Liu Ziyin

According to our database, Liu Ziyin authored at least 48 papers between 2018 and 2025.

Collaborative distances:

• Dijkstra number of four.
• Erdős number of four.

2025

Proof of a perfect platonic representation hypothesis.

Ziyin Liu, Isaac L. Chuang

CoRR, July, 2025

Emergence of Hebbian Dynamics in Regularized Non-Local Learners.

David Koplow, Tomaso A. Poggio, Liu Ziyin

CoRR, May, 2025

Neural Thermodynamics I: Entropic Forces in Deep and Universal Representation Learning.

Liu Ziyin, Yizhou Xu, Isaac L. Chuang

CoRR, May, 2025

Heterosynaptic Circuits Are Universal Gradient Machines.

Liu Ziyin, Isaac L. Chuang, Tomaso A. Poggio

CoRR, May, 2025

Understanding the Emergence of Multimodal Representation Alignment.

Megan Tjandrasuwita, Chanakya Ekbote, Liu Ziyin, Paul Pu Liang

CoRR, February, 2025

Parameter Symmetry Breaking and Restoration Determines the Hierarchical Learning in AI Systems.

Liu Ziyin, Yizhou Xu, Tomaso A. Poggio, Isaac L. Chuang

CoRR, February, 2025

Compositional Generalization Requires More Than Disentangled Representations.

Qiyao Liang, Daoyuan Qian, Liu Ziyin, Ila Fiete

CoRR, January, 2025

Three Mechanisms of Feature Learning in a Linear Network.

Yizhou Xu, Ziyin Liu

Proceedings of the Thirteenth International Conference on Learning Representations, 2025

Remove Symmetries to Control Model Expressivity and Improve Optimization.

Liu Ziyin, Yizhou Xu, Isaac L. Chuang

Proceedings of the Thirteenth International Conference on Learning Representations, 2025

Formation of Representations in Neural Networks.

Liu Ziyin, Isaac L. Chuang, Tomer Galanti, Tomaso A. Poggio

Proceedings of the Thirteenth International Conference on Learning Representations, 2025

2024

Self-Assembly of a Biologically Plausible Learning Circuit.

Qianli Liao, Liu Ziyin, Yulu Gan, Brian Cheung, Mark Harnett, Tomaso A. Poggio

CoRR, 2024

Remove Symmetries to Control Model Expressivity.

Liu Ziyin, Yizhou Xu, Isaac L. Chuang

CoRR, 2024

The Implicit Bias of Gradient Noise: A Symmetry Perspective.

Liu Ziyin, Mingze Wang, Lei Wu

CoRR, 2024

When Does Feature Learning Happen? Perspective from an Analytically Solvable Model.

Yizhou Xu, Liu Ziyin

CoRR, 2024

Parameter Symmetry and Noise Equilibrium of Stochastic Gradient Descent.

Liu Ziyin, Mingze Wang, Hongchao Li, Lei Wu

Proceedings of the Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, 2024

Symmetry Induces Structure and Constraint of Learning.

Liu Ziyin

Proceedings of the Forty-first International Conference on Machine Learning, 2024

2023

Symmetry Leads to Structured Constraint of Learning.

Liu Ziyin

CoRR, 2023

Law of Balance and Stationary Distribution of Stochastic Gradient Descent.

Liu Ziyin, Hongchao Li, Masahito Ueda

CoRR, 2023

The Probabilistic Stability of Stochastic Gradient Descent.

Liu Ziyin, Botao Li, Tomer Galanti, Masahito Ueda

CoRR, 2023

On the Stepwise Nature of Self-Supervised Learning.

James B. Simon, Maksis Knutins, Liu Ziyin, Daniel Geisz, Abraham J. Fetterman, Joshua Albrecht

Proceedings of the International Conference on Machine Learning, 2023

spred: Solving L1 Penalty with SGD.

Liu Ziyin, Zihao Wang

Proceedings of the International Conference on Machine Learning, 2023

What shapes the loss landscape of self supervised learning?

Liu Ziyin, Ekdeep Singh Lubana, Masahito Ueda, Hidenori Tanaka

Proceedings of the Eleventh International Conference on Learning Representations, 2023

2022

Sparsity by Redundancy: Solving L₁ with a Simple Reparametrization.

Liu Ziyin, Zihao Wang

CoRR, 2022

Exact Phase Transitions in Deep Learning.

Liu Ziyin, Masahito Ueda

CoRR, 2022

Stochastic Neural Networks with Infinite Width are Deterministic.

