論文輪読（2024年度）

Copied from: Public/Paper Reading 2023

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)

  • 3Q: 13:30~ (Tues.)
  • 4Q: 15:25~ (Mon.)

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)

2024-12/16 (Mon) 15:25-17:05

  • seminar
    • shimada
      • Paper
      • Slide
    • shiotani
      • Can LLMs Recognize Toxicity? A Structured Investigation Framework and Toxicity Metric paper slide
  • chair
    • takahashi

2024-12/23 (Mon) 15:25-17:05

  • seminar
    • maeda: ReCEval: Evaluating Reasoning Chains via Correctness and Informativeness (EMNLP2023) paper slides(pdf) slides(html)
    • koyama: Lookback Lens: Detecting and Mitigating Contextual Hallucinations in Large Language Models Using Only Attention Maps (EMNLP 2024) paper slides
  • chair
    • hattori

~### 2025-1/6 (Mon) 15:25-17:05~ 言語処理学会直前の為

### 2025-1/13 (Mon) 15:25-17:05 成人式の為

2025-1/20 (Mon) 15:25-17:05

  • seminar
    • david
    • hida
  • chair
    • panatchakorn

2024-01/27 (Mon) 15:25-17:05

  • seminar
    • yang
  • chair
    • maruyama

Past Seminars

2024-04/09(Tue) 17:15-18:55

  • seminar
    • wang
      • Mamba: Linear-Time Sequence Modeling with Selective State Spaces paper slide
    • hida
      • Exploring Cross-Cultural Differences in English Hate Speech Annotations: From Dataset Construction to Analysis(NAACL2024) paper slide )
        • Nayeon Lee, Chani Jung, Junho Myung, Jiho Jin, Jose Camacho-Collados, Juho Kim, Alice Oh
  • chair
    • endo maruyama

      2024-04/16(Tue) 17:15-18:55

  • seminar
    • koike
      • DNA-GPT: DIVERGENT N-GRAM ANALYSIS FOR TRAINING-FREE DETECTION OF GPT-GENERATED TEXT
        • Xianjun Yang, Wei Cheng, Yue Wu, Linda Ruth Petzold, William Yang Wang, Haifeng Chen; ICLR2024
        • Paper Slide
    • maeda
      • LLMScore: Unveiling the Power of Large Language Models in Text-to-Image Synthesis Evaluation
        • Yujie Lu et al., NeurIPS 2023 Main Conf.
        • Paper Slide
  • chair
    • hattori

      2024-04/23(Tue) 17:15-18:55

  • seminar
    • vijay
      • Semantic Sensitivities and Inconsistent Predictions Measuring the Fragility of NLI Models
    • ma
      • Prompting is not a substitute for probability measurements in large language models (EMNLP 2023)
      • Jennifer Hu, Roger Levy
      • video, paper, author slides
  • chair
    • mizuki-kentaro

      2024-04/30(Tue) 17:15-18:55

      ゴールデンウィーク中日

      2024-05/07(Tue) 17:15-18:55

  • potential research topic
  • chair
    • david

2024-05/14(Tue) 17:15-18:55

  • seminar
    • panatchakorn
      • We’re Afraid Language Models Aren’t Modeling Ambiguity (EMNLP 2023)
      • Alisa Liu, Zhaofeng Wu, Julian Michael, Alane Suhr, Peter West, Alexander Koller, Swabha Swayamdipta, Noah Smith, Yejin Choi
      • paper, slide
    • maruyama
      • Tree of Thoughts: Deliberate Problem Solving with Large Language Models (NeurIPS 2023)
      • Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Thomas L. Griffiths, Yuan Cao, Karthik R Narasimhan
      • paper, video
  • chair
    • shimada　shiotani

      2024-05/21(Tue) 17:15-18:55

  • seminar
    • sangwhan
      • Fishing for Magikarp: Automatically Detecting Under-trained Tokens in Large Language Models (arXiv preprint)
      • Sander Land, Max Bartolo
      • paper, slides
    • yang
      • Bridging the Gap between Synthetic and Authentic Images for Multimodal Machine Translation (EMNLP 2023)
      • Wenyu Guo, Qingkai Fang, Dong Yu, Yang Feng
      • paper , Slides
  • chair
    • david

