Research map / Video / Video-language understanding
SpotEM: Efficient Video Search for Episodic Memory
Santhosh Kumar Ramakrishnan, Ziad Al-Halah, Kristen Grauman
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 |
|---|---|---|---|
| Use query-conditioned recursive clip selection instead of OCSampler | InternVideo · Ego4D NLQ, η=90 | MR@1: 5.72 -> 9.62 | Improved |
| Replace ImageNet indexing features with RIO features (ablation) | SpotEM without distillation on InternVideo · Ego4D NLQ, η=90 | MR@1: 6.99 -> 7.48 | Improved |
| Add distillation losses (ablation) | SpotEM with RIO on InternVideo · Ego4D NLQ, η=90 | MR@1: 7.48 -> 9.62 | Improved |
| Remove room features (ablation) | SpotEM with RIO on InternVideo · Ego4D NLQ, η=90 | MR@5: 16.83 -> 13.36 | Hurt |
| Remove interaction features (ablation) | SpotEM with RIO on InternVideo · Ego4D NLQ, η=90 | MR@5: 16.83 -> 11.05 | Hurt |
| Remove object features (ablation) | SpotEM with RIO on InternVideo · Ego4D NLQ, η=90 | MR@5: 16.83 -> 13.12 | Hurt |
| Append RIO features directly to all-clip video features (ablation) | InternVideo · Ego4D NLQ | MR@1: 11.45 -> 10.91 | Hurt |
| Append RIO features directly to all-clip video features (ablation) | ReLER · Ego4D NLQ | MR@1: 11.50 -> 10.68 | Hurt |
| Append RIO features directly to all-clip video features (ablation) | EgoVLP · Ego4D NLQ | MR@1: 9.54 -> 9.33 | Hurt |
| Increase recursive selection from N=1 to N=4 (ablation) | SpotEM · Ego4D NLQ, multiple efficiency levels | MR@5 | Improved |
| Increase recursive selection beyond N=4 (ablation) | SpotEM · Ego4D NLQ, multiple efficiency levels | MR@5 | Hurt |
| Use RIO semantic indexing on TACoS (ablation) | SpotEM · TACoS NLG | MR | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- OCSampler — OCSampler: Compressing Videos to One Clip With Single-Step Sampling (CVPR 2022) · better on 8 of 8 shared results
- LiteEval — LiteEval: A Coarse-to-Fine Framework for Resource Efficient Video Recognition (NeurIPS 2019) · better on 1 of 1 shared result
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
- Fewer Steps, Better Performance: Efficient Cross-Modal Clip Trimming for Video Moment Retrieval Using Language (AAAI 2024 · similarity 0.89)
- NaQ: Leveraging Narrations As Queries To Supervise Episodic Memory (CVPR 2023 · similarity 0.89)
- Keep It Simple: Multi-Key Episodic Memory Retrieval for Ultra-Long Video Understanding (ECCV 2026 · similarity 0.88)
- Interactive Episodic Memory with User Feedback (CVPR 2026 · similarity 0.88)
- Video-RAG: Visually-aligned Retrieval-Augmented Long Video Comprehension (NeurIPS 2025 · similarity 0.87)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.86)
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.
