論文輪読（2023年度）

Copied from: Public/Paper Reading 2022

このセミナーについて (About this seminar)

担当者が見つけた面白い研究を紹介するセミナーです．以下の内容で構成されます． In this seminar, presenters share interesting studies that they found. The seminar consists of the following elements.

  1. 論文紹介 (Paper presentation): 面白いと思う論文（基本的にLong Paper）を選んで，その内容を著者の代わりになったつもりで発表する． A presenter reads a paper (long paper is strongly preffered) that he or she finds interesting, and explains the content as if he or she was an author of the paper.
  2. 著者プレゼンの上映と解説 (Watching talk videos): 担当者は面白いと思う研究の発表動画を選び，その論文の内容を把握する．担当者の補足説明を聞きながら，参加者みんなで発表動画を鑑賞する． Finding an interesting video presentation, a presenter reads and understands the content of the paper. Attendees watch the presentation video of the author of the paper. The presenter is expected to explain supplementary information about the research.

前者 (1) はプレゼンテーションの練習を兼ねています．後者 (2) は「よい」プレゼンテーションを鑑賞しながら，英語でのプレゼンテーションやディスカッションに慣れることを狙っています． The former (1) aims at practicing presentations. The latter (2) aims at improving listening and discussion skills in English as we enjoy good presentations.

発表者は一人目は (1) を，二人目は (1) か (2) のどちらかを担当します． The 1st presentator take (1). The 2nd one can select (1) or (2).

座長(chair)は、セミナーの司会進行をおこないます。 chair person host the seminar. 座長の仕事を参考にして，円滑に議論が進むように心がけてください． Please contribute to active discussion refering to Chair’s Job.

発表時間 (Presentation Time)

  • Presentation: 15 ~ 20 minnutes
  • QA: 5 ~ 10 minutes

日時 (Date and time)

  • 1Q: 17:30~ (Thu)

参加者 (Attendee)

  • 全員

:exclamation: 発表を登録するときは，著者名，発表年，タイトル，会議名／ジャーナル名，（巻，号，ページ番号など）を必ず記入してください．preprintの論文は発表しないでください。 :exclamation: Please include author names, publication year, title, conference/journal name, (volume and page numbers) when you add an entry for presentation. Do not introduce papers of preprints.

Tips

  • リモートミーティングでビデオプレゼンテーションを行う場合は，画面およびオーディオを共有してください．
  • When you stream the video presentation via remote meeting, please share the both screen and system audio.
    • For Zoom: You don’t have to do nothing special. Just share your screen.
    • For Microsoft Teams: Please activate Share the system audio feature.

今後の予定 (Planned Seminars)

2024-02/19 (Mon) 15:25-

  • seminar
    • loem: Are Human-generated Demonstrations Necessary for In-context Learning?
      • Rui Li, Guoyin Wang, Jiwei Li (ICLR 2024)
      • [Slides] [Paper]
  • chair
    • ishikawa

      2024-02/26 (Mon) 15:25-

      Due to entrance exam.

      2024-03/04 (Mon) 15:25-

  • seminar
  • chair
    • hattori

      Past seminars

      2023-04-20 (Thu) 17:30-

  • presenter
  • chair
    • ishikawa

2023-04-27 (Thu) 17:30-

  • presenter
    • endo: Hallucinated but Factual! Inspecting the Factuality of Hallucinations in Abstractive Summarization (Cao et al., ACL2022)
    • maeda: Transparent Human Evaluation for Image Captioning (Kasai et al., NAACL2022)
  • chair
    • koike hirai

2023-05/04 (Thu) 17:30-

2023-05/11 (Thu) 17:30-

2023-05/18 (Thu) 17:30-

2023-05/25 (Thu) 17:30-

  • seminar
    • koike
      • Adding Instructions during Pretraining: Effective Way of Controlling Toxicity in Language Models(EACL2023)
      • Paper, Slide
    • loem: PCC: Paraphrasing with Bottom-k Sampling and Cyclic Learning for Curriculum Data Augmentation
      • Hongyuan Lu,  Wai Lam; EACL2023
      • paper slide
  • chair
    • panatchakorn

2023-06/01 (Thu) 17:30-

  • seminar
    • ma: Not All Tasks Are Born Equal: Understanding Zero-Shot Generalization (ICLR 2023)
      • Jing Zhou, Zongyu Lin, Yanan Zheng, Jian Li, Zhilin Yang
      • paper, slides
    • nguyen: LISA: Learning Interpretable Skill Abstractions from Language
      • Divyansh Garg, Skanda Vaidyanath, Kuno Kim, Jiaming Song, and Stefano Ermon (NeurIPS2022)
      • paper, slides (best view in Google Slides)
  • chair
    • kato

