論文・研究業績

From Inner Randomness to Outer Stability

Nishant Sharma, Ganesh Talluri, Hiroki Naganuma

Under Review, 2026年9月（査読中）

Which Geometry on Which Layer? A Principled Criterion for Mixed-Optimizer Training

Hiroki Naganuma^† (同等貢献), Laura Gomezjurado Gonzalez^† (同等貢献), Mahdi Ghaznavi^† (同等貢献), Atsushi Nitanda, Ioannis Mitliagkas, Seng Pei Liew, Ryuichiro Hataya

Under Review, 2026年5月（査読中）

Orth-Dion: Eliminating Geometric Mismatch in Distributed Low-Rank Spectral Optimization

Tatsuhiro Nakamori^† (同等貢献), Laura Gomezjurado Gonzalez^† (同等貢献), Ganesh Talluri^† (同等貢献), Ansh Tiwari, Hideyuki Kawashima, Guillaume Rabusseau, Ioannis Mitliagkas, Hiroki Naganuma

Under Review, 2026年5月（査読中）

Convergence Bound and Critical Batch Size of Muon Optimizer

Naoki Sato^† (同等貢献), Hiroki Naganuma^† (同等貢献), Hideaki Iiduka

Under Review, 2025年7月（査読中）

When Does Alignment Help? A Comparative Study of DCCA and Fusion-Based Approaches for Multi-modal Chest X-ray Analysis

Hiroki Naganuma, Akira Uehara, Takeru Sakurai, Yoshiyuki Sankai

SSRN, 2025年4月

Augmenting NER Datasets with LLMs: Towards Automated and Refined Annotation

Yuji Naraki^† (同等貢献), Ryosuke Yamaki^† (同等貢献), Yoshikazu Ikeda, Takafumi Horie, Kotaro Yoshida, Ryotaro Shimizu, Hiroki Naganuma

arXiv, 2024年3月

Takeuchi's Information Criteria as Generalization Measures for DNNs Close to NTK Regime

Hiroki Naganuma, Taiji Suzuki, Rio Yokota, Masahiro Nomura, Kohta Ishikawa, Ikuro Sato

arXiv, 2021年9月

The Geometry of Spectral Gradient Descent: Layerwise Criteria for SignSGD vs SpecSGD

Hiroki Naganuma^† (同等貢献), Laura Gomezjurado Gonzalez^† (同等貢献), Mahdi Ghaznavi^† (同等貢献), Ioannis Mitliagkas

ICLR 2026 Workshop, 2026年2月

Smoothness-Adaptive Sharpness Aware Minimization for Finding Flatter Minima

Hiroki Naganuma^† (同等貢献), Junhyung Lyle Kim^† (同等貢献), Anastasios Kyrillidis, Ioannis Mitliagkas

ICLR 2024 Workshop, 2024年5月

workshopposter

Story-to-Images Translation: Leveraging Diffusion Models and Large Language Models for Sequence Image Generation

Haruka Kumagai, Ryosuke Yamaki, Hiroki Naganuma

NarSUM 2023, 2023年10月

workshoptravel grant

On the Interplay of Curvature, Calibration and Out-of-Distribution Generalization: Insights from SAM and Focal Loss Analyses

Hiroki Naganuma^† (同等貢献), Masanari Kimura^† (同等貢献)

UNCV 2023, 2023年10月

poster

Necessary and Sufficient Hypothesis of Curvature: Understanding Connection Between Out-of-Distribution Generalization and Calibration

Hiroki Naganuma^† (同等貢献), Masanari Kimura^† (同等貢献)

ICLR 2023 Workshop, 2023年5月

Towards Understanding the Relationship of Batch Size and Iterations in Deep Learning

Hiroki Naganuma

MLSS 2020, 2020年6月

On Empirical Analysis of Layer-wise Learning Rate Schedule

Hiroki Naganuma, Rio Yokota

ACML 2019 Workshop, 2019年11月

A Performance Improvement Approach for Second-Order Optimization in Large Mini-batch Training

Hiroki Naganuma, Rio Yokota

HPML 2019, 2019年5月

slides

Noise Injection Leads to Better Generalization in Large Mini-Batch Training

Hiroki Naganuma, Rio Yokota

Tokyo Tech-Stony Brook Joint Meeting, 2019年5月