Lune

Profiles2022-2026

Zhouhan Lin

Shanghai Jiao Tong University

31Papers
26Last author
38Unique co-authors
0First author
0Solo

First and last author come from byline position, which some venues assign alphabetically.

Recent papers31

  1. MLP Memory: A Retriever-Pretrained Memory for Large Language ModelsAuthor 7 of 7ICLR 2026 · 16 citations
  2. PonderLM-2: Pretraining LLM with Latent Thoughts in Continuous SpaceAuthor 8 of 8ICML 2026 · 10 citations
  3. FreqKV: Key-Value Compression in Frequency Domain for Context Window ExtensionAuthor 7 of 7ICLR 2026 · 7 citations
  4. Context-level Language Modeling by Learning Predictive Context EmbeddingsAuthor 9 of 9ICML 2026 · 5 citations
  5. AWM: Accurate Weight-Matrix Fingerprint for Large Language ModelsAuthor 5 of 5ICLR 2026 · 3 citations
  6. Flow of Spans: Generalizing Language Models to Dynamic Span-Vocabulary via GFlowNetsAuthor 8 of 8ICLR 2026
  7. PonderLM: Pretraining Language Models to Ponder in Continuous SpaceAuthor 9 of 9ICLR 2026
  8. Memory Decoder: A Pretrained, Plug-and-Play Memory for Large Language ModelsAuthor 7 of 7NeurIPS 2025 · 14 citations
  9. AdaptiveStep: Automatically Dividing Reasoning Step through Model ConfidenceAuthor 13 of 13ICML 2025
  10. Gumbel Reranking: Differentiable End-to-End Reranker OptimizationAuthor 8 of 8ACL 2025
  11. How to Synthesize Text Data without Model Collapse?Author 8 of 10ICML 2025
  12. Training LLMs to be Better Text Embedders through Bidirectional ReconstructionAuthor 8 of 8EMNLP 2025
  13. Training-free LLM-generated Text Detection by Mining Token Probability SequencesAuthor 5 of 7ICLR 2025
  14. Visual-informed Silent Video Identity ConversionAuthor 3 of 3ACM MM 2025
  15. HuRef: HUman-REadable Fingerprint for Large Language ModelsAuthor 8 of 8NeurIPS 2024 · 48 citations
  16. Cluster-wise Graph Transformer with Dual-granularity Kernelized AttentionAuthor 4 of 4NeurIPS 2024 · 15 citations
  17. SH2: Self-Highlighted Hesitation Helps You Decode More TruthfullyAuthor 4 of 4EMNLP Findings 2024 · 5 citations
  18. Graph Parsing NetworksAuthor 5 of 5ICLR 2024 · 4 citations
  19. Enable Fast Sampling for Seq2Seq Text DiffusionAuthor 3 of 3EMNLP Findings 2024 · 2 citations
  20. Leveraging Grammar Induction for Language Understanding and GenerationAuthor 4 of 4EMNLP Findings 2024 · 1 citation
  21. Mirror-Consistency: Harnessing Inconsistency in Majority VotingAuthor 6 of 6EMNLP Findings 2024 · 1 citation
  22. Ordered GNN: Ordering Message Passing to Deal with Heterophily and Over-smoothingAuthor 4 of 4ICLR 2023 · 18 citations
  23. Fourier Transformer: Fast Long Range Modeling by Removing Sequence Redundancy with FFT OperatorAuthor 7 of 7ACL Findings 2023 · 13 citations
  24. Tailoring Self-Attention for Graph via Rooted SubtreesAuthor 4 of 4NeurIPS 2023 · 12 citations
  25. Unsupervised Graph-Text Mutual Conversion with a Unified Pretrained Language ModelAuthor 5 of 7ACL 2023
  26. RASAT: Integrating Relational Structures into Pretrained Seq2Seq Model for Text-to-SQLAuthor 9 of 9EMNLP 2022 · 73 citations
  27. Leveraging Unimodal Self-Supervised Learning for Multimodal Audio-Visual Speech RecognitionAuthor 6 of 6ACL 2022 · 43 citations
  28. Block-Skim: Efficient Question Answering for TransformerAuthor 3 of 6AAAI 2022 · 33 citations
  29. Text Editing as Imitation GameAuthor 6 of 6EMNLP Findings 2022 · 3 citations
  30. Syntax-guided Localized Self-attention by Constituency Syntactic DistanceAuthor 8 of 8EMNLP Findings 2022 · 2 citations

Showing the 30 most recent indexed papers.

Dusk over the sea between two cliffs drawn in fine vertical lines