SoftServe: A Scalable Quasi-Newton Method for Deep Learning

Generative Cinematographer: Composing Camera and Object Motion in 3D

Current controllable video generation systems often rely on 2D motion trajectories or sparse drag signals for object motion.

2026-10-01 ·arxiv.org

I gave my coding agent a sense of taste. It picks restaurants from my music.

My coding agent can refactor a thousand lines without breaking a sweat, but ask it "where should I take a date Friday night" and it's useless. It knows my code. It knows nothing about my taste.

• Qloo coding agent gains taste sense
• Qloo Taste API connects 250M cultural entities
• Agent recommends restaurants based on music preferences

2026-10-01 ·dev.to

Watch, Infer, Coordinate: Inferring Robot Partner Constraints for Zero-Shot Coordination

Robots operating in the physical world will increasingly need to coordinate with other robots, particularly in manipulation tasks where an object may be too large or heavy for a single robot to carry…

2026-10-01 ·arxiv.org

Higher-Order Molecular Grammars for Generative and Foundation Models in Chemistry

Molecular learning models are strongly shaped by their underlying representations. Yet standard sequential and graph formalisms struggle to explicitly encode higher-order topology, such as ring…

2026-10-01 ·arxiv.org