Kazuhiro Takemoto

A morphologically diverse freshwater microalgae dataset for deep learning-based classification with transfer learning analysis Accurate identification of freshwater microalgae is critical for early detection of harmful algal blooms, water quality monitoring, and sustainable bi…

[New Paper] We developed LMFM-12, the largest public freshwater microalgae dataset. Our analysis demonstrates the importance of domain-specific datasets for microscopy-based classification. PhD student research.　www.sciencedirect.com/science/arti...

A morphologically diverse freshwater microalgae dataset for deep learning-based classification with transfer learning analysis Accurate identification of freshwater microalgae is critical for early detection of harmful algal blooms, water quality monitoring, and sustainable bi…

[論文出版] 淡水微細藻類の最大公開データセットLMFM-12を構築し、顕微鏡画像分類におけるドメイン特化データセットの重要性を示しました。博士学生の研究です。 www.sciencedirect.com/science/arti...

Preprint: We propose a method for highly accurate prediction of direct reprogramming induction factors using Graph Attention Networks with biological interpretability. www.biorxiv.org/content/10.6...

プレプリント：Graph Attention Networkを用いてダイレクトリプログラミング誘導因子を高精度で予測するとともにその生物的解釈も行える方法を提案しました。弊学科 濱野研 修士学生の成果（私はお手伝い）。 www.biorxiv.org/content/10.6...

Scaling Laws for Moral Machine Judgment in Large Language Models Autonomous systems increasingly require moral judgment capabilities, yet whether these capabilities scale predictably with model size remains unexplored. We systematically evaluate 75 large language m...

プレプリント。スケールさえすればLLMには道徳的判断能力が出現する・・・かも。 arxiv.org/abs/2601.17637

Scaling Laws for Moral Machine Judgment in Large Language Models Autonomous systems increasingly require moral judgment capabilities, yet whether these capabilities scale predictably with model size remains unexplored. We systematically evaluate 75 large language m...

New preprint: Moral capabilities may emerge in LLMs simply by scaling up. arxiv.org/abs/2601.17637

修士学生がGIW XXXIV ISCB-Asia 2025で医療分野でのVLMの敵対的脆弱性に関する口頭発表しました。Shiroshima, a master student, gave a talk on adversarial vulnerability of VLM for healthcare in GIW XXXIV ISCB-Asia 2025. academic.oup.com/bib/article/...

Results from our short-term visiting student (April-June 2025). Congrats! 短期訪問学生(2025.4-6)の成果です。おめでとう。
ieeexplore.ieee.org/document/112...

𝗗𝘂𝗻𝗰𝗮𝗻 𝗪𝗮𝘁𝘁𝘀 & 𝗦𝘁𝗲𝘃𝗲𝗻 𝗦𝘁𝗿𝗼𝗴𝗮𝘁𝘇 will give their first-ever 𝗷𝗼𝗶𝗻𝘁 𝗸𝗲𝘆𝗻𝗼𝘁𝗲 at NetSci 2026! Their groundbreaking work has shaped how we understand networks, & this session will be a highlight of NetSci’s 20th anniversary.
Call for abstracts: tinyurl.com/3tbj2v83
Call for satellites: tinyurl.com/42sru6kz

New preprint. In Prisoner's Dilemma experiments with LLMs, we quantitatively show they are "cooperative optimists", cooperating while expecting others not to. LLMs appear vulnerable to exploitation. Joint work with A. Orland (Corvinus). papers.ssrn.com/sol3/papers....

プレプリントです。LLMの囚人のジレンマ実験から、相手が協力しないと予想しながらも自分は協力する「お人好し」であることを定量的に示しました。LLMは搾取に脆弱であるようです。ブダペスト・コルヴィヌス大との共同です。
papers.ssrn.com/sol3/papers....

Chaos: Adversarial control of synchronization in complex oscillator networks
https://pubs.aip.org/aip/cha/article/35/10/103109/3365929/Adversarial-control-of-synchronization-in-complex

出版されました。 pubs.aip.org/aip/cha/arti...

published in pubs.aip.org/aip/cha/arti...

Doku & Reportage : Unser Leben mit KI: Nie wieder Stau! - hier anschauen Welche Chancen bietet Künstliche Intelligenz für das Stau-Land NRW? Kann KI dafür sorgen, dass wir schneller und sicher ankommen – oder wenigstens bequemer? Moderatorin Catherine Vogel untersucht, wi...

