Ruoyu Sun

Bridging Formal Language with Chain-of-Thought Reasoning to Geometry Problem Solving

Knapsack RL: Unlocking Exploration of LLMs via Optimizing Budget Allocation

SAC: Adaptive Learning Rate Scaling with Architectural Constraints

ORGEval: Graph-Theoretic Evaluation of LLMs in Optimization Modeling

Feature Augmentation of GNNs for ILPs: Local Uniqueness Suffices

Prune, Then Select: Select High-Quality, Important, and Diverse Data Using Training Trajectories

Online SFT for LLM Reasoning

Long CoT In-Context Learning Can Empower Pre-trained LLMs

Adam-mini: Use Fewer Learning Rates To Gain More

Teaching Language Models to Reason with Tools

Self-Evolving Critique Abilities in Large Language Models

Preserving Diversity in Supervised Fine-Tuning of Large Language Models

When GNNs meet symmetry in ILPs: an orbit-based feature augmentation approach

MoFO: Momentum-Filtered Optimizer for Mitigating Forgetting in LLM Fine-Tuning

Archilles' Heel in Semi-open LLMs: Hiding Bottom against Recovery Attacks

Towards Explaining the Power of Constant-depth Graph Neural Networks for Structured Linear Programming

In-Context Learning at Representation Level via Unlabeled Texts

OptiBench: Benchmarking Large Language Models in Optimization Modeling with Equivalence-Detection Evaluation

On the Convergence of Adam under Non-uniform Smoothness: Separability from SGDM and Beyond

On Representing Convex Quadratically Constrained Quadratic Programs via Graph Neural Networks

An Efficient Unsupervised Framework for Convex Quadratic Programs via Deep Unrolling