 |
Siyuan Li (李思远) |
News
- [2024.08] One paper on GNN theory, GNN Bottleneck, is accepted by TKDE 2024, congrats to Fang Wu.
- [2024.07] Co-authored paper on CoT reasoning, MC-CoT, is accepted by ECCV 2024, congrats to Cheng Tan.
- [2024.06] Co-authored paper on Neuro-science, NeuroBERT, is accepted by JBHI 2024, congrats to Di Wu.
- [2024.05] One paper on genomics, VQDNA, is accepted by ICML 2024.
- [2024.05] Three co-authored paper on Long-sequence (CHELA) and AI4Science are accepted by ICML 2024, congrats to Zicheng Liu, Lirong Wu, and Yufei Huang.
- [2024.04] Co-authored paper on Long-sequence model, LongVQ, is accepted by IJCAI 2024, congrats to Zicheng Liu.
- [2024.01] One paper on semi-supervised learning, SemiReward, is accepted by ICLR 2024.
- [2024.01] One paper on vision architecture, MogaNet, is accepted by ICLR 2024.
- [2024.01] Three co-authored paper on AI4Science are accepted by ICLR 2024 (MAPE-PPI wins Spotlight), congrats to Lirong Wu, Zhangyang Gao, and Cheng Tan.
- [2023.12] Two co-authored paper on protein and video are accepted by AAAI 2024, congrats to Yufei Huang and Xuesong Nie.
- [2023.10] One paper on self-supervised learning, GenURL, is accepted by TNNLS 2023.
- [2023.09] One paper on video prediction, OpenSTL, is accepted by NeurIPS 2023.
- [2023.09] One paper on data augmentation, Decoupled Mixup, is accepted by NeurIPS 2023.
- [2023.09] Two co-authored paper on AI4Science are accepted by NeurIPS 2023, congrats to Haitao Lin and Jun Xia.
- [2023.06] Build an open-source toolbox, OpenSTL, for spatio-temporal predictive learning based on PyTorch and update on arXiv (OpenSTL). On updating!
- [2023.06] One paper on label noise, GNN-Cleaner, is accepted by TKDE 2023, congrats to Jun Xia.
- [2023.04] One paper on self-supervised learning, A2MIM, is accepted by ICML 2023.
- [2023.04] Co-authored paper on GNN explanation, MatchExplainer, is accepted by ICML 2023, congrats to Fang Wu.
- [2023.03] Co-authored paper on sign language, CVT-SLR, is accepted by CVPR 2023 (Highlight), congrats to Zheng Jiangbin.
- [2023.03] Co-authored paper on video prediction, TAU, is accepted by CVPR 2023, congrats to Cheng Tan.
- [2023.01] Co-authored paper on molecular graph pre-training, Mole-BERT, is accepted by ICLR 2023, congrats to Jun Xia.
- [2022.09] One paper on fine-grained self-supervised learning, CVSA, is accepted by BMVC 2022 (Spotlight).
- [2022.09] Co-authored paper on domain adaptation, DaC, is accepted by NeurIPS 2022, congrats to Ziyi Zhang.
- [2022.09] Co-authored paper on graph representation, DMAGE, is accepted by Neurocomputing 2022, congrats to Zelin Zang.
- [2022.08] Co-authored paper on graph attack, AtkSE, is accepted by CIKM 2022, congrats to Zihan Liu.
- [2022.07] One paper on data augmentation, AutoMix, is accepted by ECCV 2022 (Oral).
- [2022.07] One paper on self-supervised learning, DLME, is accepted by ECCV 2022.
- [2022.01] Build an open-source toolbox, OpenBioSeq, for supervised and self-supervised bio-sequence representation learning based on PyTorch. On updating!
- [2022.03] Co-authored paper on semi-supervised learning, HCR, is accepted by CVPR 2022, congrats to Cheng Tan.
- [2022.01] Build an open-source toolbox, OpenMixup, for supervised, self- and semi-supervised visual representation learning based on PyTorch. On updating!
- [2021.10] Co-authored paper on deep clustering, GCML, is accepted by WACV 2022, congrats to Lirong Wu.
- [2021.07] Got my B.S. degree from Nanjing University!
- [2021.06] One paper on invertible learning, inv-ML, is accepted by ECML-PKDD 2021.
- [2020.05] One paper on visual object tracking, TLPG-Tracker, is accepted by IJCAI 2020.
Research Interest
Currently, I focus on Self-supervised Learning and Data-efficient Deep Learning:- Self-supervised Learning in Computer Vision and Bioengineering
- Network Architecture and Long-sequence Modeling
- General Data Augmentation
- Semi-supervised Learning
Education
- 2021.09-2026.06 Ph.D at Zhejiang University & Westlake University. Supervisor: Prof. Stan Z. Li
- 2017.09-2021.06 B.Sc. at Nanjing University. Supervisor: Prof. Jianxing Wu
Internship
- 2023.07-2024.08 Research on AIGC at Alibaba DAMO Academy, supervised by Researcher Baigui Sun.
- 2020.06-2021.08 Research on Dimension Reduction at Westlake University, supervised by Prof. Stan Z. Li.
