学者档案2020-2026
Jie Wang
MoE Key Laboratory of Brain-inspired Intelligent Perception and Cognition, University of Science and Technology of China
67论文数
13末位作者
74合作者人数
1一作
0独著
一作与末位作者按论文署名顺序统计,部分顶会采用姓氏字母顺序署名。
近期论文67
- Why Attention Patterns Exist: A Unifying Temporal Perspective Analysis第 2 位,共 9 位作者ICLR 2026 · 5 次引用
- Opt-Verifier: Unleashing the Power of LLMs for Optimization Modeling via Dual-Side Verification第 2 位,共 11 位作者ICML 2026 · 4 次引用
- Short Chains, Deep Thoughts: Balancing Reasoning Efficiency and Intra-Segment Capability via Split-Merge Optimization第 2 位,共 6 位作者ICML 2026 · 1 次引用
- A Hierarchical Circuit Symbolic Discovery Framework for Efficient Logic Optimization第 2 位,共 7 位作者ICLR 2026
- Adversarial Latent Embedding Repair for LLM Continual Learning第 3 位,共 11 位作者ICML 2026
- Boosting Multi-Domain Reasoning of LLMs via Curvature-Guided Policy Optimization第 3 位,共 9 位作者ICLR 2026
- D-ARL: A Distribution-Matched Asynchronous Reinforcement Learning Framework for Language Reasoning第 3 位,共 10 位作者ICML 2026
- Evolving Graph Structured Programs for Circuit Generation with Large Language Models第 2 位,共 9 位作者ICLR 2026
- Following the Navigation: Enhancing Small Language Models Contextual Reasoning with LLM Guidance第 2 位,共 7 位作者ICLR 2026
- Latent-Guided Reasoning: Empowering Small LLMs with Large-Model Thinking第 3 位,共 7 位作者ICLR 2026
- Opt-Miner: Empowering Information-Seeking Agent with Tree-Guided Data Synthesis for Optimization Modeling第 3 位,共 9 位作者ICML 2026
- Tackling Heavy-Tailed Q-Value Bias in Offline-to-Online Reinforcement Learning with Laplace-Robust Modeling第 6 位,共 7 位作者ICLR 2026
- OptiTree: Hierarchical Thoughts Generation with Tree Search for LLM Optimization Modeling第 2 位,共 6 位作者NeurIPS 2025 · 33 次引用
- Mixture-of-Experts Operator Transformer for Large-Scale PDE Pre-Training第 3 位,共 7 位作者NeurIPS 2025 · 15 次引用
- RoME: Domain-Robust Mixture-of-Experts for MILP Solution Prediction across Domains第 5 位,共 8 位作者NeurIPS 2025 · 11 次引用
- AttentionPredictor: Temporal Patterns Matter for KV Cache Compression第 2 位,共 11 位作者NeurIPS 2025 · 8 次引用
- Accurate KV Cache Eviction via Anchor Direction Projection for Efficient LLM Inference第 2 位,共 12 位作者NeurIPS 2025 · 6 次引用
- SymMaP: Improving Computational Efficiency in Linear Solvers through Symbolic Preconditioning第 2 位,共 5 位作者NeurIPS 2025 · 6 次引用
- LogicTree: Improving Complex Reasoning of LLMs via Instantiated Multi-step Synthetic Logical Data第 3 位,共 10 位作者NeurIPS 2025 · 5 次引用
- STNet: Spectral Transformation Network for Solving Operator Eigenvalue Problem第 3 位,共 6 位作者NeurIPS 2025 · 4 次引用
- High-Performance Arithmetic Circuit Optimization via Differentiable Architecture Search第 2 位,共 7 位作者NeurIPS 2025 · 3 次引用
- Dynamic Configuration for Cutting Plane Separators via Reinforcement Learning on Incremental Graph第 2 位,共 9 位作者NeurIPS 2025 · 1 次引用
- A Graph Enhanced Symbolic Discovery Framework For Efficient Logic Optimization第 2 位,共 8 位作者ICLR 2025
- Accelerating Large Language Model Reasoning via Speculative Search第 2 位,共 8 位作者ICML 2025
- Accurate and Scalable Graph Neural Networks via Message Invariance第 2 位,共 6 位作者ICLR 2025
- Apollo-MILP: An Alternating Prediction-Correction Neural Solving Framework for Mixed-Integer Linear Programming第 2 位,共 8 位作者ICLR 2025
- Boosting Multi-Domain Fine-Tuning of Large Language Models through Evolving Interactions between Samples第 3 位,共 7 位作者ICML 2025
- Computing Circuits Optimization via Model-Based Circuit Genetic Evolution第 2 位,共 9 位作者ICLR 2025
- HyperTree Planning: Enhancing LLM Reasoning via Hierarchical Thinking第 3 位,共 10 位作者ICML 2025
- Knowledge Graph Finetuning Enhances Knowledge Manipulation in Large Language Models第 3 位,共 9 位作者ICLR 2025
展示最近收录的 30 篇论文。
