Research seminar (2022)

毎週メンバー全員が参加し,研究について議論を行うミーティングです.

This is a weekly meeting where all members attend and discuss the research in the laboratory.

  • 日時 (Time slot):
    • 1Q: 12:30~ (Fri.)
    • 2Q: 12:30~ (Fri.)
    • 3Q: 13:45~ (Fri.)
    • 4Q: 17:30~ (Wed.)
  • 発表要領 (Presentation format):
    • (ひとりあたり)発表15 ~ 20分 および 議論10 ~ 15分,合計30分を目安にしてください.
    • 15 - 20 min for a presentation and 10 - 15 min for discussion, total 30 min.
    • Tipsの発表は5分程度(長くても質疑含めて15分くらい)が望ましいです
    • 5 mins for Tips presentation, upto 15 mins including discussion.
  • 言語 (Language used for presentation):
    • トークとスライドのうち,少なくともどちらか一方は英語で用意してください.
      学部生は例外的に,本人が希望すれば,トークとスライドの両方を日本語にすることができます.
    • Please prepare, between talk and slide, at least one of them must be English.
      Undergraduate students are exceptional: both talk and slide can be Japanese if they prefer it.
  • 座長 (Chair)
    • B/Mの学生は持ち回りで座長を担当します.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.
    • B/M students are assigned as a chair of the seminar. Please contribute to active discussion refering to Chair’s Job.

Planned Seminars

2023-02-08 (Wed) 17:30~

  • seminar
    • muraoka: Progresss in Unified Multimodal Model (2.0 MB)

    • sangwhan Gradient-free Inspection of LLMs

  • tips
    • niwa: Outreach Activities
  • chair
    • hirai

2023-02-15 (Wed) 17:30~

  • seminar
    • liu: Table-Logic Pretraining for Tabular Reasoning.pdf (1.6 MB)

    • maeda : QuIC-360°: Query-based Image Captioning for Multi-context 360° Images (Work @ RIKEN AIP)

  • tips
    • yang ChatGPT
  • chair
    • taniguchi

2023-03-03 (Fri) 13:00~

Rehearsal for ANLP 2023

  • oral: 15 mins presentation + 5 mins comments
  • poster: 10 mins presentation + 5 mins comments
Time Presenter Form Slide
13:00 - 13:15 iida poster slide
13:15 - 13:35 loem(スピーチコンテストで欠席ですが、Slack等でコメントをいただければ幸いです。) oral slide
13:35 - 13:55 hattori oral slide
       
14:05 - 14:20 endo poster slide
14:20 - 14:40 nguyen oral slide
14:40 - 15:00 boost oral slide
       
15:10 - 15:25 yang poster Poster.pdf
15:25 - 15:45 kaneko oral slid
15:45 - 16:05 taniguchi(引っ越しのため時間変更の可能性あり) oral Slide
       
16:15 - 16:30 niwa poster poster
16:30 - 16:50 maeda oral slide

waiting list (cancels)

Please be prepared to do your presentation anytime.

  • sangwhan (11/04)
  • liu (11/04)
  • maeda(12/02, maybe I have nothing to talk)

Past seminars

2022-04-15(Fri) 12:30~

  • Self-introduction
    • Marco Marco’s TLDR.pdf

    • Wang Self_Intro.pdf

    • Hirai self-introduction.pdf

    • Trang Nguyen Self-introduction.pdf

    • Panatchakorn : Self-Introduction

    • Endo self-introduction

    • Hattori Self Introduction.pdf

  • chair
    • ishikawa

2022-04-22(Fri) 12:30~

  • Kick off
    • Okazaki: Kick-off: Japanese slides, English slides
  • chiar
    • maeda
  • Agenda:
    • Rearrengement of goods in Room 604
    • Welcome-party
    • Laboratory tour
    • TA reports

2022-05-06(Fri) 12:30~

  • seminar
    • iida: When, Why, and Which methods of Zero-shot Domain Adaptation for Information Retrieval are Effective?
    • mizuki, Semantic Specialization for knowledge-based Word Sense Disambiguation
  • tips
    • niwa: Services and tools for research
  • chair
    • kuo

