Tag
Veridive is a tool that lets you find key moments in videos via chat, enabling quick discovery of important segments.
Interhuman.ai has launched a Streaming API for its Inter-1 model, enabling real-time detection of 12 social signals from live video streams via WebSocket, along with engagement tracking and conversation quality scoring.
Artifact-Bench is a comprehensive benchmark that evaluates multimodal large language models on detecting and analyzing artifacts in AI-generated videos, revealing significant limitations and misalignment with human perception.
Introduces Knowly AI tool, capable of interpreting YouTube videos and arXiv papers with impressive results. Interaction and interpretation quality rival NotebookLM. Comes with a Chrome extension already featured by Google. Drawbacks: limited free quota and slightly slow vector processing.
Perceptron Inc. released its flagship video analysis model Mk1, claiming 80-90% lower cost than competitors while achieving strong performance on spatial and video reasoning benchmarks.
This paper introduces three parameter-efficient methods for multi-view proficiency estimation on the Ego-Exo4D dataset, shifting from discriminative classification to generative feedback. The proposed models achieve state-of-the-art accuracy with significantly fewer parameters and training epochs than video-transformer baselines.