Research map / Video / Video-language understanding
CTEPM: Continuous-Time Event Process Memory for Long-Video Language Models
Yangyang Liu
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add CTEPM event process memory | LLaVA-NeXT-Video-7B · LongVideoBench (total) | Top-1 accuracy: 46.0 -> 49.2 | Improved |
| add CTEPM event process memory | LLaVA-NeXT-Video-7B · LongVideoBench (900-3600s videos) | Top-1 accuracy: 41.0 -> 45.6 | Improved |
| add CTEPM event process memory | LLaVA-NeXT-Video-7B · Video-MME w/ subs | Top-1 accuracy: 52.5 -> 55.0 | Improved |
| add CTEPM event process memory | LLaVA-NeXT-Video-7B · MLVU | M-Avg: 34.0 -> 37.8 | Improved |
| add CTEPM event process memory | LLaVA-NeXT-Video-7B · MLVU Action Count | accuracy: 21.0 -> 27.8 | Improved |
| double the frame budget from 64 to 128 for uniform sampling | LLaVA-NeXT-Video-7B · LongVideoBench (total) | Top-1 accuracy: 46.0 -> 47.1 | Improved |
| replace uniform sampling with segment retrieval | LLaVA-NeXT-Video-7B · MLVU | M-Avg: 34.0 -> 34.6 | Improved |
| replace uniform sampling with MA-LMM-style memory | LLaVA-NeXT-Video-7B · LongVideoBench (>30 min videos) | Top-1 accuracy: 40.8 -> 42.4 | Improved |
| use CTEPM instead of MA-LMM-style memory | LLaVA-NeXT-Video-7B · LongVideoBench (>30 min videos) | Top-1 accuracy: 42.4 -> 45.9 | Improved |
| use CTEPM instead of uniform sampling at small frame cap | LLaVA-NeXT-Video-7B · LongVideoBench (F=16) | Top-1 accuracy: +4.6 (F=16), +3.1 (F=64) | Improved |
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.
- FastVID — FastVID: Dynamic Density Pruning for Fast Video Large Language Models (NeurIPS 2025) · better on 7 of 7 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
- $\infty$-Video: A Training-Free Approach to Long Video Understanding via Continuous-Time Memory Consolidation (ICML 2025 · similarity 0.91)
- EventLens: Event-Structure Reinforcement Learning for Video Understanding (NeurIPS 2026 · similarity 0.90)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.90)
- Seq2Time: Sequential Knowledge Transfer for Video LLM Temporal Grounding (CVPR 2025 · similarity 0.89)
- Temporal Reasoning Transfer from Text to Video (ICLR 2025 · similarity 0.89)
- Streaming Long Video Understanding with Large Language Models (NeurIPS 2024 · similarity 0.89)
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.
