Oral Sessions at CPAL 2025
A select number of papers from the CPAL 2025 Proceedings Track will be presented as oral presentations at the conference. The oral presentations are listed below, in their corresponding oral sessions.
Highlight Talks 1
Time: Day 2 (Mar 25) – Tuesday – 10:00 AM to 10:30 AM
Progressive Gradient Flow for Robust N:M Sparsity Training in Transformers
Abhimanyu Rajeshkumar Bambhaniya, Amir Yazdanbakhsh, Suvinay Subramanian, Sheng-Chun Kao, Shivani Agrawal, Utku Evci, Tushar Krishna
Keywords: N:M structured sparsity, sparsity, model compression, attention-based models, sparse training recipe
Improving Neuron-level Interpretability with White-box Language Models
Hao Bai, Yi Ma
Keywords: White-box models, deep learning architectures, neuron-level interpretation
A unified framework for Sparse plus Low-Rank Matrix Decomposition for LLMs
Mehdi Makni, Kayhan Behdin, Zheng Xu, Natalia Ponomareva, Rahul Mazumder
Keywords: model compression, sparse plus low-rank, optimization, inference acceleration, 2:4 sparsity, hardware and system co-design
Approximate Nullspace Augmented Finetuning for Robust Vision Transformers
Haoyang Liu, Aditya Singh, Yijiang Li, Haohan Wang
Keywords: Robustness, Vision Transformer, Invariance
Highlight Talks 2
Time: Day 2 (Mar 25) – Tuesday – 12:00 PM to 12:30 PM
Closure Discovery for Coarse-Grained Partial Differential Equations Using Grid-based Reinforcement Learning
Jan-Philipp von Bassewitz, Sebastian Kaltenbach, Petros Koumoutsakos
Keywords: Closure Discovery, Inductive Bias, Multi-Agent Reinforcement Learning
The Computational Limits of State-Space Models and Mamba via the Lens of Circuit Complexity
Yifang Chen, Xiaoyu Li, Yingyu Liang, Zhenmei Shi, Zhao Song
Keywords: State-Space Models, Mamba, Circuit Complexity, Computational Limits
Fast John Ellipsoid Computation with Differential Privacy Optimization
Xiaoyu Li, Yingyu Liang, Zhenmei Shi, Zhao Song, Junwei Yu
Keywords: Fast Optimization, Differential Privacy, John Ellipsoid Computation
Sufficient and Necessary Explanations (and What Lies in Between)
Beepul Bharti, Paul Yi, Jeremias Sulam
Keywords: interpretability, explainability
Highlight Talks 3
Time: Day 4 (Mar 27) – Thursday – 10:00 AM to 11:00 AM
Vanishing Feature: Diagnosing Model Merging and Beyond
Xingyu Qu, Samuel Horváth
Keywords: Model Merging, Efficiency, Deep Learning, Efficient Deep Learning
A Case Study of Low Ranked Self-Expressive Structures in Neural Network Representations
Uday Singh Saini, William Shiao, Yahya Sattar, Yogesh Dahiya, Samet Oymak, Evangelos E. Papalexakis
Keywords: Subspace Clustering, Centered Kernel Alignment, Representation Similarity Measures.
Hamiltonian Mechanics of Feature Learning: Bottleneck Structure in Leaky ResNets
Arthur Jacot, Alexandre Kaiser
Keywords: Low-rank bias, NeuralODE, Hamiltonian, Bottleneck structure
You Only Debias Once: Towards Flexible Accuracy-Fairness Trade-offs at Inference Time
Xiaotian Han, Tianlong Chen, Kaixiong Zhou, Zhimeng Jiang, Zhangyang Wang, Xia Hu
Keywords: fairness, weight space, neural network subspace