distribution (3)
- token efficiency (7)
- token embeddings (5)
- token entropy (3)
- token generation (3)
- token importance (4)
- token importance scores (1)
- token prediction (4)
- token pruning (6)
- token reduction (7)
- token representations (4)
- token selection (3)
- token sequences (4)
- token usage (6)
- token-level counterfactuals (1)
- token-level reasoning (1)
- tokenization (4)
- tokenizers (3)
- tokens (3)
- tool use (3)
- top-1 accuracy (4)
- total variation distance (6)
- toxicity mitigation (3)
- traceability (3)
- tracking (6)
- tracking performance (5)
- tractability (3)
- trade-off (18)
- trade-off analysis (4)
- trade-offs (14)
- tradeoff (3)
- trainable parameters (10)
- training acceleration (5)
- training algorithm (3)
- training algorithms (3)
- training batches (3)
- training buffer (3)
- training cost (4)
- training costs (3)
- training data (30)
- training data attribution (4)
- training data efficiency (5)
- training data quality (3)
- training data scarcity (3)
- training data selection (6)
- training dataset (5)
- training datasets (8)
- training distribution (8)
- training dynamic (3)
- training dynamics (50)
- training efficiency (47)
- training examples (6)
- training flops (5)
- training framework (5)
- training instability (6)
- training iterations (4)
- training loss (5)
- training methodology (3)
- training methods (4)
- training objective (5)
- training objectives (3)
- training optimization (3)
- training overhead (5)
- training paradigm (6)
- training paradigms (5)
- training performance (5)
- training pipeline (4)
- training procedures (3)
- training recipe (3)
- training recipes (3)
- training samples (9)
- training scheme (3)
- training set (4)
- training speed (6)
- training speedup (3)
- training stability (21)
- training strategies (8)
- training strategy (3)
- training time (6)
- training time reduction (5)
- training tokens (3)
- training-based methods (3)
- training-free (12)
- training-free approach (10)
- training-free framework (12)
- training-free manner (3)
- training-free method (7)
- training-free methods (6)
- trajectories (6)
- trajectory alignment (3)
- trajectory analysis (3)
- trajectory optimization (5)
- trajectory planning (6)
- trajectory prediction (5)
- trajectory stitching (3)
- transfer learning (29)
- transfer performance (3)
- transferability (25)
- transformer (19)
- transformer architecture (28)
- transformer architectures (17)
- transformer blocks (6)
- transformer encoder (3)
- transformer language models (4)
- transformer layers (5)
- transformer model (5)
- transformer models (19)
- transformer-based (3)
- transformer-based approaches (3)
- transformer-based language models (5)
- transformer-based model (7)
- transformers (70)
- transition dynamics (4)
- transparency (9)
- transport maps (3)
- traveling salesman problem (6)
- treatment effects (7)
- tree search (4)
- tree structure (3)
- tree-based models (3)
- triangle inequality (3)
- trustworthiness (7)
- trustworthy ai (3)
- truthful reporting (4)
- tsallis entropy (3)
- turbulent flows (3)
- two-layer neural networks (3)
- two-stage framework (4)
- two-stage pipeline (3)
- two-stage pipelines (3)
- two-stage training strategy (6)
- u-net (4)
- u-net architecture (3)
- unbalanced optimal transport (3)
- unbiased estimate (3)
- uncertainties (3)
- uncertainty (9)
- uncertainty calibration (7)
- uncertainty estimates (10)
- uncertainty estimation (15)
- uncertainty modeling (6)
- uncertainty quantification (57)
- uncertainty set (5)
- unconditional generation (5)
- unconstrained optimization (4)
- underfitting (3)
- unified framework (17)
- unified model (3)
- uniform distribution (5)
- uniform sampling (3)
- uniqueness (3)
- unit tests (3)
- universal approximation property (6)
- universal approximator (3)
- universality (6)
- unlabeled data (15)
- unlabeled images (3)
- unlearning (6)
- unlearning algorithms (3)
- unlearning effectiveness (5)
- unlearning methods (5)
- unmanned aerial vehicles (4)
- unobserved confounding (3)
- unpaired data (4)
- unseen domains (4)
- unseen tasks (3)
- unstable optimization (4)
- unsupervised anomaly detection (3)
- unsupervised domain adaptation (9)
- unsupervised learning (26)
- unsupervised motion tasks (1)
- update-to-data ratio (3)
- upper and lower bounds (4)
- upper bound (12)
- upper bounds (15)
- upper confidence bound (7)
