Loading activity…
@furongh
Assoc. prof. of CS at University of Maryland. Researcher in #PhysicalAI, #TrustworthyML, #EthicalAI, AI #Alignment, #AI for ALL.
AI
Humans& says developers can access the model through its Agents API without joining a waitlist.
AI
NVIDIA researchers say Flora turns human and robot footage into timestamped captions and rewards for robotics training.
AI
The incoming professor says they're recruiting PhD students and postdocs for fall 2027 in reinforcement learning and deep learning.
AI
FlowBank’s author reports an average score of 73.40 across five benchmarks, versus 70.40 for the strongest automated baseline they evaluated.
AI
A μ₀ team member says the model predicts 3D motion traces, which they propose as an alternative to pixels or opaque latent representations.
AI
AI help with prose should not be punished, argues @furongh, who announced an ICML 2027 program chair role. But reviewers should not have to spend hours on work its authors haven't checked or understood.
AI
A post sharing a Science release says the clinical trial forecasts addressed both efficacy and avoiding side effects.
AI
The MathAdv team says the reverse—solving an equivalent reformulation but not its original—never occurred in its paired test. It calls this evidence of brittleness, not proof of memorization.
AI
The author says their blog maps 16 overlapping research traditions by what models represent and predict, how they connect to actions, and what they’re used for—not as mutually exclusive categories.
AI
The instructor says CIS 6280 will cover simulation, robotics, and video and 3D generation, with hands-on work building an environment and training a world model.
AI
One post connects the question of knowledge’s usefulness to the step-by-step automation of knowledge production.
AI
University of Maryland professor argues agents should retain experience across tasks.
AI
Cognitive scientist and LLM critic made the remark in a social post.
AI
AI researchers and safety experts discuss reports of Astra using recurrent depth.
AI
Stanford PhD-led startup building AI agents to simulate human behavior secures major venture backing.
AI
OpenAI reset limits after GPT-5.6 Sol consumed Codex faster than expected.
