I've graduated from NAIST with my master's degree! 🎓
Huge thanks to Prof. Taro Watanabe and everyone in the NLP Lab for all the guidance and support.
Same lab, new chapter: I'm staying on as a PhD student starting this October.
Our paper on humor preference analysis has been accepted to #EMNLP2026 main conference! Huge thanks to all my co-authors for their amazing support.
The final draft will be on arXiv soon. See you in Budapest! 🇭🇺✈️
https://t.co/PuYgB7pWeY
7 papers were accepted at EMNLP2026🎉🎉
Also, one paper was accepted to the Industry track🎉
EMNLPに7本採択されました!また、Industry trackでも1本採択されました。論文の詳細は主著の方々からそれぞれアナウンスあると思うので、お楽しみに。
#EMNLP2026#EMNLP#NLProc
Our work has been selected as an ACL SAC Highlight at #ACL2026 🎉🎉
Huge thanks to the SAC and reviewers for recognizing our work and for their thoughtful feedback!!
ACL 2026でSAC Highlightに選出されました。2年連続で主著論文が受賞となり、とても嬉しいです!
#ACL2026NLP#NLProc
"RePo: Language Models with Context Re-Positioning" will be presented at #ICML2026
Paper: https://t.co/Y13MORpqTB
Standard language models process information as a rigid linear sequence where the only signal for structure is a fixed token index, forcing them to treat physical proximity as semantic relevance. This is inefficient, since when key facts are buried in noise, models waste finite capacity managing disorganized inputs instead of focusing on deep reasoning. RePo breaks this bottleneck by allowing models to actively reorganize their context. Rather than using a fixed index, our module learns to assign positions based on content relevance. This flexibility yields significant gains in robustness.
https://t.co/UGS1FkQxlf
🎉 Happy to share that our paper on a Construction Grammar benchmark received an 👑Outstanding Paper Award👑 at ACL2026!
Huge thanks to @penzant !
#ACL2026
Tomorrow, July 7, at 10:30, we will present the following work done by @yusuke1997, @h_kamiga, and @tarowatanabe at ICML 2026 Poster Session 1 (Hall A): https://t.co/bfd3R9ApwO (1/5)
Tomorrow, July 7, at 10:30, we will present the following work done by @yusuke1997, @h_kamiga, and @tarowatanabe at ICML 2026 Poster Session 1 (Hall A): https://t.co/bfd3R9ApwO (1/5)
By assigning weights to these terms and evaluating them, we found that this decomposition improves natural language generation performance and that each term exhibits distinct tendencies. (4/5)