6 papers
OrderGrad: Optimizing Beyond the Mean with Order-Statistic Policy Gradient Estimation
Paavo Parmas, Yongmin Kim, Kohsei Matsutani +5
Policy-gradient methods usually optimize expected return, but many real world applications care about distributional properties of returns: tail risk, outlier robustness, or best-o…
On Advantage Estimates for Max@K Policy Gradients
Shota Takashiro, Soichiro Nishimori, Paavo Parmas +6
Reinforcement learning with verifiable rewards is widely used for post-training reasoning models, but sparse outcome rewards make exploration difficult. A complementary approach is…
Zipping the Thought: When and How Compressed Reasoning Data Works in LLM Post-Training
Kohsei Matsutani, Gouki Minegishi, Takeshi Kojima +2
Large language models (LLMs) can now solve complex problems through long chain-of-thought (CoT) reasoning, but the trade-off between performance and token cost remains a central ch…
RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMs
Kohsei Matsutani, Shota Takashiro, Gouki Minegishi +3
Large language models (LLMs) are typically trained by reinforcement learning (RL) with verifiable rewards (RLVR) and supervised fine-tuning (SFT) on reasoning traces to improve the…
MultiBanana: A Challenging Benchmark for Multi-Reference Text-to-Image Generation
Yuta Oshima, Daiki Miyake, Kohsei Matsutani +4
Recent text-to-image generation models have acquired the ability of multi-reference generation and editing; that is, to inherit the appearance of subjects from multiple reference i…
Mixture of Experts Provably Detect and Learn the Latent Cluster Structure in Gradient-Based Learning
Ryotaro Kawata, Kohsei Matsutani, Yuri Kinoshita +2
Mixture of Experts (MoE), an ensemble of specialized models equipped with a router that dynamically distributes each input to appropriate experts, has achieved successful results i…