Profiles2020-2026
Yutaka Matsuo
The University of Tokyo
18Papers
16Last author
26Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers18
- Does "Do Differentiable Simulators Give Better Policy Gradients?" Give Better Policy Gradients?Author 4 of 4ICLR 2026 · 134 citations
- Quantization-Aware Diffusion Models For Maximum Likelihood TrainingAuthor 3 of 3ICLR 2026
- Topology of Reasoning: Understanding Large Reasoning Models through Reasoning Graph PropertiesAuthor 5 of 5NeurIPS 2025 · 31 citations
- Answer When Needed, Forget When Not: Language Models Pretend to Forget via In-Context Knowledge UnlearningAuthor 6 of 6ACL Findings 2025 · 2 citations
- Dynamic Injection of Entity Knowledge into Dense RetrieversAuthor 5 of 8EMNLP Findings 2025 · 1 citation
- When Instructions Multiply: Measuring and Estimating LLM Capabilities of Multiple Instructions FollowingAuthor 7 of 7EMNLP Findings 2025 · 1 citation
- Beyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit EmergenceAuthor 5 of 5ICML 2025
- Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous WordsAuthor 4 of 4ICLR 2025
- ADOPT: Modified Adam Can Converge with Any β2 with the Optimal RateAuthor 10 of 10NeurIPS 2024 · 32 citations
- Improving Low-Resource Machine Translation for Formosan Languages Using Bilingual Lexical ResourcesAuthor 3 of 3ACL Findings 2024 · 5 citations
- Langevin Autoencoders for Learning Deep Latent Variable ModelsAuthor 4 of 4NeurIPS 2022 · 2 citations
- Test-Time Classifier Adjustment Module for Model-Agnostic Domain GeneralizationAuthor 2 of 2NeurIPS 2021 · 456 citations
- Deployment-Efficient Reinforcement Learning via Model-Based Offline OptimizationAuthor 3 of 5ICLR 2021 · 166 citations
- Subformer: Exploring Weight Sharing for Parameter Efficiency in Generative TransformersAuthor 3 of 3EMNLP Findings 2021 · 27 citations
- Group Equivariant Conditional Neural ProcessesAuthor 5 of 5ICLR 2021 · 22 citations
- AfroMT: Pretraining Strategies and Reproducible Benchmarks for Translation of 8 African LanguagesAuthor 4 of 4EMNLP 2021 · 12 citations
- Variational Inference for Learning Representations of Natural Language EditsAuthor 3 of 3AAAI 2021 · 9 citations
- VCDM: Leveraging Variational Bi-encoding and Deep Contextualized Word Representations for Improved Definition ModelingAuthor 3 of 3EMNLP 2020 · 14 citations
