Publications

publications by categories in reversed chronological order. * denotes equal contribution.

preprint

  1. Baichuan-Omni Technical Report
    Li, Yadong, Sun, Haoze, Lin, Mingan, Li, Tianpeng, Dong, Guosheng, Zhang, Tao, Ding, Bowen, Song, Wei, Cheng, Zhenglin, Huo, Yuqi, Chen, Song, Lu, Xu, Pan, Da, Zhang, Shusen, Wu, Xin, Liang, Zheng, Liu, Jun, Zhang, Tao, Lu, Keer, Zhao, Yaqi, Shen, Yanjun, Yang, Fan, Yu, Kaicheng, Lin, Tao, Xu, Jianhua, Zhou, Zenan, and Chen, Weipeng
    (arXiv:2410.08565) preprint
  2. Co-Evolution Learning
    Long, Zhenwei*, Sun, Peng*, Cheng, Zhenglin, Liu, Deyuan, and Lin, Tao
    preprint
  3. Dynamic Mixture of Experts: An Auto-tuning Approach for Efficient Transformer Models
    Guo, Yongxin*, Cheng, Zhenglin*, Tang, Xiaoying, and Lin, Tao
    (arXiv:2405.14297) preprint
    PDF Code
  4. DeFT: Decoding with Flash Tree-Attention for Efficient Tree-structured LLM Inference
    Yao, Jinwei*, Chen, Kaiqi*, Zhang, Kexun, You, Jiaxuan, Yuan, Binhang, Wang, Zeke, and Lin, Tao
    (arXiv:2404.00242), abridged in ICLR workshop AGI (Oral). preprint
  5. ELICIT: LLM Augmentation Via External In-context Capability
    Wang, Futing*, Yan, Jianhao*, Zhang, Yue, and Lin, Tao
    (arXiv:2410.09343) preprint
  6. PathGen-1.6M: 1.6 Million Pathology Image-text Pairs Generation through Multi-agent Collaboration
    Sun, Yuxuan, Zhang, Yunlong, Si, Yixuan, Zhu, Chenglu, Shui, Zhongyi, Zhang, Kai, Li, Jingxiong, Xinheng, Lyu, Lin, Tao†, and Yang, Lin†
    (arXiv:2407.00203) preprint
  7. Cognitive Insights and Stable Coalition Matching for Fostering Multi-Agent Cooperation
    Shao, Jiaqi, Yuan, Tianjun, Lin, Tao, Cao, Xuanyu, and Luo, Bing
    (arXiv:2405.18044) preprint
  8. MorphAgent: Empowering Agents through Self-Evolving Profiles and Decentralized Collaboration
    Lu, Siyuan*, Shao, Jiaqi*, Luo, Bing, and Lin, Tao
    (2410.15048) preprint
  9. GIFT: Unlocking Full Potential of Labels in Distilled Dataset at Near-zero Cost
    Shang, Xinyi*, Sun, Peng*, and Lin, Tao
    (arXiv:2405.14736) preprint
    PDF Code
  10. Information Compensation: A Fix for Any-scale Dataset Distillation
    Sun, Peng, Shi, Bei, Shang, Xinyi, and Lin, Tao
    ICLR workshop DMLR preprint
  11. Privacy as a Free Lunch: Crafting Initial Distilled Datasets through the Kaleidoscope
    Shi, Shuo, Sun, Peng, Shang, Xinyi, Du, Tianyu, Zhang, Xuhong, Yin, Jianwei, and Lin, Tao
    preprint
  12. Client2Vec: Improving Federated Learning by Distribution Shifts Aware Client Indexing
    Guo, Yongxin, Wang, Lin, Tang, Xiaoying, and Lin, Tao
    (arXiv:2405.16233) preprint
  13. Find Your Optimal Assignments On-the-fly: A Holistic Framework for Clustered Federated Learning
    Guo, Yongxin, Tang, Xiaoying, and Lin, Tao
    (arXiv:2310.05397) preprint
    PDF Code
  14. Federated Unlearning: a Perspective of Stability and Fairness
    Shao, Jiaqi, Lin, Tao, Cao, Xuanyu, and Luo, Bing
    ICLR workshop PML preprint
    PDF
  15. Improving Group Connectivity for Generalization of Federated Deep Learning
    Li, Zexi*, Lin, Jie*, Li, Zhiqi*, Zhu, Didi, Ye, Rui, Shen, Tao, Lin, Tao†, and Wu, Chao†
    (arXiv:2402.18949), abridged in FL@FM-NeurIPS’24 (Oral). preprint
  16. CollabEdit: Towards Non-destructive Collaborative Knowledge Editing
    Zheng, Jiamu, Zhang, Jinghuai, Du, Tianyu, Zhang, Xuhong, Yin, Jianwei, and Lin, Tao
    (arXiv:2410.09508), ICLR 2024 workshop SeT LLM and ICLR 2024 workshop DPFM. preprint
  17. Switch EMA: A Free Lunch for Better Flatness and Sharpness
    Li, Siyuan, Liu, Zicheng, Juanxi, Tian, Wang, Ge, Wang, Zedong, Jin, Weiyang, Wu, Di, Tan, Cheng, Lin, Tao, Liu, Yang, Sun, Baigui, and Li, Stan Z.
    (arXiv:2402.09240) preprint

