About Me

I am a first-year Ph.D. student in Computer Science at the University of California, Berkeley, advised by Profs. Jason D. Lee and Song Mei. My interests lie in the mathematical foundations of deep learning, with a focus on nonconvex optimization, dynamical analysis, and computational-statistical guarantees for neural networks. I am also interested in understanding emergent capabilities of foundation models such as chain-of-thought reasoning.

I will be interning at Google DeepMind during the summer of 2026, hosted by Hossein Mobahi.

Previously, I received my B.Sc. in Mathematics and Statistics at Seoul National University as Valedictorian of ‘23, and my M.Sc. in Mathematical Informatics at the University of Tokyo advised by Prof. Taiji Suzuki, where I received the Dean’s Award for outstanding research.

Selected Publications

Sharp Capacity Scaling of Spectral Optimizers in Learning Associative Memory
Juno Kim*, Eshaan Nichani*, Denny Wu, Alberto Bietti, Jason D. Lee. Under review.

All Publications

When Does Online Imitation Learning Help in LLM Post-Training? The Role of (Non-)Realizability Beyond Horizon
Huaqing Zhang, Jingchu Gai, Juno Kim, Bingbin Liu, Andrej Risteski. Under review.
Sharp Capacity Thresholds in Linear Associative Memory: From Winner-Take-All to Listwise Retrieval
Nicholas Barnfield, Juno Kim, Eshaan Nichani, Jason D. Lee, Yue M. Lu. Preprint.
Sharp Capacity Scaling of Spectral Optimizers in Learning Associative Memory
Juno Kim*, Eshaan Nichani*, Denny Wu, Alberto Bietti, Jason D. Lee. Under review.
Coverage Improvement and Fast Convergence of On-policy Preference Learning
Juno Kim, Jihun Yun, Jason D. Lee, Kwang-Sung Jun. ICML 2026.
Alignment as Distribution Learning: Your Preference Model is Explicitly a Language Model
Jihun Yun*, Juno Kim*, Jongho Park, Junhyuck Kim, Jongha Jon Ryu, Jaewoong Cho, Kwang-Sung Jun. ICML 2025 MoFA Workshop.
Mirror Mean-Field Langevin Dynamics
Anming Gu*, Juno Kim*. ICML 2026.
Hessian-guided Perturbed Wasserstein Gradient Flows for Escaping Saddle Points
Naoya Yamamoto, Juno Kim, Taiji Suzuki. NeurIPS 2025.
Optimality and Adaptivity of Deep Neural Features for Instrumental Variable Regression
Juno Kim, Dimitri Meunier, Arthur Gretton, Taiji Suzuki, Zhu Li. ICLR 2025.
Transformers are Minimax Optimal Nonparametric In-Context Learners
Juno Kim, Tai Nakamaki, Taiji Suzuki. NeurIPS 2024 and ICML 2024 TF2M Workshop, Best Paper Award.
Symmetric Mean-field Langevin Dynamics for Distributional Minimax Problems
Juno Kim, Kakei Yamamoto, Kazusato Oko, Zhuoran Yang, Taiji Suzuki. ICLR 2024 Spotlight.
t3-Variational Autoencoder: Learning Heavy-tailed Data with Student's t and Power Divergence
Juno Kim*, Jaehyuk Kwon*, Mincheol Cho*, Hyunjong Lee, Joong-Ho Won. ICLR 2024.
Hessian Based Smoothing Splines for Manifold Learning
Juno Kim, Otto van Koert. Preprint.
Reeb Flows without Simple Global Surfaces of Section
Juno Kim, Yonghwan Kim, Otto van Koert. Involve, 15(5), pp. 813–842, 2022. (*equal contribution)

Education

University of California, Berkeley

2025 - current

Ph.D. student in EECS

University of Tokyo

2023 - 2025

M.S. in Mathematical Informatics
Thesis: Statistical and Dynamical Analysis of Transformers: In-Context Learning and Chain-of-Thought Reasoning

Seoul National University

2018 - 2023

B.S. in Statistics
B.S. in Mathematics
Graduated Valedictorian of the College of Natural Sciences (GPA 4.28/4.3)
Thesis: Token and Corpus Imputation in Statistical Language Modeling via Semantic Embeddings, Hessian Based Smoothing Splines for Manifold Learning

Awards

Dean's Award for Research Achievement, IST, University of Tokyo

2025

Doctoral Course (DC1) Research Fellowship, JSPS

declined

Japanese Government Scholarship

2023 - 2025

President Award, Highest Honors, Seoul National University

2023

President Award, Korean Statistical Society

2023

National Scholarship, Kwanjeong Educational Foundation

2020 - 2023

4th Place, Simon Marais Mathematics Competition

2020

Eminence Scholarship, Seoul National University

2018 - 2020

Gold Prize, College Mathematics Competition

2019

Experience

Google DeepMind

2026
Student Researcher

KRAFTON AI

2025
Deep Learning Div. Core Research Team, Intern

RIKEN Center for Advanced Intelligence Project

2023 - 2025
Part-time Researcher

Seoul National University

2019 - 2020, 2022 - 2023
Undergraduate Research Intern

Reviewer

AISTATS'24, ICML'24,25,26 (Gold Reviewer), NeurIPS'24,25,26, ICLR'25 (Notable Reviewer),26

Talks and Visits

KAIST AI (Host: Chulhee Yun)

Jul 14, 2026

[Oral] ICML 2026 HiLD Workshop (Seoul, South Korea)

Jul 10, 2026

Tsinghua University (Host: Kaifeng Lyu) (online)

May 15, 2026

Flatiron Institute (Host: Alberto Bietti)

May 15, 2026

University of Pennsylvania (Host: Yuxin Chen)

May 14, 2026

Kempner Institute, Harvard University (Host: Bingbin Liu)

May 13, 2026

Carnegie Mellon University (Host: Andrej Risteski)

May 11, 2026

NLP Colloquium, Japan (Host: Sho Yokoi) (online)

May 21, 2025

[Oral] ICLR 2025 (Singapore)

Apr 26, 2025

Flatiron Institute (Host: Denny Wu)

Mar 15, 2025

Vector Institute (Host: Anastasis Kratsios) (online)

Jan 24, 2025

[Visit] Simons Institute for the Theory of Computing

Nov 11 - Dec 9, 2024

[Oral] ICML 2024 TF2M Workshop (Vienna, Austria)

Jul 27, 2024

[Oral] ICML 2024 (Vienna, Austria)

Jul 23, 2024

[Visit] UCL Gatsby Computational Neuroscience Unit (Host: Arthur Gretton)

Aug 5 - 19, 2024

[Visit] NYU Center for Data Science

Dec 16 - 19, 2023