Xiaolin Hu
Professor
Works on adversarial attacks and defenses for deep learning and brain-inspired computational models.
Xiaolin Hu's lab at Tsinghua works on trustworthy & safe ML and computational neuroscience & cognition. 59 papers since 2023, 9 at top venues, mostly TPAMI and ACM MM. Applications close .
At a glance
- papers, 2023–2026
- 15/ year
- at T1 venues
- 15%
- PhD students
- —
- MS students
- —
- alumni listed
- —
Papers by yearfrom the lab homepage, 2023–2026
- 14 papers2026 · T1 1 · collecting
- 15 papers2025 · T1 1
- 17 papers2024 · T1 2
- 13 papers2023 · T1 5
T1 venuesOtherVenues: TPAMI, ACM MM, TASLP, CoRL, Interspeech
Research focuspaper titles that use each topic's words
- Trustworthy & safe ML3 papers
- attacks3 papers
- Computational neuroscience & cognition0 papers
- defenses0 papers
- computational models0 papers
Paper list
From the lab homepage.
202614 papers
- The International Conference on Machine Learning (ICML), Seoul, South Korea, July 6-11A semantically consistent dataset for data-efficient query-based universal sound separation (opens in a new tab)
- Proc. of the International Conference on Learning Representations (ICLR), Rio de Janeiro, Brazil, AprAudioTrust: benchmarking the multifaceted trustworthiness of audio large language models (opens in a new tab)
- Proc. of the Annual Meeting of the Association for Computational Linguistics (ACL), San Diego, California, from July 2 - 7Between a rock and a hard place: the tension between ethical reasoning and safety alignment in LLMs (opens in a new tab)
- ACM MMBeyond language priors: diagnosing and fixing visual-origin hallucinations in multimodal LLMs
- The International Conference on Machine Learning (ICML), Seoul, South Korea, July 6-11CoPE: continual probe-guided expansion for large vision-language models (opens in a new tab)
- IEEE/CAA Journal of Automatica SinicaCrafting physical adversarial examples by combining differentiable and physically based renders (opens in a new tab)
- IEEE Transactions on Information Forensics and SecurityDefending against patch-based and texture-based adversarial attacks with spectral decomposition (opens in a new tab)
- Proc. of the International Conference on Learning Representations (ICLR), Rio de Janeiro, Brazil, AprEfficient audio-visual speech separation with discrete lip semantics and multi-scale global-local attention (opens in a new tab)
- The Annual AAAI Conference on Artificial Intelligence (AAAI), Singapore, Jan. - Jan.FGNet: leveraging feature-guided attention to refine sam2 for 3D EM neuron segmentation (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Denver CO, June - , .Mirror illusion art (opens in a new tab)
- The International Conference on Machine Learning (ICML), Seoul, South Korea, July 6-11MiVE: multiscale vision-language features for reference-guided video editing (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Denver CO, JunePhysical adversarial clothing evades visible-thermal detectors via non-overlapping RGB-T pattern (opens in a new tab)
202515 papers
- The European Conference on Artificial Intelligence (ECAI), Bologna, Italy, OctoberA fast and lightweight model for causal audio-visual speech separation (opens in a new tab)
- Proc. of the International Conference on Learning Representations (ICLR), Singapore, AprADBM: adversarial diffusion bridge model for reliable adversarial purification (opens in a new tab)
- Pattern RecognitionCEDNet: A cascade encoder-decoder network for dense prediction (opens in a new tab)
- Advances in Neural Information Processing (NeurIPS), San Diego, USA, Nov - DecConnectome-based modelling reveals orientation maps in the drosophila optic lobe， (opens in a new tab)
- Proc. of the International Conference on Learning Representations (ICLR), Singapore, AprEfficient neuron segmentation in electron microscopy by affinity-guided queries (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Nashville TN, JuneGlyphMastero: a glyph encoder for high-fidelity scene text editing (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Nashville TN, JuneImproving accuracy and calibration via differentiated deep mutual learning (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Nashville TN, JuneMTADiffusion: mask text alignment diffusion model for object inpainting (opens in a new tab)
- IEEE Transactions on Information Forensics and SecurityOn the Importance of Backbone to the Adversarial Robustness of Object Detectors (opens in a new tab)
- Proc. of the International Conference on Computer Vision (ICCV), Honolulu, Hawaii, OctPBCAT: patch-based composite adversarial training against physically realizable attacks on object detection (opens in a new tab)