2024

  1. NeurIPS 2024
    Efficiency for Free: Ideal Data Are Transportable Representations
    Sun, Peng, Jiang, Yi, and Lin, Tao
    In Advances in Neural Information Processing Systems (NeurIPS) 2024
    PDF Code
  2. NeurIPS 2024
    Cooperative Hardware-Prompt Learning for Snapshot Compressive Imaging
    Wang, Jiamian, Wu, Zongliang, Zhang, Yulun, Yuan, Xin, Lin, Tao, and Tao, Zhiqiang
    In Advances in Neural Information Processing Systems (NeurIPS) 2024
    PDF
  3. ECCV 2024 (Oral)
    PathMMU: A Massive Multimodal Expert-Level Benchmark for Understanding and Reasoning in Pathology
    Sun, Yuxuan, Wu, Hao, Zhu, Chenglu, Zheng, Sunyi, Chen, Qizi, Zhang, Kai, Zhang, Yunlong, Wan, Dan, Lan, Xiaoxiao, Zheng, Mengyue, Li, Jingxiong, Lyu, Xinheng, Lin, Tao†, and Yang, Lin†
    In European Conference on Computer Vision (ECCV), Best Paper Candidate @ ECCV 2024
    PDF Code Website
  4. ICML 2024
    FedRC: Tackling Diverse Distribution Shifts Challenge in Federated Learning by Robust Clustering
    Guo, Yongxin, Tang, Xiaoying, and Lin, Tao
    In International Conference on Machine Learning (ICML), abridged in ICLR Workshop ML4IoT 2024
    PDF Code
  5. CVPR 2024
    On the Diversity and Realism of Distilled Dataset: An Efficient Dataset Distillation Paradigm
    Sun, Peng, Shi, Bei, Yu, Daiwei, and Lin, Tao
    In Computer Vision and Pattern Recognition Conference (CVPR) 2024
    PDF Code
  6. ICLR 2024
    Towards Robust Multi-Modal Reasoning via Model Selection
    Liu, Xiangyan*, Li, Rongxue*, Ji, Wei, and Lin, Tao
    In International Conference on Learning Representations (ICLR) 2024
    PDF Code
  7. ICLR 2024
    Increasing Model Capacity for Free: A Simple Strategy for Parameter Efficient Fine-tuning
    Song, Haobo*, Zhao, Hao*, Majumder, Soumajit, and Lin, Tao
    In International Conference on Learning Representations (ICLR) 2024
    PDF Code
  8. Journal
    Decentralized Gradient Tracking with Local Steps
    Liu, Yue, Lin, Tao, Koloskova, Anastasia, and Stich, Sebastian U.
    Optimization Methods and Software 2024
    PDF