      2024-05/28(Tue) 17:15-18:55

  • seminar
    • hattori
      • Rho-1: Not All Tokens Are What You Need (arXiv preprint)
      • Paper, Slides
    • endo
  • chair
    • takahashi

      2024-06/04(Tue) 17:15-18:55

  • seminar
    • yoshikawa
      • On Large Language Models’ Hallucination with Regard to Known Facts (NAACL 2024)
      • paper, slides
    • erick
  • chair
    • endo

      2024-06/11(Tue) 17:15-18:55

  • seminar
    • koyama
      • Exploring Human-Like Translation Strategy with Large Language Models (TACL 2024)
      • Zhiwei He, Tian Liang, Wenxiang Jiao, Zhuosheng Zhang, Yujiu Yang, Rui Wang, Zhaopeng Tu, Shuming Shi, Xing Wang
      • paper, slide
    • cognetta
  • chair
    • panatchakorn

      2024-06/18(Tue) 17:15-18:55

  • seminar
  • chair
    • shiotani

      2024-06/25(Tue) 17:15-18:55

  • seminar
  • chair
    • maruyama

      2024-07/02(Tue) 17:15-18:55

  • seminar
    • saito
      • Red Teaming Language Model Detectors with Language Models (TACL 2024)
      • Zhouxing Shi, Yihan Wang, Fan Yin, Xiangning Chen, Kai-Wei Chang, Cho-Jui Hsieh
      • Paper, Slides
    • takahashi
      • Beyond Memorization: Violating Privacy via Inference with Large Language Models (ICLR 2024)
      • Robin Staab · Mark Vero · Mislav Balunovic · Martin Vechev
      • paper, video
  • chair
    • mizuki-kentaro

      2024-07/09(Tue) 17:15-18:55

      absence of presenters

      2024-07/16(Tue) 17:15-18:55

      Shifted to following week ### 2024-07/23(Tue) 17:15-18:55

  • seminar
  • chair
    • shimada ### 2024-10/22 (Tue) 13:30-15:10
  • seminar
    • panatchakorn
      • Confidence Under the Hood: An Investigation into the Confidence-Probability Alignment in Large Language Models (ACL’24)
      • paper, slide
    • maruyama
      • Grokked Transformers are Implicit Reasoners: A Mechanistic Journey to the Edge of Generalization
      • paper, slide
  • chair
    • endo

2024-10/29 (Tue) 13:30-15:10

  • seminar
    • hattori
      • Generative Verifiers: Reward Modeling as Next-Token Prediction (arXiv)
      • Paper, Slides
    • takahashi
      • Do Membership Inference Attacks Work on Large Language Models? (COLM 2024)
      • Paper, Slides
  • chair
    • panatchakorn

### 2024-11/5 (Tue) 13:30-15:10 大学祭のため

2024-11/12 (Tue) 13:30-15:10

2024-11/19 (Tue) 13:30-15:10

2024-11/26 (Tue) 13:30-15:10

  • seminar
    • LUO (arase lab)
      • Paper https://aclanthology.org/2024.acl-long.715.pdf
      • Slides paper_reading_luo_241126.pdf (1.1 MB)
    • endo
      • Paper https://aclanthology.org/2024.naacl-long.95.pdf
      • Slides Ghostbuster_ Detecting Text Ghostwritten by Large Language Models.pdf (580.0 kB)
  • chair
    • hattori ~### 2024-12/3 (Tue) 13:30-15:10~ absense of presenters

2024-12/9 (Mon) 15:25-17:05

  • seminar
    • sangwhan
      • Tokenization is More Than Compression - Schmidt et al., 2024 (EMNLP 2024)
      • Paper: anthology
      • Slides: master, pdf
    • marco
      • CUTE: Measuring LLMs’ Understanding of Their Tokens (Edman et al., EMNLP 2024)
      • Paper
      • Slides
  • chair
    • david