Our paper was accepted for NeurIPS 2022. [1本目] ◼︎書誌情報 Shohei Taniguchi, Yusuke Iwasawa, Wataru Kumagai, Yutaka Matsuo. “Langevin Autoencoders for Learning Deep Latent Variable Models”, Neural Information Processing Systems (NeurIPS 2022). ◼︎概要 Markov chain Monte Carlo (MCMC), such as Langevin dynamics, is valid for approximating intractable distributions. However, its usage is limited in the context…
当研究室の発表が、第25回画像の認識・理解シンポジウム (MIRU2022) にて「MIRU優秀賞」を受賞しました。
タイトル：Pixel vs. Object: 変化キャプショニングにおける最適な画像表現についての研究
著者： 土居健人，濱口竜平，岩澤有祐，大西正輝，松尾豊，櫻田健
https://sites.google.com/view/miru2022/home/award#h.8ea1ax6uvq4w
当研究室の論文がACL2022のWorkshop（Insights from Negative Results in NLP）に採録されました。 ■書誌情報 Itsuki Okimura, Machel Reid, Makoto Kawano and Yutaka Matsuo, On the Impact of Data Augmentation on Downstream Performance in Natural Language Processing, the Third Workshop on Insights from Negative Results in NLP, ACL 2022, May 2022 ■概要 With in the broader scope of machine learning, data augmentation is…
当研究室沖村さん，マシェルさん，河野さんの論文が現在開催されているACLのWorkshopでBest Paper Awardを受賞しました。
◼︎書誌情報 タイトル：Fixing the train-test objective discrepancy: Iterative Image Inpainting for Unsupervised Anomaly Detection 著 者：Hitoshi Nakanishi, Masahiro Suzuki, Yutaka Matuo 掲載号：J-Stage in August Vol.30 ◼︎概要 Autoencoders have emerged as popular methods for unsupervised anomaly detection, but they still have difficulty detecting local anomalies in real-world images due to lack of modeling small details. We have assessed…
Our paper was accepted for IJCAI-ECAI 2022 (Short). ■書誌情報 Takeshi Kojima, Yusuke Iwasawa, and Yutaka Matsuo. “Robustifying Vision Transformer without Retraining from Scratch by Test-Time Class-Conditional Feature Alignment”, the 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence (IJCAI-ECAI 2022) ■概要 Vision Transformer (ViT) is becoming more popular in…
Our paper was accepted for NAACL 2022 (main). ◼︎書誌情報 David Ifeoluwa Adelani, Jesujoba Oluwadara Alabi, Angela Fan, Julia Kreutzer, Xiaoyu Shen, Machel Reid, Dana Ruiter, Dietrich Klakow, Peter Nabende, Ernie Chang, Tajuddeen Gwadabe, Freshia Sackey, Bonaventure F. P. Dossou, Chris Chinenye Emezue, Colin Leong, Michael Beukman, Shamsuddeen Hassan Muhammad, Guyo Dub Jarso, Oreen Yousuf, Andre…
Our paper was accepted for NAACL 2022 (main). ◼︎書誌情報 Machel Reid and Mikel Artetxe “PARADISE: Exploiting Parallel Data for Multilingual Sequence-to-Sequence Pretraining”. The 2022 Conference of the North American Chapter of the Association for Computational Linguistics (NAACL 2022). July 2022. Association for Computational Linguistics. ◼︎概要 Despite the success of multilingual sequence-to-sequence pretraining, most existing approaches…
Our paper was accepted for presentation at NeurIPS2022 (Spotlight) . ◼︎書誌情報 Hiroki Furuta, Yutaka Matsuo, Shixiang Shane Gu. “Generalized Decision Transformer for Offline Hindsight Information Matching”, International Conference on Learning Representations (ICLR2022). ◼︎概要 How to extract as much learning signal from each trajectory data has been a key problem in reinforcement learning (RL), where sample…
◼Information タイトル：Transformerと自己教師あり学習を用いたシーン解釈手法の提案 著 者： 小林 由弥 鈴木 雅大 松尾 豊 掲載号：第37巻2号 J-STAGE ◼Overview Ability to understand surrounding environment based on its components, namely objects, is one of the most important cognitive ability for intelligent agents. Human beings are able to decompose sensory input, i.e. visual stimulation, into some components based on its meaning or relationships between…
日本語