AI Paper Atlas

Research map / Video / Video-language understanding

Perception Test: A Diagnostic Benchmark for Multimodal Video Models

Viorica Patraucean, Lucas Smaira, Ankush Gupta, Adria Recasens Continente, Larisa Markeeva, Dylan Sunil Banarse, Skanda Koppula, Joseph Heyward, Mateusz Malinowski, Yi Yang, Carl Doersch, Tatiana Matejovicova and 12 more

NeurIPS 2023 · Video › Video-language understanding · approach: Event and anomaly reasoning in video · 17 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
fine-tune SeViLA on Perception Test train splitSeViLA · Perception Test mc-vQA valtop-1 accuracy: 46.2 -> 62.0Improved
8-shot prompting instead of 0-shotFlamingo-3B · Perception Test mc-vQA valtop-1 accuracy: 43.6 -> 45.8Improved
scale Flamingo from 3B to 80BFlamingo · Perception Test mc-vQA val 0-shottop-1 accuracy: 43.6 -> 41.6Hurt
use video-language models instead of simple frequency baselineFlamingo / SeViLA 0-shot or 8-shot · Perception Test mc-vQA valtop-1 accuracy: 45.8 vs 51.0 (8-shot frequency)Hurt
add audio features to video features (ablation)ActionFormer · Perception Test action localisationmAP: 14.74 -> 15.56Improved
add audio features to video features (ablation)ActionFormer · Perception Test sound localisationmAP: 13.04 -> 15.46Improved
use simple SiamFC tracker instead of more complex Starkobject tracking · Perception Testaverage IoU: 0.56 -> 0.67Improved
use TAP-Net instead of static baselinepoint tracking · Perception Testaverage Jaccard: 0.361 -> 0.401Improved
moving camera instead of static cameraSiamFC · Perception Test object trackingaverage IoU: 0.70 / 0.69 -> 0.42 / 0.54 (static dummy / SiamFC)Hurt

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