ドイツ公共放送のドキュメンタリー番組「Unser Leben mit KI (Our Life with AI)」で、自動運転AIの道徳的判断について少しだけお話ししました。 www.ardmediathek.de/video/Y3JpZD...

Doku & Reportage : Unser Leben mit KI: Nie wieder Stau! - hier anschauen Welche Chancen bietet Künstliche Intelligenz für das Stau-Land NRW? Kann KI dafür sorgen, dass wir schneller und sicher ankommen – oder wenigstens bequemer? Moderatorin Catherine Vogel untersucht, wi...

I had the opportunity to briefly explain moral decision-making in autonomous driving AI on the WDR's documentary "Unser Leben mit KI (Our Life with AI)". www.ardmediathek.de/video/Y3JpZD...

Künstliche Intelligenz im Verkehr: Wie die KI bei einem Unfall über die Opfer entscheidet Das autonom fahrende Auto ist längst in der Entwicklung. Wie entscheidet die KI bei einem Autounfall, wer überlebt und wer nicht?

私たちのLLMのモラルマシン実験がWDR（西ドイツ放送）のニュースで取り上げられました。Moral Machine experiments on LLMs, our study, is featured in WDR. www1.wdr.de/nachrichten/...

Managed Healthcare Executiveが研究を紹介してくれました. www.managedhealthcareexecutive.com/view/skin-ca...

Managed Healthcare Executive featured this study. www.managedhealthcareexecutive.com/view/skin-ca...

Adversarial control of synchronization in complex oscillator networks This study investigates adversarial attacks, a concept from deep learning, designed to control synchronization dynamics through strategically crafted minimal perturbations. We propose a gradient-based...

[Preprint] Applied AI adversarial attacks to oscillator network sync control. Small phase perturbations effectively control synchronization. Useful for power grid stabilization & abnormal brain sync suppression. arxiv.org/abs/2506.02403

Adversarial control of synchronization in complex oscillator networks This study investigates adversarial attacks, a concept from deep learning, designed to control synchronization dynamics through strategically crafted minimal perturbations. We propose a gradient-based...

【プレプリント】AIの敵対的攻撃手法を振動子ネットワークの同期制御に応用しました。小さな位相摂動で振動子ネットワークの同期を効果的に制御できることを示しています。電力網の安定化や脳の異常同期の抑制など、広く同期制御に役立つかも知れません。博士&修士学生の研究です。 arxiv.org/abs/2506.02403

出版されました！ www.nature.com/articles/s41...

Published! www.nature.com/articles/s41...

🤖LLMのモラルマシン大規模実験が公開されました。その時、AIは誰を見殺しにするのか。 journals.plos.org/plosone/arti...

Large-scale moral machine experiment on large language models The rapid advancement of Large Language Models (LLMs) and their potential integration into autonomous driving systems necessitates understanding their moral decision-making capabilities. While our pre...

Our large-scale Moral Machine experiment across 52 LLMs is out. 🤖 When AI decides who lives and who dies. journals.plos.org/plosone/arti...

Great new paper by Stagnaro & Amsalem in NatComm:
Detailed fact-based informational module about gun control leads to persistent reductions in polarization (attitudes move towards midpoint). Seeing the evidence cited by other side makes opinions more moderate www.nature.com/articles/s41...

Ecological Networks Ecosystems are often represented as networks. The structure and function of ecological networks are significant in ecology. The availability of ecolog…

"Encyclopedia of Bioinformatics and Computational Biology第2版の「Ecological Networks」を更新しました。ネットワーク科学がどのように複雑な生態系を解明し、環境評価に応用されているかを概説しています。
www.sciencedirect.com/science/arti...

Ecological Networks Ecosystems are often represented as networks. The structure and function of ecological networks are significant in ecology. The availability of ecolog…

Excited to share our updated chapter on Ecological Networks in the Encyclopedia of Bioinformatics (2nd Ed). Discover how network science illuminates complex ecosystems and environmental assessment applications. www.sciencedirect.com/science/arti...

New preprint: We found that skin cancer detection apps can be easily fooled using transparent camera stickers with ASR of 50-80%. This vulnerability could affect clinical diagnoses and highlights critical security concerns in medical AI applications.
www.researchsquare.com/article/rs-5...