ICML 2025 Workshop HiLD Submissions

Fundamental Limits of Learning Single-Index Models under Structured Data
Jivan Waber, Alireza Mousavi-Hosseini, Murat A Erdogdu
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Oral
- Readers: Everyone
Data Free Metrics Are Not Reparameterisation Invariant Under the Critical and Robust Layer Phenomena
Gabryel Mason-Williams, Israel Mason-Williams, Fredrik Dahlqvist
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Understanding Deep Learning Requires Rethinking Sharpness
Israel Mason-Williams, Gabryel Mason-Williams, Helen Yannakoudakis
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
The Silent Helper: How Implicit Regularization Enhances Group Robustness
Nahal Mirzaie, Mahdi Ghaznavi, Hosna Oyarhoseini, Alireza Alipanah, Erfan Sobhaei, Ali Abbasi, Amirmahdi Farzane, Hossein Jafarinia, Parsa Sharifi Sedeh, Arefe Boushehrian, Mahdieh Soleymani Baghshah, Mohammad Hossein Rohban
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Adam Reduces a Unique Form of Sharpness: Theoretical Insights Near the Minimizer Manifold
Xinghan Li, Haodong Wen, Kaifeng Lyu
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Learning curves theory of hierarchically compositional data with power-law distributed features
Francesco Cagnetta, Hyunmo Kang, Matthieu Wyart
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Spectral Dynamics of Contrastive Learning with Spurious Features
Naghmeh Ghanooni, Dennis Wagner, Waleed Mustafa, Anthony Widjaja Lin, Sophie Fellenz, Marius Kloft
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Catalyst: Structured Pruning with Robust Bifurcation Dynamics
Jaeheun Jung, Donghun Lee
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
A Random Matrix Perspective on the Learning Dynamics of Multi-Head Latent Attention
Nandan Kumar Jha, Brandon Reagen
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Reactivation: Empirical NTK Dynamics Under Task Shifts
Yuzhi LIU, Zixuan Chen, Zirui zhang, Yufei Liu, Giulia Lanzillotta
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Implicit Bias and Loss of Plasticity in Matrix Completion: Depth Promotes Low-Rank Solutions
Baekrok Shin, Chulhee Yun
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
The Price of Robustness: Stable Classifiers Need Overparameterization
Jonas von Berg, Adalbert Fono, Massimiliano Datres, Sohir Maskey, Gitta Kutyniok
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
The Cost of Robustness: Tighter Bounds on Parameter Complexity for Robust Memorization in ReLU Nets
Yujun Kim, Chaewon Moon, Chulhee Yun
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Bayes optimal learning of attention-indexed models
Fabrizio Boncoraglio, Emanuele Troiani, Vittorio Erba, Lenka Zdeborova
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Oral
- Readers: Everyone
Symmetries in Weight Space Learning: To Retain or Remove?
Fynn Kiwitt, Behrooz Tahmasebi, Stefanie Jegelka
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Understanding Lookahead Dynamics Through Laplace Transforms
Aniket Sanyal, Tatjana Chavdarova
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Emergent Specialization: Rare Token Neurons in Language Models
Jing Liu, Haozheng Wang, Yueheng Li
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Information-Geometric Neural Granger Causality
Pauline Bourigault, Danilo Mandic
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Adapting to High Dimensional Concepts with Metalearning
Max Gupta
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
When Do Transformers Outperform Feedforward and Recurrent Networks? A Statistical Perspective
Alireza Mousavi-Hosseini, Clayton Sanford, Denny Wu, Murat A Erdogdu
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Oral
- Readers: Everyone
Rethinking Memorization–Generalization Trade-Off in Generative Models
Jiseok Chae, Kyuwon Kim, Donghwan Kim
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Theoretical Guarantees and Training Dynamics of Contrastive Learning: How Misaligned Data Influence Feature Purity
Jiawei Sun, Shuai Zhang, Hongkang Li, Meng Wang
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
How Compositional Generalization and Creativity Improve as Diffusion Models are Trained
Alessandro Favero, Antonio Sclocchi, Francesco Cagnetta, Pascal Frossard, Matthieu Wyart
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Gluon: Making Muon & Scion Great Again! (Bridging Theory and Practice of LMO-based Optimizers for LLMs)
Artem Riabinin, Egor Shulgin, Kaja Gruntkowska, Peter Richtárik
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Langevin Learning Dynamics in Lazy and Non-Lazy Wide Neural Networks
Yehonatan Avidan, Haim Sompolinsky
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone