[ GitHub ] [ arXiv ] [ HuggingFace ] [ ModelScope ] Grounded in Time is a benchmark and dataset for temporal grounding under perceptual ambiguity in robotic manipulation. Its tasks are deliberately built so that the object a command refers to cannot be identified from the current frame alone. Each scene contains several perceptually near-identical objects, and what separates them is not appearance but the temporal structure of the episode — which test tube was contaminated first, which container was filled last, which part was handed over just now. A model that grounds the instruction on the present frame alone will pick the wrong object; resolving the reference requires tracking and…
Independent publisher
XGQ
XGQ12345
Undergraduate student at South China University of Technology, working on embodied robotics.
Models in Library0
Datasets in Library1
Models on Hugging Face—
Followers—
Compare the published evidence
This profile connects the 1 model and dataset record shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Robotics.
Reported licenses: apache-2.0. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.