9 papers · 1 filter
Bayes-Optimal BER and AUC: Estimation and Evaluation of Estimators
Ryota Ushio, Takashi Ishida, Masashi Sugiyama
A fundamental quantity in machine learning is the optimal performance achievable by any model on a given task. Estimating this quantity allows us to distinguish the irreducible par…
Do Coding Agents Deceive Us? Detecting and Preventing Cheating via Capped Evaluation with Randomized Tests
Thanawat Lodkaew, Johannes Ackermann, Soichiro Nishimori +3
A growing failure mode in agent evaluation and training is that models can achieve high evaluation scores by exploiting shortcuts instead of solving the intended task, producing de…
Proteo-R1: Reasoning Foundation Models for De Novo Protein Design
Fang Wu, Weihao Xuan, Heli Qi +26
Deep learning in de novo protein design has achieved atomic-level fidelity. However, existing models remain largely non-deliberative: they directly synthesize molecular geometries…
Mitigating Reward Hacking in RLHF via Advantage Sign Robustness
Shinnosuke Ono, Johannes Ackermann, Soichiro Nishimori +2
Reward models (RMs) used in reinforcement learning from human feedback (RLHF) are vulnerable to reward hacking: as the policy maximizes a learned proxy reward, true quality plateau…
Gradient Regularization Mitigates Reward Hacking in Reinforcement Learning from Human Feedback and Verifiable Rewards
Johannes Ackermann, Michael Noukhovitch, Takashi Ishida +1
Reinforcement Learning from Human Feedback (RLHF) or Verifiable Rewards (RLVR) are two key steps in the post-training of modern Language Models (LMs). A common problem is reward ha…
Towards Scalable Oversight via Partitioned Human Supervision
Ren Yin, Takashi Ishida, Masashi Sugiyama
As artificial intelligence (AI) systems approach and surpass expert human performance across a broad range of tasks, obtaining high-quality human supervision for evaluation and tra…
LLM Routing with Dueling Feedback
Chao-Kai Chiang, Takashi Ishida, Masashi Sugiyama
We study LLM routing, the problem of selecting the best model for each query while balancing user satisfaction, model expertise, and inference cost. We formulate routing as context…
Off-Policy Corrected Reward Modeling for Reinforcement Learning from Human Feedback
Johannes Ackermann, Takashi Ishida, Masashi Sugiyama
Reinforcement Learning from Human Feedback (RLHF) allows us to train models, such as language models (LMs), to follow complex human preferences. In RLHF for LMs, we first train an…
Practical estimation of the optimal classification error with soft labels and calibration
Ryota Ushio, Takashi Ishida, Masashi Sugiyama
While the performance of machine learning systems has experienced significant improvement in recent years, relatively little attention has been paid to the fundamental question: to…