Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Xiaoou TangSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works76 from public data
- Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition5,594
A novel model of dynamic skeletons called Spatial-Temporal Graph Convolutional Networks (ST-GCN), which moves beyond the limitations of previous methods by automatically learning both the spatial and temporal patterns from data.
- 4,387
- Unsupervised Feature Learning via Non-parametric Instance Discrimination4,044
This work forms this intuition as a non-parametric classification problem at the instance-level, and uses noise-contrastive estimation to tackle the computational challenges imposed by the large number of instance classes.
- Temporal Action Detection with Structured Segment Networks992
This paper addresses an important and challenging task, namely detecting the temporal intervals of actions in untrimmed videos by presenting a framework called structured segment network (SSN), built on temporal proposals of actions.
- Temporal Segment Networks for Action Recognition in Videos989
The proposed TSN framework, called temporal segment network (TSN), aims to model long-range temporal structure with a new segment-based sampling and aggregation scheme and won the video classification track at the ActivityNet challenge 2016 among 24 teams.
- UntrimmedNets for Weakly Supervised Action Recognition and Detection546
This paper presents a new weakly supervised architecture, called UntrimmedNet, which is able to directly learn action recognition models from untrimmed videos without the requirement of temporal annotations of action instances.
- Learning Self-Consistency for Deepfake Detection405
A novel representation learning approach, called pair-wise self-consistency learning (PCL), is introduced for training ConvNets to extract source features and detect deepfake images.
- 340
- MeMOT: Multi-Object Tracking with Memory275
An online tracking algorithm that performs the object detection and data association under a common framework, capable of linking objects after a long time span is proposed, by preserving a large spatio-temporal memory to store the identity embeddings of the tracked objects.
- Motion Guided 3D Pose Estimation from Videos256
A new graph convolutional network architecture, U-shaped GCN (UGCN), which captures both short-term and long-term motion information to fully leverage the additional supervision from the motion loss.
- Alpha-CLIP: A CLIP Model Focusing on Wherever you Want223
Alpha-CLIp, an enhanced version of CLIP with an auxiliary alpha channel to suggest attentive regions and fine-tuned with constructed millions of RGBA region-text pairs, is introduced, which not only preserves the visual recognition ability of CLIP but also enables precise control over the emphasis of image contents.
- 176
- 163
- Knowledge Guided Disambiguation for Large-Scale Scene Classification With Multi-Resolution CNNs159
A multi-resolution CNN architecture that captures visual content and structure at multiple levels is proposed and two knowledge guided disambiguation techniques to deal with the problem of label ambiguity are designed.
- 131
- 126
- TubeR: Tubelet Transformer for Video Action Detection120
TubeR is a simple solution to directly detect an action tubelet in a video by simultaneously performing action localization and recognition from a single representation and outperforms the previous state-of-the-art on commonly used action detection datasets AVA, UCF101-24 and JHMDB51-21.
- Image and Video Tokenization with Binary Spherical Quantization119
By learning an autoregressive prior for adaptive arithmetic coding, BSQ-ViT achieves comparable results on video compression with state-of-the-art video compression standards and enables masked language models to achieve competitive image synthesis quality to GAN- and diffusion-based methods.
- Optimizing Video Object Detection via a Scale-Time Lattice117
This paper presents a unified framework that integrates detection, temporal propagation, and across-scale refinement on a Scale-Time Lattice and develops an adaptive scheme with the detector invoked on demand and thus obtain improved tradeoff.
- Towards Backward-Compatible Representation Learning110
A framework to train embedding models, called backward-compatible training (BCT), is proposed as a first step towards backward compatible representation learning, and models trained with BCT successfully achieve backward compatibility without sacrificing accuracy, thus enabling backfill-free model updates of visual embeddings.
- 110
- 108
- Omni-Sourced Webly-Supervised Learning for Video Recognition107
Experiments show that by utilizing data from multiple sources and formats, OmniSource is more data-efficient in training, and several good practices, including data balancing, resampling, and cross-dataset mixup are adopted in joint training.
- 105
- 103
- Towards Causal Benchmarking of Bias in Face Analysis Algorithms100
An experimental method for measuring algorithmic bias of face analysis algorithms, which manipulates directly the attributes of interest, e.g., gender and skin tone, in order to reveal causal links between attribute variation and performance change, is developed.
- Trajectory Convolution for Action Recognition95
This work proposes a new CNN architecture TrajectoryNet, which incorporates trajectory convolution, a new operation for integrating features along the temporal dimension, to replace the existing temporal convolution.
- 90
- Positive-Congruent Training: Towards Regression-Free Model Updates78
A simple approach for PC training, Focal Distillation, is proposed, which enforces congruence with the reference model by giving more weights to samples that were correctly classified, and it is found that negative flips can be further reduced without affecting the new model’s accuracy.
- Action Recognition With Spatial-Temporal Discriminative Filter Banks73
This work hypothesizes that current architectures have poor sensitivity to finer details and exploits recent advances in the fine-grained recognition literature to improve the model in this aspect, and obtains state-of-the-art performance on Kinetics-400 and Something-Something-V1, the two major large-scale action recognition benchmarks.
- A Full-duplex Speech Dialogue Scheme Based On Large Language Models54
In automatic quality evaluations simulating real-life interaction, the proposed system reduces the average conversation response latency by more than threefold compared with LLM-based half-duplex dialogue systems while responding within less than 500 milliseconds in more than 50% of evaluated interactions.
- Learning Hierarchical Graph Neural Networks for Image Clustering44
This work proposes a hierarchical graph neural network model that learns how to cluster a set of images into an unknown number of identities using a training set of image annotated with labels belonging to a disjoint set of identities.
- MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models33
Multi-Image Augmented Direct Preference Optimization (MIA-DPO) mitigates the scarcity of diverse multi-image training data by extending single-image data with unrelated images arranged in grid collages or pic-in-pic formats, significantly reducing the costs associated with multi-image data annotations.
- SSCAP: Self-supervised Co-occurrence Action Parsing for Unsupervised Temporal Action Segmentation33
An unsupervised method that operates on a corpus of unlabeled videos and predicts a likely set of temporal segments across the videos, which achieves state-of-the-art performance on all datasets and can even outperform some weakly-supervised approaches, demonstrating its effectiveness and generalizability.
- RAR: Retrieving and Ranking Augmented MLLMs for Visual Recognition31
RAR, a Retrieving And Ranking augmented method for Multimodal Large Language Models, demonstrates a significant improvement in performance on 5 fine-grained visual recognition benchmarks, 11 few-shot image recognition datasets, and the 2 object detection datasets under the zero-shot recognition setting.
- Compatibility-aware Heterogeneous Visual Search31
Compared to ordinary visual search using the largest embedding model (paragon), Cmp-NAS achieves 80-fold and 23-fold cost reduction while maintaining accuracy within 0.3% and 1.6% of the paragon on DeepFashion2 and IJB-C respectively.
- 30
- 23
- 22
- Regression Bugs Are In Your Model! Measuring, Reducing and Analyzing Regressions In NLP Model Updates22
This work focuses on quantifying, reducing and analyzing regression errors in the NLP model updates, using negative flip rate as regression measure, and shows that regression has a prevalent presence across tasks in the GLUE benchmark.
- RetinaGS: Scalable Training for Dense Scene Rendering with Billion-Scale 3D Gaussians21
A general model parallel training method for 3DGS, named RetinaGS, which uses a proper rendering equation and can be applied to any scene and arbitrary distribution of Gaussian primitives, and observes a clear positive trend of increasing visual quality when increasing primitive numbers.
- Learning to Recognize Patch-Wise Consistency for Deepfake Detection.21
This work proposes a novel representation learning approach for Deepfake detection, called patch-wise consistency learning (PCL), which learns by measuring the consistency of image source features, resulting to representation with good interpretability and robustness to multiple forgery methods.
- On Improving Temporal Consistency for Online Face Liveness Detection System20
This paper focuses on improving the online face liveness detection system to enhance the security of the downstream face recognition system, and proposes a simple yet effective solution based on temporal consistency that is more robust against several presentation attacks in various scenarios.
- 20
- Stochastic Backpropagation: A Memory Efficient Strategy for Training Video Models18
Stochastic Backpropagation reduces the GPU memory cost by eliminating the need to cache activation values corresponding to the dropped backward paths, whose amount can be controlled by an adjustable keep-ratio.
- X-Prompt: Towards Universal In-Context Image Generation in Auto-Regressive Vision Language Foundation Models15
This work introduces X-Prompt, a purely auto-regressive large-vision language model designed to deliver competitive performance across a wide range of both seen and unseen image generation tasks, all within a unified in-context learning framework.
- 15
- Harnessing Unrecognizable Faces for Improving Face Recognition13
Accounting for recognizability in the design and implementation of face recognition systems reduces the error rate of single-image face recognition by 58% at FAR=1e-5 on the IJB-C Covariate Verification benchmark, and reduces the verification error rate by 24%.
- 11
- 10
- 9
- Elodi: Ensemble Logit Difference Inhibition for Positive-Congruent Training8
The role of ensembles in reducing NFR is analyzed and it is observed that they remove negative flips that are typically not close to the decision boundary, but often exhibit large deviations in the distance among their logits.
- 7
- Bootstrap3D: Improving Multi-View Diffusion Model with Synthetic Data6
Bootstrap3D, a novel framework that automatically generates filtered multi-view images to assist in training multi-view diffusion models, is proposed and a Training Timestep Reschedule (TTR) strategy that leverages the denoising process to learn multi-view consistency while maintaining the original 2D diffusion prior is presented.
- 6
- KnapFormer: An Online Load Balancer for Efficient Diffusion Transformers Training4
KnapFormer builds on the insight that strong synergy exists between sequence parallelism and the need to address the significant token imbalance across ranks, and achieves minimal communication overhead and less than 1% workload discrepancy in real-world training workloads with sequence length varying from a few hundred to tens of thousands.
- 4
- 4
- Zeta Hull Pursuits: Learning Nonconvex Data Hulls4
A novel approach is proposed based on a graph-theoretic measure that leverages graph cycles to characterize the structural complexities of input data points to achieve state-of-the-art accuracy in text and image classification tasks.
- Small target disease detection based on YOLOv5 framework for intelligent bridges3
A small target disease detection method using YOLOv5 framework for detecting small apparent diseases on intelligent bridges, aiming to address the problem of missed and false detection.
- Lightweight concrete crack segmentation method based on Deeplabv3+3
A lightweight concrete crack segmentation method based on DeeplabV3+, which effectively reduces the number of model parameters and improves Crack segmentation accuracy by replacing the backbone and adding the SSH (Single Stage Headless) context module and ECA (Efficient Channel Attention).
- 3
- Complete-muE: Optimal Hyperparameter Transfer and Scaling for MoE Models2
Extensive language model and diffusion model pretraining experiments confirm that complete-muE yields relatively stable hyperparameter optima across model architectures and parameter counts -- with only minor drift consistent with the non-strict SDE behavior of Bridge~II.
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.