Skip to content
View ymgw55's full-sized avatar

Highlights

  • Pro

Block or report ymgw55

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ymgw55/README.md

GitHub Stats Card

Portfolio

https://ymgw55.github.io/

Biography

My name is Hiroaki Yamagiwa. I'm a final year P.h.D student at Shimodaira Lab, Kyoto University.

Research Interests

  • The properties of embeddings, especially their distributions and interpretability.
  • Understanding the internal structure of language models.

Contact

  • hiroaki.yamagiwa [at] sys.i.kyoto-u.ac.jp (Please replace [at] with @)
  • Alternatively, you can reach out to me via direct message on X or LinkedIn.

Research Profiles

Social Profiles

Education

  • PhD student in Informatics, Graduate School of Informatics, Kyoto University, 2022-present.
  • MSc in Informatics, Graduate School of Informatics, Kyoto University, 2020-2022.
  • BSc in Mathematics, Department of Mathematics, Kyoto University, 2015-2020.

Pinned Loading

  1. shimo-lab/Universal-Geometry-with-ICA shimo-lab/Universal-Geometry-with-ICA Public

    Discovering Universal Geometry in Embeddings with ICA (Published in EMNLP 2023)

    Python 16

  2. WSMD WSMD Public

    Improving word mover’s distance by leveraging self-attention matrix (Published in EMNLP 2023 Findings)

    Python 9

  3. Axis-Tour Axis-Tour Public

    Axis Tour: Word Tour Determines the Order of Axes in ICA-transformed Embeddings (Published in EMNLP 2024 Findings)

    Python

  4. Cosine-Similarity-via-ICA Cosine-Similarity-via-ICA Public

    Revisiting Cosine Similarity via Normalized ICA-transformed Embeddings (Published in COLING 2025)

    Mathematica 2

  5. Norm-and-Variance Norm-and-Variance Public

    Norm of Mean Contextualized Embeddings Determines their Variance (Published in COLING 2025)

    Python 1

  6. SCESAME SCESAME Public

    Zero-Shot Edge Detection with SCESAME: Spectral Clustering-based Ensemble for Segment Anything Model Estimation (Published in WACV 2024 Workshop)

    C++ 17 1