AI Paper Atlas

Research map / Video / Video-language understanding

ST-LLM: Large Language Models Are Effective Temporal Learners

Ruyang Liu, Chen Li, Haoran Tang, Yixiao Ge, Ying Shan, Ge Li

ECCV 2024 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 82 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
feed all spatial-temporal tokens into the LLM instead of mean pooling over time (ablation)ST-LLM with InstructBLIP, full fine-tuning · MVBenchaverage accuracy: 47.8 -> 54.9Improved
feed all spatial-temporal tokens into the LLM instead of mean pooling over time (LoRA) (ablation)ST-LLM with InstructBLIP, LoRA · MVBenchaverage accuracy: 45.2 -> 51.9Improved
use ST tokens with MiniGPT-4 base instead of mean pooling (ablation)ST-LLM with MiniGPT-4 · MVBenchaverage accuracy: 46.4 -> 50.5Improved
add dynamic masking (vs fixed frames or static 50% mask) during training (ablation)ST-LLM · MVBench, varying test frame countsaccuracy robustnessImproved
add static 50% masking instead of dynamic masking (ablation)ST-LLM · MVBench, varying test frame countsaccuracy robustnessNo clear effect
add MVM loss on top of dynamic masking (ablation)ST-LLM · MVBenchaccuracyImproved
add global branch via zero-initialized adapter to local input (ablation)ST-LLM · VideoChatGPT-Bench Temporalscore: 2.75 -> 2.93Improved
add global branch via zero-initialized adapter to local input (ablation)ST-LLM · MVBenchaverage accuracy: 54.9 -> 54.7No clear effect
add global tokens by simple addition instead of adapter (ablation)ST-LLM · MVBenchaverage accuracy: 54.9 -> 50.1Hurt
add separator tokens between frames or modalities (ablation)ST-LLM · MVBenchaccuracyNo clear effect
add absolute positional encodings to visual tokens (ablation)ST-LLM · MVBenchaccuracyHurt
add image data to video instruction tuning (InstructBLIP base) (ablation)ST-LLM with InstructBLIP · MVBenchaverage accuracy: 54.9 -> 53.1Hurt

Later papers that compared with this method

7 later accepted papers report ST-LLM as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.

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