2023-06/08 (Thu) 17:30-

  • seminar
    • kaneko
    • sangwhan
  • chair
    • wata

2023-06/12 (Mon) 17:30-

  • seminar
    • yang; BIGVIDEO: A Large-scale Video Subtitle Translation Dataset for Multimodal Machine Translation
    • paper, slides
    • iida Ho et al. 2023 Wiki-Why ANSWERING AND EXPLAINING CAUSE-AND-EFFECT QUESTIONS　ICLR2023 - paper, slide
  • chair
    • hirai

2023-06/19 (Mon) 17:30-

2023-06/26 (Mon) 17:30-

  • seminar
    • koyama
      • Renliang Sun, Zhixian Yang, Xiaojun Wan. Exploiting Summarization Data to Help Text Simplification. In EACL 2023.
      • paper, slide
    • mizuki-sakae
      • AKYÜREK, Ekin, et al. What learning algorithm is in-context learning? investigations with linear models. In: ICLR 2023. 2023.
      • paper, video, summary
  • chair
    • hattori

2023-07/03 (Mon) 17:30-

  • seminar
    • vijay - Song, Kaitao et al. “MPNet: Masked and Permuted Pre-training for Language Understanding.” In NeurIPS 2020 - Paper, Slides
    • cognetta - Tokenization and the Noiseless Channel (ACL 2023): preprint (ACL Anthology link not yet available) - Slides
  • chair
    • panatchakorn

2023-07/10 (Mon) 17:30-

  • seminar
    • yoshikawa: Why do Nearest Neighbor Language Models Work?
      • Frank F. Xu, Uri Alon, Graham Neubig (ICML 2023)
      • paper, slides
    • ao:
      • Selection-Inference: Exploiting Large Language Models for Interpretable Logical Reasoning
      • Antonia Creswell, Murray Shanahan, Irina Higgins (ICLR 2023)
      • paper, slides
  • chair
    • hirai

2023-07/17 (Mon) 17:30-

2023-07/24 (Mon) 17:30-

  • seminar
    • sangwhan
    • mizuki-kentaro - Retentive Network: A Successor to Transformer for Large Language Models - paper, slides
  • chair
    • maeda

2023-07/31 (Mon) 17:30-

### 2023-09/25 (Mon) 17:30-

2023-10/02 (Mon) 15:25-

  • seminar
    • wata: Chain-of-Verification Reduces Hallucination in Large Language Models
    • muraoka: When and Why Vision-Language Models Behave like Bags-Of-Words, and What to Do About It? (ICLR2023)
  • chair
    • kato

      2023-10/09 (Mon) 15:25-

  • Sport’s day ( National Holiday )

### 2023-10/16 (Mon) 15:25-

2023-10/23 (Mon) 15:25-

2023-11/06 (Mon) 15:25-

2023-12/11 (Mon) 15:25-

  • seminar
    • yoshikawa
      • DisentQA: Disentangling Parametric and Contextual Knowledge with Counterfactual Question Answering (ACL2023)
        • Ella Neeman, Roee Aharoni, Or Honovich, Leshem Choshen, Idan Szpektor, Omri Abend
      • paper, slides
    • erick: - Paper: Toolformer: Language Models Can Teach Themselves to Use Tools. - Authors: Timo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Luke Zettlemoyer, Nicola Cancedda, Thomas Scialom - Conference: Oral presentation at Neurips 2023. - Video: Poster video
  • chair
    • endo

      2023-12/18 (Mon) 15:25-

  • seminar
    • sangwhan - Text Embeddings Reveal (Almost) As Much As Text - John X. Morris, Volodymyr Kuleshov, Vitaly Shmatikov, Alexander M. Rush - EMNLP 2023 - paper slides
    • wang
  • chair
    • hirai

      2023-12/25 (Mon) 15:25-

      Talk by Roberto Navigli.

      2024-01/01 (Mon) 15:25-

      National Holiday

      2024-01/08 (Mon) 15:25-

      National Holiday

      2024-01/15 (Mon) 15:25-

  • seminar
    • hirai
    • marco
  • chair
    • ishikawa

2024-01/22 (Mon) 15:25-