computational costs
The resources (time, memory, processing power) required to train or run an AI model, influencing scalability and feasibility of implementations.
- $\text{S}^2$Q-VDiT: Accurate Quantized Video Diffusion Transformer with Salient Data and Sparse Token Distillation
- A Near-optimal, Scalable and Parallelizable Framework for Stochastic Bandits Robust to Adversarial Corruptions and Beyond
- AutoEdit: Automatic Hyperparameter Tuning for Image Editing
- AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration
- BEAST: Efficient Tokenization of B-Splines Encoded Action Sequences for Imitation Learning
- Beyond Greedy Exits: Improved Early Exit Decisions for Risk Control and Reliability
- Cancer Survival Analysis via Zero-shot Tumor Microenvironment Segmentation on Low-resolution Whole Slide Pathology Images
- Classical Planning with LLM-Generated Heuristics: Challenging the State of the Art with Python Code
- Conditional Representation Learning for Customized Tasks
- Constrained Linear Thompson Sampling
- Cost-aware LLM-based Online Dataset Annotation
- DartQuant: Efficient Rotational Distribution Calibration for LLM Quantization
- DeltaFlow: An Efficient Multi-frame Scene Flow Estimation Method
- DyMU: Dynamic Merging and Virtual Unmerging for Efficient Variable-Length VLMs
- Efficient Prompt Compression with Evaluator Heads for Long-Context Transformer Inference
- Efficiently Maintaining the Multilingual Capacity of MCLIP in Downstream Cross-Modal Retrieval Tasks
- FastLongSpeech: Enhancing Large Speech-Language Models for Efficient Long-Speech Processing
- FedRACE: A Hierarchical and Statistical Framework for Robust Federated Learning
- Flash Invariant Point Attention
- FlowCut: Rethinking Redundancy via Information Flow for Efficient Vision-Language Models
- FlowMoE: A Scalable Pipeline Scheduling Framework for Distributed Mixture-of-Experts Training
- GLID$^2$E: A Gradient-Free Lightweight Fine-tune Approach for Discrete Biological Sequence Design
- Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization
- Hawaii: Hierarchical Visual Knowledge Transfer for Efficient Vision-Language Models
- Improving Diffusion-based Inverse Algorithms under Few-Step Constraint via Linear Extrapolation
- Improving Task-Specific Multimodal Sentiment Analysis with General MLLMs via Prompting
- Language Ranker: A Lightweight Ranking framework for LLM Decoding
- Momentum-SAM: Sharpness Aware Minimization without Computational Overhead
- Multiplication-Free Parallelizable Spiking Neurons with Efficient Spatio-Temporal Dynamics
- NEP: Autoregressive Image Editing via Next Editing Token Prediction
- Neural Networks for Learnable and Scalable Influence Estimation of Instruction Fine-Tuning Data
- Opinion Maximization in Social Networks by Modifying Internal Opinions
- Optimized Minimal 3D Gaussian Splatting
- PAROAttention: Pattern-Aware ReOrdering for Efficient Sparse and Quantized Attention in Visual Generation Models
- Planning without Search: Refining Frontier LLMs with Offline Goal-Conditioned RL
- Pragmatic Heterogeneous Collaborative Perception via Generative Communication Mechanism
- Quartet: Native FP4 Training Can Be Optimal for Large Language Models
- SMRS: advocating a unified reporting standard for surrogate models in the artificial intelligence era.
- Sequential Attention-based Sampling for Histopathological Analysis
- Single-Teacher View Augmentation: Boosting Knowledge Distillation via Angular Diversity
- Sparse Gaussian Processes: Structured Approximations and Power-EP Revisited
- Split conformal classification with unsupervised calibration
- Time-Masked Transformers with Lightweight Test-Time Adaptation for Neural Speech Decoding
- Toward Efficient Inference Attacks: Shadow Model Sharing via Mixture-of-Experts
- Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator
- VCM: Vision Concept Modeling with Adaptive Vision Token Compression via Instruction Fine-Tuning
- Visual Instruction Bottleneck Tuning
- zip2zip: Inference-Time Adaptive Tokenization via Online Compression