training-free-video-temporal-grounding-frozen-vision-language-embedding-backbones-matching CLM-0249In training-free video temporal grounding with frozen vision-language embedding backbones, matching a query only against frames sampled at a fixed rate leaves temporal context unused: adding clip-level (sliding-window) similarities projected onto the frame timeline raises localization accuracy even though clip-only similarity is worse than frame-only on its own.
Capability: Reasoning about time in video
Observed on
Frozen video-capable VLM embedders (Qwen3-VL-Embedding, BLIP-2, RzenEmbed) used for similarity-based grounding; no task-specific training; short-to-medium videos (Charades-STA ~30s.
Sources
- Validated by: Charades-STA mIoU 51.30 vs 45.69 for the strongest prior training-free method (TAG), and vs 45.41/44.03 when TAG/TFVTG are re-run on the same backbone; QVHighlights 38.64 mAP vs 33.15. Component ablation: 45.69 base, 47.06 with Q-TTA only, 50.25 with MSF only, 51.30 both. Granularity ablation on the same backbone: frame-only 47.06, clip-only 42.82, fused 51.30. Also OOD splits Charades-CG and Charades-CD.. [truncated]
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- In training-free video temporal grounding that localizes segments by cosine similarity between a text query embedding and video features, clip-level features alone localize worse than frame-level features alone, yet projecting clip similarities onto the frame timeline and mixing the two streams beats either alone, because frames supply boundary resolution and clips supply temporal context.Reasoning about time in video · reviewed by AI
- For streaming video understanding with 7-8B multimodal backbones, question-relevant historical frames can be retrieved from the keys of only the first few language-Transformer layers (layer ~3-4 of 28-32), so continuous frames need not be prefilled through full model depth: indexing shallow and re-prefilling only the retrieved units at full depth cuts per-frame prefill up to 52x and 20-second end-to-end latency up to 15x while scoring about 1.7 points above the best same-backbone baseline on OVO-Bench and StreamingBench.Reasoning about time in video · reviewed by AI
- For long-video QA under a fixed frame budget with a frozen 7B multimodal model, organizing frames into a question-agnostic hierarchical temporal tree and searching it coarse-to-fine (answer, then issue a search statement, then expand the relevant branch) raises accuracy over uniform sampling and over agent methods that score frames independently, with the gain largest on the longest-video benchmarks.Reasoning about time in video · reviewed by AI
- For long-video QA under a fixed 32-frame budget with 7B-8B multimodal backbones, selecting frames from the structural shape of the frame-query similarity curve (peak, rising, falling, transition regions) beats ranking frames by similarity magnitude, and the margin grows with video length — a few points on short clips and roughly 8 points on hour-long videos — while on short/medium-clip benchmarks it is near-indistinguishable from other selectors.Reasoning about time in video · reviewed by AI
- For 4-8B multimodal models on long-video QA, RL training on questions authored from short scene segments but answered over the whole video improves benchmark accuracy more than RL on questions authored from the whole video, because segment-authored questions target details a whole-video view overlooks; including the segment timestamp in the training instruction adds a further gain by removing referential ambiguity, and is not needed at inference.Reasoning about time in video · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: medium. Falsifier as drafted: On the same backbone and benchmarks, adding kernel-projected clip-level similarities to frame-level similarities fails to improve mIoU or R@0.7 over frame-only, or the reported gain disappears once the frame-only baseline is tuned with the same interval-decoding hyperparameters.