Publications
Selected Preprints:
|
|
Unveiling the Backbone-Optimizer Coupling Bias in Visual Representation Learning |
|
|
Switch EMA: A Free Lunch for Better Flatness and Sharpness |
Conferences (As First Author):
|
|
VQDNA: Unleashing the Power of Vector Quantization for Multi-Species Genomic Sequence Modeling |
 |
SemiReward: A General Reward Model for Semi-supervised Learning |
 |
MogaNet: Efficient Multi-order Gated Aggregation Network |
 |
OpenSTL: A Comprehensive Benchmark of Spatio-Temporal Predictive Learning |
 |
Harnessing Hard Mixed Samples with Decoupled Regularizer |
 |
Architecture-Agnostic Masked Image Modeling -- From ViT back to CNN |
 |
Exploring Localization for Self-supervised Fine-grained Contrastive Learning |
 |
AutoMix: Unveiling the Power of Mixup for Stronger Classifiers |
 |
DLME: Deep Local-flatness Manifold Embedding |
 |
Invertible Manifold Learning for Dimension Reduction |
 |
TLPG-Tracker: Joint Learning of Target Localization and Proposal Generation for Visual Tracking |
Conferences (As Co-author):
|
|
Boosting the Power of Small Multimodal Reasoning Models to Match Larger Models with Self-Consistency Training |
|
|
Short-Long Convolutions Help Hardware-Efficient Linear Attention to Focus on Long Sequences |
|
|
Re-Dock: Towards Flexible and Realistic Molecular Docking with Diffusion Bridge |
|
|
Learning to Predict Mutation Effects of Protein-Protein Interactions by Microenvironment-aware Hierarchical Prompt Learning |
|
|
LongVQ: Long Sequence Modeling with Vector Quantization on Structured Memory |
|
|
MAPE-PPI: Towards Effective and Efficient Protein-Protein Interaction Prediction via Microenvironment-Aware Protein Embedding |
|
|
Hierarchical Data-efficient Representation Learning for Tertiary Structure-based RNA Design |
|
|
KW-Design: Pushing the Limit of Protein Deign via Knowledge Refinement |
|
|
Protein 3D Graph Structure Learning for Robust Structure-based Protein Property Prediction |
|
|
Wavelet-Driven Spatiotemporal Predictive Learning: Bridging Frequency and Time Variations |
 |
Functional-Group-Based Diffusion for Pocket-Specific Molecule Generation and Elaboration |
 |
Explaining Graph Neural Networks via Non-parametric Subgraph Matching |
 |
CVT-SLR: Contrastive Visual-Textual Transformation for Sign Language Recognition with Variational Alignment |
 |
Temporal attention unit: Towards efficient spatiotemporal predictive learning |
 |
Mole-BERT: Rethinking Pre-training Graph Neural Networks for Molecules |
 |
Divide and Contrast: Source-free Domain Adaptation via Adaptive Contrastive Learning |
 |
Are Gradients on Graph Structure Reliable in Gray-box Attacks? |
 |
Hyperspherical Consistency Regularization |
 |
Generalized Clustering and Multi-Manifold Learning with Geometric Structure Preservation |
Journals:
 |
GenURL: A General Framework for Unsupervised Representation Learning |
|
|
Discovering the Representation Bottleneck of Graph Neural Networks |
|
|
Neuro-BERT: Rethinking Masked Autoencoding for Self-Supervised Neurological Pretraining |
 |
GNN Cleaner: Label Cleaner for Graph-structured Data |
 |
Deep Manifold Attributed Graph Embedding |
Services
Membership:
- IEEE Membership, Graduate Student Member, 2022.01-present
- IEEE Young Professionals, 2022.01-present
- China Society of Image and Graphics (CSIG), Student Member, 2023-2026
- China Computer Federation (CCF), Student Member, 2024-2026
- British Machine Vision Association (BMVA), Student Member, 2024-present
Invited Talk:
- 2024/03/27: Talk on "Modern Convolutional Neural Networks" @ Chengdu Institute of Computer Application, Chinese Academy of Sciences [PPT]
- 2024/03/12: Talk on "Convolution Kernel Design and Gated Attention for Modern Convolutional Neural Networks" @ ShuZiHuanYu Platform [Video] [PPT]
- 2023/12/14: Talk on "Mixup Data Augmentation for Computer Vision" @ Chongqing Technology and Business University [PPT]
- 2023/12/14: Talk on "Introduction to AI Research and Experience Sharing" @ Chongqing Technology and Business University
Program Committee Member | Reviewer:
- Conference Reviewer / PC Member: International Conference on Learning Representations (ICLR), 2024
- Journal Reviewer: IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
International Conference on Machine Learning (ICML), 2022-2024
Conference and Workshop on Neural Information Processing Systems (NeurIPS), 2022-2024
Dataset and Benchmark Track on Neural Information Processing Systems (NeurIPS DB Track), 2023-2024
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022-2024
International Conference on Computer Vision (ICCV), 2023
Conference on European Conference on Computer Vision (ECCV), 2022-2024
AAAI Conference on Artificial Intelligence (AAAI), 2022-2025
International Joint Conference on Artificial Intelligence (IJCAI), 2023
ACM Multimedia (ACMMM), 2022
IEEE Conference on Multimedia Expo (ICME), 2024
British Machine Vision Conference (BMVC), 2024
Asian Conference on Computer Vision (ACCV), 2024
IEEE Winter Conference on Applications of Computer Vision (WACV), 2022-2025
International Journal of Computer Vision (IJCV)
IEEE Transactions on Neural Networks and Learning Systems (TNNLS)
IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)
Neural Computing & Applications (NCA)
Electronic Research Archive (ERA)
Teaching Assistant:
- Deep Learning Course, Westlake University, 2024.03-2024.06
Online Project:
- OpenMixup: Open mixup toolbox and benchmark for visual representation learning [Code]
- OpenSTL: Open-source project for video prediction benchmarks [Code]
- Awesome-MIM: Masked Modeling for Self-supervised Representation Learning on Vision and Beyond [Code]
- Awesome-Mixup: Awesome List of Mixup Augmentation and Beyond [Code]
Social Awards:
- 2024: ICML 2024 Scholarship (Financial Aid Awardees)
- 2024: ICLR 2024 Scholarship (Finantial Assistance Awardees)
- 2023: ICML 2023 Participation Grant