Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Research landscape, hotspots, and evolutionary trends of knee osteoarthritis and oxidative stress: a multidimensional bibliometric analysis.

Frontiers in medicine·2026

Same author

Reconciling sensitivity and breathability in flexible iontronic pressure sensors <i>via</i> a hierarchical fibrous architecture.

Materials horizons·2026

Same author

A Randomized Controlled Trial of Yizhi Kaiqiao Formula Combined With Repetitive Transcranial Magnetic Stimulation on Neurocognitive and Social Outcomes in Preschool Children With Autism Spectrum Disorder.

Developmental neurobiology·2026

Same author

Advancing high-altitude medicine: a model for the future.

Signal transduction and targeted therapy·2026

Same author

Comparison of staple line reinforcement with nonreinforcement in Roux-en-Y gastric bypass surgery: a meta-analysis.

Surgery for obesity and related diseases : official journal of the American Society for Bariatric Surgery·2026

Same author

<sup>68</sup>Ga-Labeled LLP2A for PET Imaging of Very Late Antigen-4 in Acute Cardiac Rejection.

Molecular pharmaceutics·2026

Same journal

Relation DETR+: Exploring Explicit Position Relation Prior for Dense Prediction.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

RBF++: Quantifying and Optimizing Reasoning Boundaries across Measurable and Unmeasurable Capabilities for Chain-of-Thought Reasoning.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

CAFE: Cross-View Adaptive Fusion and Cluster Center Enhancement for Robust Multi-View Clustering.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Driving.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

Ethics-Aware Safe Reinforcement Learning for Rare-Event Risk Control in Interactive Urban Driving.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

Learning Shape Anchors for Holistic Indoor Scene Understanding.

IEEE transactions on pattern analysis and machine intelligence·2026

See all related articles

Search research articles

Related Experiment Video

Updated: Jul 26, 2025

Localizing Function-specific Targets for Transcranial Magnetic Stimulation in the Absence of Navigation Equipment

Localizing Function-specific Targets for Transcranial Magnetic Stimulation in the Absence of Navigation Equipment

Published on: May 23, 2025

Bilateral Relation Distillation for Weakly Supervised Temporal Action Localization.

Zhe Xu, Kun Wei, Erkun Yang

IEEE Transactions on Pattern Analysis and Machine Intelligence

|June 15, 2023

Summary

This summary is machine-generated.

Weakly supervised temporal action localization (WSTAL) models struggle with accurate action interval identification. Our Bilateral Relation Distillation (BRD) method improves WSTAL by modeling category and sequence relations for better results.

More Related Videos

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

A Novel Experimental and Analytical Approach to the Multimodal Neural Decoding of Intent During Social Interaction in Freely-behaving Human Infants

A Novel Experimental and Analytical Approach to the Multimodal Neural Decoding of Intent During Social Interaction in Freely-behaving Human Infants

Published on: October 4, 2015

Related Experiment Videos

Last Updated: Jul 26, 2025

Localizing Function-specific Targets for Transcranial Magnetic Stimulation in the Absence of Navigation Equipment

Localizing Function-specific Targets for Transcranial Magnetic Stimulation in the Absence of Navigation Equipment

Published on: May 23, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

A Novel Experimental and Analytical Approach to the Multimodal Neural Decoding of Intent During Social Interaction in Freely-behaving Human Infants

A Novel Experimental and Analytical Approach to the Multimodal Neural Decoding of Intent During Social Interaction in Freely-behaving Human Infants

Published on: October 4, 2015

Area of Science:

Computer Science
Artificial Intelligence
Machine Learning

Background:

Weakly supervised temporal action localization (WSTAL) uses video-level labels for action interval detection in untrimmed videos.
Existing WSTAL models often produce inaccurate and incomplete localization due to focusing on dominant segments.

Purpose of the Study:

To propose a novel method, Bilateral Relation Distillation (BRD), for improving WSTAL.
To address the limitations of current WSTAL models by enhancing relation modeling.

Main Methods:

BRD jointly models relations at both category and sequence levels.
Category-level relations are captured by distilling knowledge from a pre-trained language model using correlation alignment and category-aware contrast.
Sequence-level relations are modeled using a gradient-based feature augmentation method for representation consistency.

Main Results:

The proposed BRD method achieves state-of-the-art performance.
Experiments were conducted on the THUMOS14 and ActivityNet1.3 datasets.

Conclusions:

Bilateral Relation Distillation (BRD) effectively improves weakly supervised temporal action localization.
Jointly modeling category and sequence relations is crucial for accurate action localization in videos.