- TPAMI (T1)Physical adversarial examples for person detectors in thermal images based on 3D modeling (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Nashville TN, JuneSAM-REF: introducing image-prompt synergy during interaction for detail enhancement in the segment anything model (opens in a new tab)
202417 papers
- TPAMI (T1)An audio-visual speech separation model inspired by cortico-thalamo-cortical circuits (opens in a new tab)
- CoRL (T1)Bridging the sim-to-real gap from the information bottleneck perspective (opens in a new tab)
- The European Conference on Computer Vision (ECCV), MiCo Milano, Italy, Sep -OctControllable navigation instruction generation with chain of thought prompting (opens in a new tab)
- Advances in Neural Information Processing (NeurIPS), Vancouver, Dec 10-15, .CooHOI: learning cooperative human-object interaction with manipulated object dynamics (opens in a new tab)
- Computer Vision and Image UnderstandingDHS-DETR: Efficient DETRs with dynamic head switching (opens in a new tab)
- Advances in Neural Information Processing (NeurIPS), Vancouver, Dec 10-15Full-distance evasion of pedestrian detectors in the physical world (opens in a new tab)
- NeurocomputingHiding from thermal imaging pedestrian detectors in the physical world (opens in a new tab)
- National Science ReviewHow variable progenitor clones construct a largely invariant neocortex (opens in a new tab)
- The International Conference on Machine Learning (ICML), Vienna, Austria, July 21-27IIANet: an intra- and inter-modality attention network for audio-visual speech separation (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, USA, June 17-21Infrared adversarial car stickers (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, USA, June 17-21Language-driven anchors for zero-shot adversarial robustness (opens in a new tab)
- ACM MMNatural language induced adversarial images (opens in a new tab)
202313 papers
- TASLP (T1)A fast high-fidelity source-filter vocoder with lightweight neural modules (opens in a new tab)
- TPAMI (T1)Adjacency constraint for efficient hierarchical reinforcement learning (opens in a new tab)
- Proc. of the International Conference on Learning Representations (ICLR), Kigali, Rwanda, May 1-5An efficient encoder-decoder architecture with top-down attention for speech separation (opens in a new tab)
- Interspeech (T1)Audio-visual speech separation in noisy environments with a lightweight iterative model (opens in a new tab)
- TPAMI (T1)Extracting semantic knowledge from GANs with unsupervised learning (opens in a new tab)
- IEEE Transactions on Intelligent Transportation SystemsFocal distillation from high-resolution data to low-resolution data for 3D object detection (opens in a new tab)
- Advances in Neural Information Processing (NeurIPS), New Orleans, Dec 10-16HEDNet: A Hierarchical Encoder-Decoder Network for 3D Object Detection in Point Clouds (opens in a new tab)
- Applied IntelligenceHiding from infrared detectors in real world with adversarial clothes (opens in a new tab)
- Advanced MaterialsMemristors-based dendritic neuron for high-efficiency spatial-temporal information processing (opens in a new tab)
- Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Vancouver, Canada, June 18-22Physically realizable natural-looking clothing textures evade person detectors via 3D modeling (opens in a new tab)
- TPAMI (T1)Recognizing object by components with human prior knowledge enhances adversarial robustness of deep neural networks (opens in a new tab)
- Proceedings of the International Conference on Information Science and Technology (ICIST), Cairo, Egypt, December 8-14TDFNet: an efficient audio-visual speech separation model with top-down fusion (opens in a new tab)
Where alumni went
Alumni are not listed on the lab homepage.
Apply to TsinghuaFall 2027
Computer Science PhD (international applicants)
- Round
- Main
- Opens
- GRE
- See program page
- English
- Per school catalog
- Fee
- RMB 800
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USUC Berkeley
Anil Aswani
- Learning theory & optimization
- CNPeking
Ruosong Wang
- Learning theory & optimization
- USGeorgia Tech
Vidya K Muthukumar
- Reinforcement learning
- SGNTU
Zhiping Lin
- Learning theory & optimization
- USCMU
Geoffrey J. Gordon
- Learning theory & optimization
- USUW
Natasha Jaques
- Reinforcement learning
Your notes
Save this lab to keep notes.
Data
Some numbers could not be verified:
- The member list could not be found on the lab site.
- The alumni list could not be found on the lab site.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Not collected
- Alumni
- Not collected
Tiers are venue classifications, not lab rankings.
Tsinghuacloses