state-of-the-art baselines
State-of-the-art baselines serve as reference points to evaluate the performance of new models or methods against existing, established benchmarks in specific tasks. They provide context for understanding improvements or advancements.
- AceSearcher: Bootstrapping Reasoning and Search for LLMs via Reinforced Self-Play
- COLA: Towards Efficient Multi-Objective Reinforcement Learning with Conflict Objective Regularization in Latent Space
- CoUn: Empowering Machine Unlearning via Contrastive Learning
- Conformal Prediction for Time-series Forecasting with Change Points
- Constrained Entropic Unlearning: A Primal-Dual Framework for Large Language Models
- Defining and Discovering Hyper-meta-paths for Heterogeneous Hypergraphs
- Eulerian Neural Network Informed by Chemical Transport for Air Quality Forecasting
- FedQS: Optimizing Gradient and Model Aggregation for Semi-Asynchronous Federated Learning
- FlowRefiner: A Robust Traffic Classification Framework against Label Noise
- GD$^2$: Robust Graph Learning under Label Noise via Dual-View Prediction Discrepancy
- Graph Few-Shot Learning via Adaptive Spectrum Experts and Cross-Set Distribution Calibration
- Group-Level Data Selection for Efficient Pretraining
- How Different from the Past? Spatio-Temporal Time Series Forecasting with Self-Supervised Deviation Learning
- LBMKGC: Large Model-Driven Balanced Multimodal Knowledge Graph Completion
- Listwise Preference Diffusion Optimization for User Behavior Trajectories Prediction
- Meta Guidance: Incorporating Inductive Biases into Deep Time Series Imputers
- NeuroPath: Neurobiology-Inspired Path Tracking and Reflection for Semantically Coherent Retrieval
- Non-Markovian Discrete Diffusion with Causal Language Models
- SYMPHONY: Synergistic Multi-agent Planning with Heterogeneous Language Model Assembly
- SaFiRe: Saccade-Fixation Reiteration with Mamba for Referring Image Segmentation
- Synergy over Discrepancy: A Partition-Based Approach to Multi-Domain LLM Fine-Tuning
- TV-Rec: Time-Variant Convolutional Filter for Sequential Recommendation
- Tight High-Probability Bounds for Nonconvex Heavy-Tailed Scenario under Weaker Assumptions
- TimeEmb: A Lightweight Static-Dynamic Disentanglement Framework for Time Series Forecasting
- Towards Unsupervised Open-Set Graph Domain Adaptation via Dual Reprogramming
- TrajMamba: An Efficient and Semantic-rich Vehicle Trajectory Pre-training Model
- Uncertain Knowledge Graph Completion via Semi-Supervised Confidence Distribution Learning