AI Paper Atlas

Research map / Video / Video-language understanding

STRIDE: When to Speak Meets Sequence Denoising for Streaming Video Understanding

Junho Kim, Hosu Lee, James Rehg, Minsu Kim, Yong Man Ro

ECCV 2026 · Video › Video-language understanding · approach: Long video understanding · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace autoregressive binary activation with masked diffusion (STRIDE)Qwen3-VL-2B activation model · ET-Benchavg F1: 24.3 -> 32.6Improved
STRIDE vs Baseline-AR on Qwen3-VL-8BQwen3-VL-8B · OVO-Benchoverall avg: 52.81 -> 59.07Improved
STRIDE vs Baseline-AR on Qwen3-VL-8BQwen3-VL-8B · StreamingBenchoverall avg: 57.12 -> 59.29Improved
STRIDE vs Baseline-AR on forward active respondingQwen3-VL-8B · OVO-Benchforward active responding avg: 42.73 -> 59.70Improved
add STRIDE to Gemma3-4BGemma3-4B · OVO-Benchoverall avg: 42.13 -> 50.51Improved
add STRIDE to InternVL3-8BInternVL3-8B · StreamingBenchoverall avg: 53.97 -> 57.58Improved
use structured span masking instead of independent masking (ablation)STRIDE · ET-Benchavg F1: 7.2 -> 21.1Improved
add full masking and span unmasking to span masking (ablation)STRIDE · ET-Benchavg F1: 21.1 -> 32.6Improved
add sequence duplication (ablation)STRIDE · ET-Benchavg F1: 22.9 -> 32.6Improved
use selective re-masking instead of last-only (ablation)STRIDE · ET-Benchavg F1: 22.6 -> 32.6Improved
set retention threshold tau to 0 (retain all) (ablation)STRIDE · ET-Bench TVGF1: -19.7 ptHurt
increase denoising steps K beyond 8 (ablation)STRIDE · ET-Benchmean F1No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy