AI Science Tracker
EN

Predictive Spatio-Temporal Scene Graphs for Semi-Static Scenes

Not stated in the reviewed arXiv full text · 30 Apr 2026 · computer science

Discovery

Predictive Spatio-Temporal Scene Graphs for Semi-Static Scenes

Source check: Source checked. Not independent scientific confirmation.

Finding: PredictiveGraphs combines open-vocabulary scene mapping, language-model assistance, and state persistence modeling to predict where semi-static objects will be. The paper evaluates simulation tasks and 25 real-world queries during a three-week laboratory collection; this is a bounded robotics demonstration, not long-term deployment.

Technology: CLIP embeddings and an LLM verifier/planner support object recognition and scene queries; a Bayesian persistence filter predicts object-receptacle states.

Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.

AI contribution
The learned scene-state system predicts object locations over time and enables time-aware navigation queries.
Test and result
Evaluation covered 250 predictive and 125 adaptive simulation queries and 25 open-vocabulary real-world queries during three weeks in a laboratory.
Limitation / still unknown
Bounded laboratory evaluation; not sustained deployment, and the paper notes potential language-model hallucinations.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://arxiv.org/html/2605.00121v1 · https://arxiv.org/html/2605.00121v3

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 22 Nov 2026. Review scope: Primary source full-text Methods and Results; version date and reported validation scope checked..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse computer science discoveries.