Full list on Google Scholar.
Under review
MIPS: Mitigating Mode Collapse in Reinforcement Learning for Large Outcome Spaces
Under review at ICLR 2027
Extends reward-proportional RL to massive discrete outcome spaces using trajectory-based marginal-probability estimation.
Density-Aware Reward Scaling: A Reinforcement Learning Framework for Reward-Proportional Sampling
Under review at AAAI 2027
DARS is a critic-free reward-processing method that inversely scales rewards by local outcome density while keeping the standard policy-gradient objective.
2026
Expected Return Causes Outcome-Level Mode Collapse in Reinforcement Learning and How to Fix It with Inverse Probability Scaling
International Conference on Machine Learning, 2026
Explains why expected-return objectives concentrate policies on a narrow set of outcomes, and introduces inverse probability scaling to recover diverse high-reward outcomes.
How Does My Model Fail? Automatic Identification and Interpretation of Physical Plausibility Failure Modes with Matryoshka Transcoders
CVPR Workshop on Explainable AI for Computer Vision (XAI4CV), 2026 Spotlight
Matryoshka Transcoders learn hierarchical sparse features for coarse-to-fine analysis of physical-plausibility failures in generative models.
2023
Brain-Inspired Attention Model for Object Counting
International Conference on Neural Information Processing, 2023
A Q-learning-based sequential attention model with recurrent visual glimpses that reaches 92.1% object-counting accuracy.