Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks196 from public data
- 211
- Matcher: Segment Anything with One Shot Using All-Purpose Feature Matching199
This work presents Matcher, a novel perception paradigm that utilizes off-the-shelf vision foundation models to address various perception tasks and demonstrates impressive generalization performance across various segmentation tasks, all without training.
- Improved Algorithms for the Feedback Vertex Set Problems160
Improved parameterized algorithms for the Feedback Vertex Set problem on both unweighted and weighted graphs are presented, both of which run in time O(5kkn2).
- TeachText: CrossModal Generalized Distillation for Text-Video Retrieval158
This work proposes a novel generalized distillation method, TeachText, which leverages complementary cues from multiple text encoders to provide an enhanced supervisory signal to the retrieval model and extends the method to video side modalities and shows that it can effectively reduce the number of used modalities at test time without compromising performance.
- Cross Modal Retrieval with Querybank Normalisation149
This work formulates a simple but effective framework called Querybank Normalisation (QB-NORM) that re-normalises query similarities to account for hubs in the embedding space, and shows that QB- NORM improves retrieval performance without requiring retraining.
- 139
- 131
- Weakly Supervised Video Moment Localization with Contrastive Negative Sample Mining125
A novel weakly supervised solution to video moment localization by introducing Contrastive Negative sample Mining (CNM), which uses a learnable Gaussian mask to generate positive samples, highlighting the video frames most related to the query, and considers other frames of the video and the whole video as easy and hard negative samples respectively.
- Amplifying Key Cues for Human-Object-Interaction Detection102
Two methods to amplify key cues in the image and a method to combine these and other cues when considering the interaction between a human and an object, which exceeds prior HOI methods across standard benchmarks by a considerable margin.
- Synthetically Supervised Feature Learning for Scene Text Recognition84
This work designs a multi-task network with an encoder-discriminator-generator architecture to guide the feature of the original image toward that of the clean image, and significantly outperforms the state-of-the-art methods on standard scene text recognition benchmarks in the lexicon-free category.
- 73
- Understanding and Mitigating Overfitting in Prompt Tuning for Vision-Language Models71
Subspace Prompt Tuning (Sub PT) is proposed to project the gradients in back-propagation onto the low-rank subspace spanned by the early-stage gradient flow eigenvectors during the entire training process and successfully eliminate the overfitting problem.
- Comparing HMM, maximum entropy, and conditional random fields for disfluency detection71
Overall it is found that that the conditional modeling approaches (Maxent and CRF) provide benefit over the HMM approach and provide a more principled way to model correlated features.
- Overcoming Catastrophic Forgetting During Domain Adaptation of Seq2seq Language Generation67
This work proposes an innovative framework, RMR_DSE that leverages a recall optimization mechanism to selectively memorize important parameters of previous tasks via regularization, and uses a domain drift estimation algorithm to compensate the drift between different do-mains in the embedding space.
- Fused GRU with semantic-temporal attention for video captioning56
This work proposes an end-to-end pipeline named Fused GRU with Semantic-Temporal Attention (STA-FG), which can explicitly incorporate the high-level visual concepts to the generation of semantic-temporal attention for video captioning.
- Recognizing Unseen Attribute-Object Pair with Generative Model55
This paper proposes a generative model with the encoder-decoder mechanism that bridges visual and linguistic information in a unified end-to-end network that presents the impressive potential to find an intrinsic attribute-object feature representation.
- Towards Generalisable Video Moment Retrieval: Visual-Dynamic Injection to Image-Text Pre-Training52
This work explores multi-modal correlations derived from large-scale image-text data to facilitate generalisable VMR and proposes a generic method, referred to as Visual-Dynamic Injection (VDI), to empower the model's understanding of video moments.
- 51
- Measuring and Reducing LLM Hallucination without Gold-Standard Answers50
This work proposes Factualness Evaluations via Weighting LLMs (FEWL), an innovative hallucination metric that is specifically designed for the scenario when gold-standard answers are absent and leverages the answers from off-the-shelf LLMs that serve as a proxy of gold-standard answers.
- 50
- DIURNAL VARIATION IN THE P300 COMPONENT OF HUMAN COGNITIVE EVENT-RELATED POTENTIAL47
The results suggest the existence of diurnal variation in human cognitive functions and the P300 latency was correlated positively with subjectively rated sleepiness and negatively correlated withsubjectively rated attention level.
- Uncertainty-Induced Transferability Representation for Source-Free Unsupervised Domain Adaptation46
A novel Uncertainty-induced Transferability Representation (UTR), which leverages uncertainty as the tool to analyse the channel-wise transferability of the source encoder in the absence of the sources data and target labels, and a novel Calibrated Adaption Framework for SFUDA.
- 44
- 42
- Exploring the Potential of Large Foundation Models for Open-Vocabulary HOI Detection39
A novel end-to-end open vocabulary HOI detection framework with conditional multi-level decoding and fine-grained semantic enhancement (CMD-SE), harnessing the potential of Visual-Language Models (VLMs) by leveraging large language models such as GPT models.
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 2026-10-14. 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.