Archives AI News

TimeScope: Towards Task-Oriented Temporal Grounding In Long Videos

arXiv:2509.26360v1 Announce Type: cross Abstract: Identifying key moments in long videos is essential for downstream understanding and reasoning tasks. In this paper, we introduce a new problem, Taskoriented Temporal Grounding ToTG, which aims to localize time intervals containing the necessary…

The Ever-Evolving Science Exam

arXiv:2507.16514v3 Announce Type: replace-cross Abstract: As foundation models grow rapidly in capability and deployment, evaluating their scientific understanding becomes increasingly critical. Existing science benchmarks have made progress towards broad Range, wide Reach, and high Rigor, yet they often face two…

Find the Fruit: Zero-Shot Sim2Real RL for Occlusion-Aware Plant Manipulation

arXiv:2505.16547v2 Announce Type: replace-cross Abstract: Autonomous harvesting in the open presents a complex manipulation problem. In most scenarios, an autonomous system has to deal with significant occlusion and require interaction in the presence of large structural uncertainties (every plant is…

All hail the new Fat Bear Champion

The votes are in, and the winner of Fat Bear Week 2025 is the indomitable Bear 32, also known as Chunk. The public votes for their favorite brown bear each year in the March Madness-style tournament held by Katmai National…

Compute-Optimal Quantization-Aware Training

Quantization-aware training (QAT) is a leading technique for improving the accuracy of quantized neural networks. Previ- ous work has shown that decomposing training into a full-precision (FP) phase followed by a QAT phase yields superior accuracy compared to QAT alone.…