empirical analysis
The study and evaluation of AI models and systems based on experimental data and observations rather than solely on theoretical foundations.
- A Closer Look to Positive-Unlabeled Learning from Fine-grained Perspectives: An Empirical Study
- Accelerating Parallel Diffusion Model Serving with Residual Compression
- Ada-R1: Hybrid-CoT via Bi-Level Adaptive Reasoning Optimization
- DataSIR: A Benchmark Dataset for Sensitive Information Recognition
- Demystifying Language Model Forgetting with Low-rank Example Associations
- Don’t Think Longer, Think Wisely: Optimizing Thinking Dynamics for Large Reasoning Models
- Fast-Slow Thinking GRPO for Large Vision-Language Model Reasoning
- Generalization Bounds for Model-based Algorithm Configuration
- Homogeneous Keys, Heterogeneous Values: Exploiting Local KV Cache Asymmetry for Long-Context LLMs
- Horizon Reduction Makes RL Scalable
- How Does Topology Bias Distort Message Passing in Graph Recommender? A Dirichlet Energy Perspective
- How to Scale Second-Order Optimization
- IGD: Token Decisiveness Modeling via Information Gain in LLMs for Personalized Recommendation
- Impact of Dataset Properties on Membership Inference Vulnerability of Deep Transfer Learning
- Mixture of Scope Experts at Test: Generalizing Deeper Graph Neural Networks with Shallow Variants
- On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity
- On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity
- One SPACE to Rule Them All: Jointly Mitigating Factuality and Faithfulness Hallucinations in LLMs
- Pessimistic Data Integration for Policy Evaluation
- ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models
- Rethinking Neural Combinatorial Optimization for Vehicle Routing Problems with Different Constraint Tightness Degrees
- Rethinking Residual Distribution in Locate-then-Edit Model Editing
- Scalable Fingerprinting of Large Language Models
- The Curse of Depth in Large Language Models
- Theoretical Guarantees for the Retention of Strict Nash Equilibria by Coevolutionary Algorithms
- Token Embeddings Violate the Manifold Hypothesis
- Unveiling Chain of Step Reasoning for Vision-Language Models with Fine-grained Rewards