NeurIPS 2025 Workshop on Constrained Optimization for Machine Learning

Accepted Contributions

Contributed Talks

We are glad to present the following accepted contributed talks. See the full schedule on the Schedule page.

1. [11:30 - 11:45] Active-Set Identification by Stochastic and Noisy Optimization Algorithms for Constrained Learning
   Lara Zebiane, Frank Edward Curtis, Daniel Robinson

2. [11:45 - 12:00] Alignment of Large Language Models with Constrained Learning
   Botong Zhang, Shuo Li, Ignacio Hounie, Osbert Bastani, Dongsheng Ding, Alejandro Ribeiro

3. [14:10 - 14:25] Global Solutions to Non-Convex Functional Constrained Problems via Hidden Convexity
   Ilyas Fatkhullin, Niao He, Guanghui Lan, Florian Wolf

4. [14:25 - 14:40] HardNet: Hard-Constrained Neural Networks with Universal Approximation Guarantees
   Youngjae Min, Navid Azizan

5. [14:40 - 14:55] Infinity Embeddings: Representation learning with ultrametric structure for faster search via Constrained Learning
   Antonio Pariente Granero, Ignacio Hounie, Alejandro Ribeiro

Poster Session #1

• Learning (Approximately) Equivariant Networks via Constrained Optimization
• A Bilevel Optimization Approach for Computing Synthetic Data to Mitigate Unfairness in Collaborative Machine Learning
• ErA: Error-Aware Deep Unrolling Network for Single Image Defocus Debluring
• Hypercube-Constrained Graph Learning for Protein Fitness with Dynamic Laplacian Regularization
• Sample Complexity Bounds for Linear Constrained MDPs with a Generative Model
• Compressing Vision Transformers in Geospatial Transfer Learning with Manifold-Constrained Optimization
• Online Statistical Inference for Proximal Stochastic Gradient Descent under Markovian Sampling
• The Explore-Exploit Tradeoff Redefined: Balancing Regret and Treatment Effects in Contextual Bandits
• Exploring Black-box Adversarial Attacks on Low-rank Constrained Neural Networks
• Statistical Inference of Constrained Model Estimation via Derivative-Free Stochastic Sequential Quadratic Programming
• Active-Set Identification by Stochastic and Noisy Optimization Algorithms for Constrained Learning
• Fixed-Order Lexicographic Optimization via the $\lambda$-ladder Exponential Loss
• Fair Supervised Learning Through Constraints on Smooth Nonconvex Unfairness-Measure Surrogates
• Online Omniprediction with Long-Term Constraints
• humancompatible.train: Implementing Optimization Algorithms for Stochastically-Constrained Stochastic Optimization Problems
• Inexact Moreau Envelope Augmented Lagrangian Method for Nonconvex Robust Constrained Optimization
• A Constrained Multi-Agent Reinforcement Learning Approach to Autonomous Traffic Signal Control
• Improving Transfer Learning via Uniform Boundedness Prior
• Unrolled Neural Networks for Constrained Optimization
• Statistical Inference for Responsiveness Verification
• A Constrained Optimization Perspective of Unrolled Transformers

Poster Session #2

• FPS: Feedforward-based Parameter Selection For Efficient Fine-Tuning
• Cooper: A Library for Constrained Optimization in Deep Learning
• Alignment of Large Language Models with Constrained Learning
• Machine Unlearning Meets Adversarial Robustness via Constrained Interventions on LLMs
• Stochastic Safe-Set Projection
• Certified Training with Branch-and-Bound: A Case Study on Lyapunov-stable Neural Control
• Infinity Embeddings: Representation learning with ultrametric structure for faster search via Constrained Learning.
• FineAMP: Optimization-Based Automatic Mixed Precision Quantization for Efficient Diffusion Model Inference
• Constrained Active Regression via Ellipsoid Regularized Leverage Scores
• Generalized Constrained Flow Matching for Constraint-Aware Generative Modeling
• The Lagrangian Method for Solving Locally Constrained Markov Games
• Combining Online CUR Decomposition and Matrix Sketching for Data Streams in Open Feature Spaces
• Causal-Informed Hybrid Online Adaptive Optimization for Ad Load Personalization in Large-Scale Social Networks
• Global Solutions to Non-Convex Functional Constrained Problems via Hidden Convexity
• CAFL-L: Constraint-Aware Federated Learning with Lagrangian Dual Optimization for On-Device Language Models
• RL-Guided Data Selection for Language Model Finetuning
• HardNet: Hard-Constrained Neural Networks with Universal Approximation Guarantees
• Augmented Lagrangian for Constrained Learning
• SIMU: Selective Influence Machine Unlearning
• Sharpness-Aware Minimization Meets Spectral Norm
• Composition and Alignment of Diffusion Models using Constrained Learning