distribution shift
Distribution shift occurs when the statistical properties of the training data do not match those of the test data. This can lead to degraded model performance and is an important consideration in deploying models in dynamic environments.
- A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspective
- ADPretrain: Advancing Industrial Anomaly Detection via Anomaly Representation Pretraining
- Adaptive Neighborhood-Constrained Q Learning for Offline Reinforcement Learning
- Bridging Critical Gaps in Convergent Learning: How Representational Alignment Evolves Across Layers, Training, and Distribution Shifts
- Conformal Prediction in The Loop: A Feedback-Based Uncertainty Model for Trajectory Optimization
- DreamPRM: Domain-reweighted Process Reward Model for Multimodal Reasoning
- Epistemic Uncertainty for Generated Image Detection
- Escaping the SpuriVerse: Can Large Vision-Language Models Generalize Beyond Seen Spurious Correlations?
- Generalization vs Specialization under Concept Shift
- Incentivizing Reasoning for Advanced Instruction-Following of Large Language Models
- Increasing the Utility of Synthetic Images through Chamfer Guidance
- Monitoring Risks in Test-Time Adaptation
- NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache
- Optimal Single-Policy Sample Complexity and Transient Coverage for Average-Reward Offline RL
- PARTONOMY: Large Multimodal Models with Part-Level Visual Understanding
- PolyJuice Makes It Real: Black-Box, Universal Red Teaming for Synthetic Image Detectors
- Quantifying Uncertainty in the Presence of Distribution Shifts
- Robust LLM Alignment via Distributionally Robust Direct Preference Optimization
- Secure and Confidential Certificates of Online Fairness
- TOMCAT: Test-time Comprehensive Knowledge Accumulation for Compositional Zero-Shot Learning
- The Power of Iterative Filtering for Supervised Learning with (Heavy) Contamination
- Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness
- What Does It Take to Build a Performant Selective Classifier?