- user engagement (4)
- user intent (3)
- user preferences (5)
- user studies (4)
- user study (6)
- user-specific preferences (3)
- utility (4)
- utility function (8)
- utility functions (4)
- utility guarantees (3)
- utility maximization (3)
- validation loss (6)
- validity (3)
- value function (11)
- value function approximation (3)
- value functions (6)
- vanishing gradients (3)
- variable selection (4)
- variance (4)
- variance reduction (9)
- variational autoencoder (13)
- variational autoencoders (4)
- variational framework (4)
- variational inference (17)
- variational lower bound (5)
- vc dimension (5)
- vector databases (3)
- vector quantization (5)
- velocity field (5)
- velocity fields (3)
- verifiable rewards (15)
- versatility (3)
- vertical federated learning (4)
- video anomaly detection (6)
- video comprehension (4)
- video data (3)
- video diffusion models (10)
- video editing (3)
- video generation (25)
- video generation models (4)
- video generative models (4)
- video inpainting (3)
- video large language models (10)
- video object segmentation (3)
- video quality (3)
- video question answering (7)
- video reasoning (4)
- video synthesis (4)
- video temporal grounding (6)
- video understanding (15)
- video-text retrieval (3)
- view consistency (3)
- view synthesis (4)
- viewpoint variations (3)
- virtual reality (4)
- virtual try-on (4)
- vision benchmarks (5)
- vision encoder (5)
- vision encoders (7)
- vision foundation models (8)
- vision language model (5)
- vision language models (16)
- vision tasks (7)
- vision transformer (7)
- vision transformers (37)
- vision-and-language navigation (3)
- vision-language alignment (3)
- vision-language benchmarks (4)
- vision-language model (20)
- vision-language models (173)
- vision-language navigation (4)
- vision-language reasoning (5)
- vision-language tasks (5)
- vision-language understanding (4)
- vision-language-action (16)
- vision-language-action models (7)
- visual artifacts (6)
- visual attention (3)
- visual captioning (3)
- visual complexity (6)
- visual concepts (3)
- visual consistency (6)
- visual cortex (4)
- visual cues (3)
- visual data (3)
- visual details (3)
- visual effects (3)
- visual embeddings (3)
- visual experts (1)
- visual features (6)
- visual fidelity (21)
- visual foundation models (3)
- visual generation (6)
- visual grounding (10)
- visual hallucinations (6)
- visual information processing (3)
- visual input (3)
- visual inputs (4)
- visual language models (5)
- visual modification (1)
- visual observations (3)
- visual perception (10)
- visual perception tasks (3)
- visual perturbations (4)
- visual priors (4)
- visual quality (21)
- visual question answering (18)
- visual question-answering (5)
- visual realism (3)
- visual reasoning (18)
- visual recognition (4)
- visual representation learning (3)
- visual representations (11)
- visual search (2)
- visual similarity (3)
- visual storytelling (3)
- visual tasks (4)
- visual token compression (3)
- visual token pruning (3)
- visual tokenizer (3)
- visual tokens (16)
- visual understanding (10)
- visualization (3)
- visualizations (3)
- visuomotor policies (5)
- vits (3)
- vllm (3)
- vlms (5)
- vocabulary size (4)
- vulnerabilities (8)
- vulnerability detection (3)
- wasserstein distance (8)
- wasserstein space (5)
- watermark detection (3)
- watermarking (8)
- watermarking schemes (3)
- waymo open dataset (3)
- weak supervision (8)
- weak-to-strong generalization (5)
- weakly supervised learning (7)
- weakly-supervised learning (3)
- weather forecasting (3)
- web agents (3)
- weight decay (12)
- weight matrices (11)
- weight matrix (3)
- weight updates (3)
- weighted average (5)
- whole slide image (3)
- whole-body control (3)
- word error rate (3)
- working memory (3)
- world model (9)
- world modeling (3)
- world models (26)
- world simulation (3)
- xlstm (6)
- zero-shot (5)
- zero-shot accuracy (4)
- zero-shot capabilities (6)
- zero-shot classification (13)
- zero-shot detection (4)
- zero-shot forecasting (4)
- zero-shot generalization (34)
- zero-shot inference (10)
- zero-shot learning (28)
- zero-shot methods (3)
- zero-shot performance (11)
- zero-shot prediction (3)
- zero-shot setting (4)
- zero-shot settings (7)
- zero-shot transfer (4)
- zero-sum games (6)
- zeroth-order methods (5)
- zeroth-order optimization (6)