2023

  1. NeurIPS 2023
    DELTA: Diverse Client Sampling for Fasting Federated Learning
    Wang, Lin, Guo, Yongxin, Lin, Tao, and Tang, Xiaoying
    In Advances in Neural Information Processing Systems (NeurIPS) 2023
    PDF
  2. ICCV 2023
    No Fear of Classifier Biases: Neural Collapse Inspired Federated Learning with Synthetic and Fixed Classifier
    Li, Zexi, Shang, Xinyi, Rui, He, Lin, Tao†, and Wu, Chao†
    In International Conference on Computer Vision (ICCV) 2023
    PDF Code
  3. ICML 2023
    On Pitfalls of Test-time Adaptation
    Zhao, Hao*, Liu, Yuejiang*, Alahi, Alexandre, and Lin, Tao
    In International Conference on Machine Learning (ICML), abridged in ICLR Workshop on Trustworthy ML 2023 and ICLR Workshop on DG (spotlight) 2023
    PDF Code
  4. ICML 2023
    FedBR: Improving Federated Learning on Heterogeneous Data via Local Learning Bias Reduction
    Guo, Yongxin, Tang, Xiaoying, and Lin, Tao
    In International Conference on Machine Learning (ICML) 2023
    PDF Code
  5. ICML 2023
    Revisiting Weighted Aggregation in Federated Learning with Neural Networks
    Li, Zexi, Lin, Tao†, Shang, Xinyi, and Wu, Chao†
    In International Conference on Machine Learning (ICML) 2023
    PDF Code
  6. ICLR 2023
    Test-Time Robust Personalization for Federated Learning
    Jiang, Liangze*, and Lin, Tao*†
    In International Conference on Learning Representations (ICLR) 2023
    PDF Code

2022

  1. IMWUT 2022
    Learning Disentangled Behaviour Patterns for Wearable-based Human Activity Recognition
    Su, Jie, Wen, Zhenyu, Lin, Tao, and Guan, Yu
    Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT) 2022
    PDF Code
  2. OPT 2022
    Decentralized Stochastic Optimization with Client Sampling
    Liu, Ziwei, Koloskova, Anastasia, Jaggi, Martin, and Lin, Tao
    Optimization and machine learning (OPT) workshop, in NeurIPS 2022
    PDF

2021

  1. OPT 2021
    Understanding Memorization from the Perspective of Optimization via Efficient Influence Estimation
    Liu, Futong, Lin, Tao†, and Jaggi, Martin
    Optimization and machine learning (OPT) workshop, in NeurIPS 2021
    PDF
  2. NeurIPS 2021
    RelaySum for Decentralized Deep Learning on Heterogeneous Data
    Vogels, Thijs*, He, Lie*, Koloskova, Anastasia, Lin, Tao, Karimireddy, Sai Praneeth, Stich, Sebastian U., and Jaggi, Martin
    In Advances in Neural Information Processing Systems (NeurIPS) 2021
    PDF Code Video Slides
  3. NeurIPS 2021
    An Improved Analysis of Gradient Tracking for Decentralized Machine Learning
    Koloskova, Anastasia, Lin, Tao, and Stich, Sebastian U.
    In Advances in Neural Information Processing Systems (NeurIPS) 2021
    PDF
  4. ICML 2021
    Quasi-global Momentum: Accelerating Decentralized Deep Learning on Heterogeneous Data
    Lin, Tao, Karimireddy, Sai Praneeth, Stich, Sebastian U., and Jaggi, Martin
    In International Conference on Machine Learning (ICML) 2021
    PDF Code
  5. ICML 2021
    Consensus Control for Decentralized Deep Learning
    Kong, Lingjing*, Lin, Tao*†, Koloskova, Anastasia, Jaggi, Martin, and Stich, Sebastian U.
    In International Conference on Machine Learning (ICML) 2021
    PDF
  6. preprint
    Towards Federated Learning on Time-Evolving Heterogeneous Data
    Guo, Yongxin*, Lin, Tao*, and Tang, Xiaoying
    preprint (arXiv:2112.13246), appeared in FL-ICML workshop 2021
    PDF