Liu Ziyin, Hanlin Zhang, Xiangming Meng, Yuting Lu, Eric P. Xing, Masahito Ueda

CoRR, 2022

Exact Solutions of a Deep Linear Network.

Liu Ziyin, Botao Li, Xiangming Meng

Proceedings of the Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, 2022

Posterior Collapse of a Linear Latent Variable Model.

Zihao Wang, Liu Ziyin

Proceedings of the Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, 2022

Power-Law Escape Rate of SGD.

Takashi Mori, Liu Ziyin, Kangqiao Liu, Masahito Ueda

Proceedings of the International Conference on Machine Learning, 2022

SGD Can Converge to Local Maxima.

Liu Ziyin, Botao Li, James B. Simon, Masahito Ueda

Proceedings of the Tenth International Conference on Learning Representations, 2022

Strength of Minibatch Noise in SGD.

Liu Ziyin, Kangqiao Liu, Takashi Mori, Masahito Ueda

Proceedings of the Tenth International Conference on Learning Representations, 2022

Theoretically Motivated Data Augmentation and Regularization for Portfolio Construction.

Liu Ziyin, Kentaro Minami, Kentaro Imajo

Proceedings of the 3rd ACM International Conference on AI in Finance, 2022

2021

SGD May Never Escape Saddle Points.

Liu Ziyin, Botao Li, Masahito Ueda

CoRR, 2021

What Data Augmentation Do We Need for Deep-Learning-Based Finance?

Liu Ziyin, Kentaro Minami, Kentaro Imajo

CoRR, 2021

Logarithmic landscape and power-law escape rate of SGD.

Takashi Mori, Liu Ziyin, Kangqiao Liu, Masahito Ueda

CoRR, 2021

On Minibatch Noise: Discrete-Time SGD, Overparametrization, and Bayes.

Liu Ziyin, Kangqiao Liu, Takashi Mori, Masahito Ueda

CoRR, 2021

On the distributional properties of adaptive gradients.

Zhiyi Zhang, Ziyin Liu

Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, 2021

Cross-Modal Generalization: Learning in Low Resource Modalities via Meta-Alignment.

Paul Pu Liang, Peter Wu, Liu Ziyin, Louis-Philippe Morency, Ruslan Salakhutdinov

Proceedings of the MM '21: ACM Multimedia Conference, Virtual Event, China, October 20, 2021

Noise and Fluctuation of Finite Learning Rate Stochastic Gradient Descent.

Kangqiao Liu, Liu Ziyin, Masahito Ueda

Proceedings of the 38th International Conference on Machine Learning, 2021

2020

Stochastic Gradient Descent with Large Learning Rate.

Kangqiao Liu, Liu Ziyin, Masahito Ueda

CoRR, 2020

An Investigation of how Label Smoothing Affects Generalization.

Blair Chen, Liu Ziyin, Zihao Wang, Paul Pu Liang

CoRR, 2020

Volumization as a Natural Generalization of Weight Decay.

Liu Ziyin, Zihao Wang, Makoto Yamada, Masahito Ueda

CoRR, 2020

Learning Not to Learn in the Presence of Noisy Labels.

Liu Ziyin, Blair Chen, Ru Wang, Paul Pu Liang, Ruslan Salakhutdinov, Louis-Philippe Morency, Masahito Ueda

CoRR, 2020

LaProp: a Better Way to Combine Momentum with Adaptive Gradient.

Liu Ziyin, Zhikang Wang, Masahito Ueda

CoRR, 2020

Think Locally, Act Globally: Federated Learning with Local and Global Representations.

Paul Pu Liang, Terrance Liu, Liu Ziyin, Ruslan Salakhutdinov, Louis-Philippe Morency

CoRR, 2020

Neural Networks Fail to Learn Periodic Functions and How to Fix It.

Liu Ziyin, Tilman Hartwig, Masahito Ueda

Proceedings of the Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, 2020

2019

Deep Gamblers: Learning to Abstain with Portfolio Theory.

Liu Ziyin, Zhikang Wang, Paul Pu Liang, Ruslan Salakhutdinov, Louis-Philippe Morency, Masahito Ueda

Proceedings of the Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, 2019

2018

BlockPuzzle - A Challenge in Physical Reasoning and Generalization for Robot Learning.

Yixiu Zhao, Ziyin Liu

CoRR, 2018

Multimodal Language Analysis with Recurrent Multistage Fusion.

Paul Pu Liang, Ziyin Liu, Amir Zadeh, Louis-Philippe Morency

Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussels, Belgium, October 31, 2018