Aws Albarghouthi
2 papers · Latest:
Artificial Intelligence
U-Define: Designing User Workflows for Hard and Soft Constraints in LLM-Based Planning
U-Define helps users control LLM-generated plans by categorizing natural language constraints as hard rules or soft preferences, improving reliability and satisfaction.
2605.02765
Machine LearningTest-Time Scaling Makes Overtraining Compute-Optimal
New $T^2$ scaling laws show that overtraining LLMs is compute-optimal when considering test-time inference costs, leading to stronger performance.
2604.01411
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.