2020

  1. NeurIPS 2020
    Ensemble Distillation for Robust Model Fusion in Federated Learning
    Lin, Tao*, Kong, Lingjing*, Stich, Sebastian U, and Jaggi, Martin
    In Advances in Neural Information Processing Systems (NeurIPS) 2020
    PDF Code
  2. NeurIPS 2020
    On the Loss Landscape of Adversarial Training: Identifying Challenges and How to Overcome Them
    Liu, Chen, Salzmann, Mathieu, Lin, Tao, Tomioka, Ryota, and Süsstrunk, Sabine
    In Advances in Neural Information Processing Systems (NeurIPS) 2020
    PDF Code
  3. ICML 2020
    Extrapolation for Large-batch Training in Deep Learning
    Lin, Tao*, Kong, Lingjing*, Stich, Sebastian U., and Jaggi, Martin
    In International Conference on Machine Learning (ICML) 2020
    PDF
  4. EMNLP 2020
    Masking as an Efficient Alternative to Finetuning for Pretrained Language Models
    Zhao, Mengjie*, Lin, Tao*, Jaggi, Martin, and Schütze, Hinrich
    In Empirical Methods in Natural Language Processing (EMNLP) 2020
    PDF Code
  5. ICLR 2020
    Decentralized Deep Learning with Arbitrary Communication Compression
    Koloskova, Anastasia*, Lin, Tao*, Stich, Sebastian U, and Jaggi, Martin
    In International Conference on Learning Representations (ICLR) 2020
    PDF Code
  6. ICLR 2020
    Don’t Use Large Mini-batches, Use Local SGD
    Lin, Tao, Stich, Sebastian U, Patel, Kumar Kshitij, and Jaggi, Martin
    In International Conference on Learning Representations (ICLR) 2020
    PDF Code
  7. ICLR 2020
    Dynamic Model Pruning with Feedback
    Lin, Tao, Stich, Sebastian U, Barba, Luis, Dmitriev, Daniil, and Jaggi, Martin
    In International Conference on Learning Representations (ICLR) 2020
    PDF Code
  8. CLVision 2020
    Generalized Class Incremental Learning
    Mi, Fei*, Kong, Lingjing*, Lin, Tao, Yu, Kaicheng, and Faltings, Boi
    In Conference on Computer Vision and Pattern Recognition (CVPR) Workshop on Continual Learning 2020
    PDF

2019

  1. ICML 2019
    Exploring Interpretable LSTM Neural Networks over Multi-variable Data
    Guo, Tian, Lin, Tao, and Antulov-Fantulin, Nino
    In International Conference on Machine Learning (ICML) 2019
    PDF Code
  2. TPDS 2019
    Ga-par: Dependable Microservice Orchestration Framework for Geo-distributed Clouds
    Wen, Zhenyu, Lin, Tao, Yang, Renyu, Ji, Shouling, Ranjan, Rajiv, Romanovsky, Alexander, Lin, Changting, and Xu, Jie
    IEEE Transactions on Parallel and Distributed Systems (TPDS) 2019
    PDF

2018

  1. NeurIPS 2018
    Training DNNs with Hybrid Block Floating Point
    Drumond, Mario, Lin, Tao, Jaggi, Martin, and Falsafi, Babak
    In Advances in Neural Information Processing Systems (NeurIPS) 2018
    PDF Code

2017

  1. IJCAI 2017
    Hybrid Neural Networks for Learning the Trend in Time Series
    Lin, Tao*, Guo, Tian*, and Aberer, Karl
    In Proceedings of the twenty-sixth international joint conference on artificial intelligence (IJCAI) 2017
    PDF
  2. IEEE 2017
    Fog Orchestration for IoT Services: Issues, Challenges and Directions
    Wen, Zhenyu, Yang, Renyu, Garraghan, Peter, Lin, Tao, Xu, Jie, and Rovatsos, Michael
    IEEE Internet Computing 2017
    PDF
  3. VLDB 2017
    Heterogeneous Recommendations: What You Might Like to Read after Watching Interstellar
    Guerraoui, Rachid, Kermarrec, Anne-Marie, Lin, Tao, and Patra, Rhicheek
    Proceedings of the Very Large Data Base Endowment (VLDB) 2017
    PDF Code