ICML 2025 Workshop HiLD Submissions

Bayesian Influence Functions for Scalable Data Attribution
Philipp Alexander Kreer, Wilson Wu, Maxwell Adam, Zach Furman, Jesse Hoogland
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Quantization and the Bottom of the Loss Landscape
Luca Di Carlo, Daniel T. Bernstein, David J. Schwab
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Two-point deterministic equivalence for SGD in random feature models
Alexander Atanasov, Blake Bordelon, Jacob A Zavatone-Veth, Courtney Paquette, Cengiz Pehlevan
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Probing Geometry of Next Token Prediction Using Cumulant Expansion of the Softmax Entropy
Karthik Viswanathan, Sang Eon Park
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
When Can You Get Away with Low Memory Adam?
Dayal Singh Kalra, John Kirchenbauer, Maissam Barkeshli, Tom Goldstein
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Latent Concept Disentanglement in Transformer-based Language Models
Guan Zhe Hong, Bhavya Vasudeva, Vatsal Sharan, Cyrus Rashtchian, Prabhakar Raghavan, Rina Panigrahy
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Benignity of loss landscape with weight decay requires both large overparametrization and initialization
Etienne Boursier, Matthew Bowditch, Matthias Englert, Ranko Lazic
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
From Information to Generative Exponent: Learning Rate Induces Phase Transitions in SGD
Konstantinos Christopher Tsiolis, Alireza Mousavi-Hosseini, Murat A Erdogdu
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Jacobian Alignment Explains Grokking and Centroid Alignment Identifies It
Thomas Walker, Ahmed Imtiaz Humayun, Randall Balestriero, Richard Baraniuk
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Topology-Aware Robust Representation Balancing for Estimating Causal Effects
Amirhossein Farzam, Ahmed Aloui, Vahid Tarokh, Guillermo Sapiro
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
On the Interaction of Noise, Compression, and Adaptivity under $(L_0,L_1)$-Smoothness: An SDE Approach
Enea Monzio Compagnoni, Rustem Islamov, Antonio Orvieto, Eduard Gorbunov
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Better Rates for Private Linear Regression in the Proportional Regime via Aggressive Clipping
Simone Bombari, Inbar Seroussi, Marco Mondelli
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
New Evidence of the Two-Phase Learning Dynamics of Neural Networks
Zhanpeng Zhou, Yongyi Yang, Mahito Sugiyama, Junchi Yan
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Universal Dynamics of Warmup Stable Decay: understanding WSD beyond Transformers
Annalisa Belloni, Lorenzo Noci, Antonio Orvieto
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Reduce and Conquer: Independent Component Analysis at linear sample complexity
Fabiola Ricci, Lorenzo Bardone, Sebastian Goldt
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Oral
- Readers: Everyone
Feature learning is decoupled from generalization in high capacity neural networks
Niclas Alexander Göring, Charles London, Abdurrahman Hadi Erturk, Chris Mingard, Yoonsoo Nam, Ard A. Louis
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Studying Data Complexity and Learned Structure in Neural Networks with Bayesian Probes
Maxwell Adam, Zach Furman, Wilson Wu, Philipp Alexander Kreer, Jesse Hoogland
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
On the Existence of Hidden Subnetworks Within a Randomly Weighted Multi-Head Attention Mechanism
Hikari Otsuka, Yasuyuki Okoshi, Daichi Fujiki, Susumu Takeuchi, Masato Motomura, Daiki Chijiwa
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
How Transformers Get Rich: Training Dynamics Analysis
Mingze Wang, Ruoxi Yu, Weinan E, Lei Wu
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Lyapunov Learning at the Onset of Chaos
Alessandro Londei, Denise Lanzieri, Matteo Benati, Vittorio Loreto
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Neural Collapse is Globally Optimal in Deep Regularized ResNets and Transformers
Peter Súkeník, Christoph H. Lampert, Marco Mondelli
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Implicit Bias of Spectral Descent and Muon on Multiclass Separable Data
Chen Fan, Mark Schmidt, Christos Thrampoulidis
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
A solvable generative model with a linear, one-step denoiser
Indranil Halder
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Exact Learning of Permutations for Nonzero Binary Inputs with Logarithmic Training Size and Quadratic Ensemble Complexity
George Giapitzakis, Artur Back de Luca, Kimon Fountoulakis
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone
Quantitative Bounds for Length Generalization in Transformers
Zachary Izzo, Eshaan Nichani, Jason D. Lee
- Published: 09 Jun 2025, Last Modified: 09 Jun 2025
- HiLD at ICML 2025 Poster
- Readers: Everyone