2022-05-13(Fri) 12:30~

  • seminar
    • kaneko: Interpretability for Language Learners Using Example-Based Grammatical Error Correction, Paper
    • yoshikawa: Selective Prediction for Evaluating Confidence of Knowledge in LanguageModels
  • tips
    • yang OmegaConf.pdf (135.4 kB)
  • chair
    • taniguchi

2022-05-20(Fri) 12:30~

  • seminar
    • maruyama
    • vijay Acceptability Evaluation
  • tips
    • mizuki. Tips for controlling a difficulty of your research project

2022-05-27(Fri) 12:30~

  • seminar
    • kuo Learning Causality and Part Relationships in Car Recall Text

    • sangwhan

  • tips
    • koyama ABCI trivial tips
  • chair
    • maeda

2022-06-03(Fri) 12:30~

  • seminar
    • maeda IMPARA: Impact-based Metric for GEC using Parallel Data

    • ishikawa: Robust Parsing for Sentences with Case Markers Omitted Using Pseudo Training Data

  • tips
    • taniguchi
      • What to do for job hunting
  • chair
    • kuo

2022-06-10(Fri) 12:30~

  • seminar
    • loem : Multi-Head N-gram: Efficient Alternative to Self-Attention

    • niwa: Nearest Neighbor Non-autoregressive Text Generation

  • tips
    • ma
  • chair
    • Ishikawa

2022-06-17(Fri) 12:30~

  • seminar
    • liu: Table-to-Logic Pretraining for Logical Table-to-Text Generation

    • yang 2022_06_17_progress_presentation.pdf (1.5 MB)

  • tips
    • sangwhan - Tips on Trips
  • chair
    • nguyen

2022-06-24(Fri) 12:30~

  • seminar
    • cognetta
    • erick: 6060-Research Seminar
  • tips
    • ma: My Tool for Paperworks
  • chair
    • panatchakorn

2022-07-01(Fri) 12:30~

  • seminar
    • 研究シーズの共有会
      • muraoka: research_seed_muraoka.pdf (499.9 kB)
      • marco: Some Neat (IMO) Research Topics.pdf (186.9 kB)
      • yoshikawa: research_theme_yoshikawa.pdf (488.2 kB)
      • takase: pruning neural networks without training
      • kaneko: https://nlp.esa.io/posts/1958
      • okazaki: okazaki-2022
      • koyama: slide
      • Mizuki: Rare and Zero-shot Word Sense Disambiaguation using Language Model Few-shot Learning
        • No presentation due to the missing deadline

2022-07-08(Fri) 12:30~

  • seminar
    • taniguchi
      • Reverse-GPT-2
    • ma
      • Document-Level Relation Extraction with Human-Machine Evidence Alignment
  • tips
    • kaneko
      • How to come up with a research topic (survey part for input)
      • とりあえず日本語のスライドを貼りました🙏
      • 途中まで英語: slide
  • chair
    • hirai

2022-07-15(Fri) 12:30~

  • seminar
    • muraoka: Towards Multimodal Foundation Model (3.3 MB)
    • takase
  • tips
    • vijay
  • chair
    • endo

2022-07-22(Fri) 12:30~

  • seminar
    • sangwhan: Surrogate Switches for Decoder-side Vocab Patching

    • kuo (master thesis presentation rehearsal): slides

  • chair
    • maruyama

2022-07-29(Fri) 12:30~

  • seminar
    • cognetta: Parameter Efficient Korean Character Level Language Modeling
      • slides
    • wang: Comparing and Combining of Discriminator and Generator for Aspect-based Sentiment Analysis
      • slides
  • tips
    • ishikawa: SSH Config Tips
  • chair
    • loem

2022-09-16(Fri) 12:30\~

  • canceled

2022-09-23(Fri) 12:30\~

  • (秋分の日)

2022-09-30(Fri) 12:30~

  • seminar
    • koyama
    • nguyen TrangNguyen_GeneralizationInVQA.pdf
  • tips
    • vijay TIPS FOR PRODUCTIVITY - DEEP WORK 2022 JULY
  • chair
    • panatchakorn

2022-10-07 (Fri) 13:45~

  • seminar
    • hirai: Improving Vokenization Process by Learning Token Vectors Using Contrastive Learning
      • Slide
      • Poster(※10/4の合同研向けに作成)
    • hattori: Query Generation for Query-Focused Summarization of Long Documents
      • Slide
      • Poster (※10/4の合同研向けに作成)
  • tips
    • loem : Tip on handily distributed training
  • chair
    • ishikawa

2022-10-14 (Fri) 13:45~

  • seminar
    • panatchakorn: Measuring Gender Bias in Japanese Pre-trained Language Models from NLI Task

    • endo Decompression of JPEG compressed images using text information.pdf (5.5 MB)

  • tips
    • iida: SCQ Framework for writing introduction
  • chair
    • loem

2022-10-21 (Fri) 13:45~

  • seminar
    • iida Unsupervised Domain Adaptation for Sparse Retrieval by Filling Vocabulary and Word Frequency Gaps

    • cognetta

  • tips
    • yoshikawa Preparation for better annotation
  • chair
    • hirai

2022-10-28 (Fri) 13:45~

  • seminar
    • niwa: Does Example-based Text Generation Really Need to Use the Training History?
    • koyama Toward corpus construction and prediction of error span and type for japanese error correction
  • tips
    • muraoka
  • chair
    • nguyen

2022-11-04 (Fri) 13:45\~

  • seminar
    • sangwhan
    • liu
  • tips
    • maeda
  • chair
    • endo

2022-11-11 (Fri) 13:45~

  • seminar

    • yang: A study on challenges of scientific figure captioning.pdf (1.6 MB)
    • mizuki: Semantic Specialization for Knowledge-based Word Sense Disambiguation
  • tips
    • wang: Tips
  • chair
    • hattori

2022-11-18 (Fri) 13:45~

  • seminar
    • yoshikawa Selective Prediction for Confidence Aware Evaluation of Language Models
    • vijay CoLA Dataset is not a correct baseline for Acceptability Evaluation
  • tips
    • liu beamer_for_Presentation.pdf
  • chair
    • taniguchi

2022-11-25 (Fri) 13:45~

  • seminar
    • ma: Guiding Attention with Evidence for DocRE

    • emil

  • tips
    • erick Tips-Creating quick demos
  • chair
    • maeda

2022-12-02 (Fri) 13:45~

2022-12-07 (Wed) 17:30~

  • seminar
    • wang: Toward Improving Contextual Diversity: Data Augmentation for Aspect Sentiment Quad Prediction with Paraphrase Generation
    • loem: Can we guide specific task settings to pre-trained models via prompts in GEC tasks?
  • tips
    • maeda
  • chair
    • panatchakorn

2022-12-14 (Wed) 17:30~

  • seminar
    • taniguchi: Testing dependence on sentence generation direction in large-scale language models

    • panatchakorn: Measuring Gender Bias in Japanese Pre-trained Language Models from NLI Task

  • tips
    • muraoka: Tips for PowerPoint
  • chair
    • loem

2022-12-21 (Wed) 17:30~

  • seminar
    • endo
      • 画像キャプション生成におけるJPEG圧縮の軽減
    • hattori
      • Query Generation for “QA-style” Summarization of Long Documents
  • tips
    • skipped
  • chair
    • ishikawa

2023-01-18 (Wed) 17:30~

  • seminar
    • nguyen - DAS_VQA: Dual Awareness Systems for Improving Generalization in Visual Question Answering

    • erick - 6060-Research Seminar, Erick Mendieta.pdf (1.6 MB)

  • tips
    • ma
      • Marp: Markdown Tool for Fast Slides Making
      • slides, src
  • chair
    • endo

2023-01-25 (Wed) 17:30~

  • seminar
    • hirai: Improvement of mask language model by Vokenization considering diversity of assigned images
    • ishikawa: slides
  • tips
    • mizuki: Consequence of AI-assisted writing towards paper writing practices
  • chair
    • nguyen

2023-02-01 (Wed) 17:30~

  • seminar
    • cognetta - TrimmedBPE: You Are Probably Wasting A Lot of Parameters on Embeddings

    • koyama Evaluation of Japanese Error Correction

  • tips
    • taniguchi: Permission in Linux
  • chair
    • hattori