Profiles2025-2026
Linfeng Zhang
Shanghai Jiao Tong University
30Papers
28Last author
17Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers30
- dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive CachingAuthor 9 of 9ICML 2026 · 156 citations
- The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMsAuthor 16 of 16ICLR 2026 · 32 citations
- Prune2Drive: A Plug-and-Play Framework for Accelerating Vision-Language Models in Autonomous DrivingAuthor 12 of 12CVPR 2026 · 17 citations
- Mixing Importance with Diversity: Joint Optimization for KV Cache Compression in Large Vision-Language ModelsAuthor 4 of 4ICLR 2026 · 16 citations
- OPUS: Towards Efficient and Principled Data Selection in Large Language Model Pre-training in Every IterationAuthor 12 of 12ICML 2026 · 12 citations
- HiCache: A Plug-in Scaled-Hermite Upgrade for Taylor-Style Cache-then-Forecast Diffusion AccelerationAuthor 11 of 11ICLR 2026 · 7 citations
- Grounding and Enhancing Informativeness and Utility in Dataset DistillationAuthor 8 of 8ICLR 2026 · 3 citations
- Accelerating Diffusion-based Video Editing via Heterogeneous Caching: Beyond Full Computing at Sampled Denoising TimestepAuthor 3 of 8CVPR 2026 · 2 citations
- AgentSlimming: Towards Efficient and Cost-Aware Multi-Agent SystemsAuthor 6 of 6ACL 2026 · 1 citation
- Beyond Fixed Formulas: Data-Driven Linear Predictor for Efficient Diffusion ModelsAuthor 9 of 9CVPR 2026 · 1 citation
- D²Pruner: Debiased Importance and Structural Diversity for MLLM Token PruningAuthor 8 of 8AAAI 2026 · 1 citation
- Bypassing Neural Evaluations for Fast Audio Editing via Adaptive Trajectory ExtrapolationAuthor 10 of 11ACL Findings 2026
- Flash-Unified: A Training-Free and Task-Aware Acceleration Framework for Native Unified ModelsAuthor 9 of 9CVPR Findings 2026
- Socratic-Geo: Synthetic Data Generation and Cross-Modal Geometric Reasoning via Multi-Agent InteractionAuthor 7 of 7CVPR 2026
- StreamMeCo: Long-Term Agent Memory Compression for Efficient Streaming Video UnderstandingAuthor 9 of 9ACL Findings 2026
- Training-Free and Hardware-Friendly Acceleration for Diffusion Models via Similarity-based Token PruningAuthor 4 of 4AAAI 2025 · 36 citations
- Efficient Multi-modal Large Language Models via Progressive Consistency DistillationAuthor 11 of 11NeurIPS 2025 · 26 citations
- From Reusing to Forecasting: Accelerating Diffusion Models With TaylorseersAuthor 5 of 5ICCV 2025 · 12 citations
- Token Pruning in Multimodal Large Language Models: Are We Solving the Right Problem?Author 5 of 5ACL Findings 2025 · 12 citations
- Compute Only 16 Tokens in One Timestep: Accelerating Diffusion Transformers with Cluster-Driven Feature CachingAuthor 5 of 5ACM MM 2025 · 4 citations
- Stop Looking for "Important Tokens" in Multimodal Language Models: Duplication Matters MoreAuthor 8 of 8EMNLP 2025 · 4 citations
- SpeCa: Accelerating Diffusion Transformers with Speculative Feature CachingAuthor 7 of 7ACM MM 2025 · 3 citations
- Video Compression Commander: Plug-and-Play Inference Acceleration for Video Large Language ModelsAuthor 4 of 4EMNLP 2025 · 3 citations
- Accelerating Diffusion Transformers with Token-wise Feature CachingAuthor 5 of 5ICLR 2025
- Data Whisperer: Efficient Data Selection for Task-Specific LLM Fine-Tuning via Few-Shot In-Context LearningAuthor 11 of 11ACL 2025
- Dataset Distillation with Neural Characteristic Function: A Minmax PerspectiveAuthor 7 of 7CVPR 2025
- Decouple-Then-Merge: Finetune Diffusion Models as Multi-Task LearningAuthor 5 of 5CVPR 2025
- Gnothi Seauton: Empowering Faithful Self-Interpretability in Black-Box TransformersAuthor 8 of 8ICLR 2025
- GraphKV: Breaking the Static Selection Paradigm with Graph-Based KV Cache EvictionAuthor 3 of 3EMNLP 2025
- ProReflow: Progressive Reflow with Decomposed VelocityAuthor 10 of 10CVPR 2025
