PPaperPicks

Shiwei Liu

University of Oxford, Mathematical Institute, UK

28 papers at tracked venues · 23 at CORE A* · active 2024–2026

DBLP profile ↗ORCID 0009-0001-1255-4436 ↗Google Scholar ↗Homepage ↗

Venues

ICML ×8ICLR ×6NeurIPS ×6ACL ×2EMNLP ×2InterSpeech ×2AAAI ×1MLSys ×1

Frequent coauthors

Ajay Jaiswal
DBLP profile ↗ORCID search ↗
×2
Pengxiang Li
DBLP profile ↗ORCID search ↗
×2
Yuxin Zhang
DBLP profile ↗ORCID search ↗
×2
Zhenyu Zhang
DBLP profile ↗ORCID search ↗
×2
Lu Yin
DBLP profile ↗ORCID search ↗
×2
Haotian Luo
DBLP profile ↗ORCID search ↗
×1
Di He
DBLP profile ↗ORCID search ↗
×1
Arinbjörn Kolbeinsson
DBLP profile ↗ORCID search ↗
×1
Tianhao Chen
DBLP profile ↗ORCID search ↗
×1
Zihang Liu
DBLP profile ↗ORCID search ↗
×1
Xialie Zhuang
DBLP profile ↗ORCID search ↗
×1
Tianjin Huang
DBLP profile ↗ORCID search ↗
×1

Papers

  1. O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning
    ACL 2026 ·
    Haotian Luo
    DBLP profile ↗ORCID search ↗
  2. AlphaDecay: Module-wise Weight Decay for Heavy-Tailed Balancing in LLMs
    NeurIPS 2025 ·
    Di He
    DBLP profile ↗ORCID search ↗
  3. Composable Interventions for Language Models
    ICLR 2025 ·
    Arinbjörn Kolbeinsson
    DBLP profile ↗ORCID search ↗
  4. From Low Rank Gradient Subspace Stabilization to Low-Rank Weights: Observations, Theories, and Applications
    ICML 2025 ·
    Ajay Kumar Jaiswal
    DBLP profile ↗ORCID search ↗
  5. GPAS: Accelerating Convergence of LLM Pretraining via Gradient-Preserving Activation Scaling
    NeurIPS 2025 ·
    Tianhao Chen
    DBLP profile ↗ORCID search ↗
  6. LIFT the Veil for the Truth: Principal Weights Emerge after Rank Reduction for Reasoning-Focused Supervised Fine-Tuning
    ICML 2025 ·
    Zihang Liu
    DBLP profile ↗ORCID search ↗
  7. Mask-Enhanced Autoregressive Prediction: Pay Less Attention to Learn More
    ICML 2025 ·
    Xialie Zhuang
    DBLP profile ↗ORCID search ↗
  8. Mix-LN: Unleashing the Power of Deeper Layers by Combining Pre-LN and Post-LN
    ICLR 2025 ·
    Pengxiang Li
    DBLP profile ↗ORCID search ↗
  9. Outlier-weighed Layerwise Sampling for LLM Fine-tuning
    ACL 2025 ·
    Pengxiang Li
    DBLP profile ↗ORCID search ↗
  10. SPAM: Spike-Aware Adam with Momentum Reset for Stable LLM Training
    ICLR 2025 ·
    Tianjin Huang
    DBLP profile ↗ORCID search ↗
  11. The Curse of Depth in Large Language Models
    NeurIPS 2025 ·
    Wenfang Sun
    DBLP profile ↗ORCID search ↗
  12. Visual Prompting Upgrades Neural Network Sparsification: A Data-Model Perspective
    AAAI 2025 ·
    Can Jin
    DBLP profile ↗ORCID search ↗
  13. AdaMerging: Adaptive Model Merging for Multi-Task Learning
    ICLR 2024 ·
    Enneng Yang
    DBLP profile ↗ORCID search ↗
  14. Advancing Dynamic Sparse Training by Exploring Optimization Opportunities
    ICML 2024 ·
    Jie Ji
    DBLP profile ↗ORCID search ↗
  15. AlphaPruning: Using Heavy-Tailed Self Regularization Theory for Improved Layer-wise Pruning of Large Language Models
    NeurIPS 2024 ·
    Haiquan Lu
    DBLP profile ↗ORCID search ↗
  16. CaM: Cache Merging for Memory-efficient LLMs Inference
    ICML 2024 ·
    Yuxin Zhang
    DBLP profile ↗ORCID search ↗
  17. Dynamic Data Pruning for Automatic Speech Recognition
    InterSpeech 2024 ·
    Qiao Xiao
    DBLP profile ↗ORCID search ↗
  18. Dynamic Sparse No Training: Training-Free Fine-tuning for Sparse LLMs
    ICLR 2024 ·
    Yuxin Zhang
    DBLP profile ↗ORCID search ↗
  19. E2ENet: Dynamic Sparse Feature Fusion for Accurate and Efficient 3D Medical Image Segmentation
    NeurIPS 2024 ·
    Boqian Wu
    DBLP profile ↗ORCID search ↗
  20. FFN-SkipLLM: A Hidden Gem for Autoregressive Decoding with Adaptive Feed Forward Skipping
    EMNLP 2024 ·
    Ajay Jaiswal
    DBLP profile ↗ORCID search ↗
  21. Found in the Middle: How Language Models Use Long Contexts Better via Plug-and-Play Positional Encoding
    NeurIPS 2024 ·
    Zhenyu Zhang
    DBLP profile ↗ORCID search ↗
  22. Is C4 Dataset Optimal for Pruning? An Investigation of Calibration Data for LLM Pruning
    EMNLP 2024 ·
    Abhinav Bandari
    DBLP profile ↗ORCID search ↗
  23. Junk DNA Hypothesis: Pruning Small Pre-Trained Weights Irreversibly and Monotonically Impairs "Difficult" Downstream Tasks in LLMs
    ICML 2024 ·
    Lu Yin
    DBLP profile ↗ORCID search ↗
  24. MSRS: Training Multimodal Speech Recognition Models from Scratch with Sparse Mask Optimization
    InterSpeech 2024 ·
    Adriana Fernandez-Lopez
    DBLP profile ↗ORCID search ↗
  25. NeurRev: Train Better Sparse Neural Network Practically via Neuron Revitalization
    ICLR 2024 ·
    Gen Li
    DBLP profile ↗ORCID search ↗
  26. Outlier Weighed Layerwise Sparsity (OWL): A Missing Secret Sauce for Pruning LLMs to High Sparsity
    ICML 2024 ·
    Lu Yin
    DBLP profile ↗ORCID search ↗
  27. Q-Hitter: A Better Token Oracle for Efficient LLM Inference via Sparse-Quantized KV Cache
    MLSys 2024 ·
    Zhenyu Zhang
    DBLP profile ↗ORCID search ↗
  28. Sparse Cocktail: Every Sparse Pattern Every Sparse Ratio All At Once
    ICML 2024 ·
    Zhangheng Li
    DBLP profile ↗ORCID search ↗