IndicDB: Multilingual Text-to-SQL Benchmark for Indian Languages
Discover IndicDB, a benchmark evaluating multilingual Text-to-SQL across Indian languages, addressing real-world data and complex schemas for AI models.
Automated Assessment of Teachers’ Geometric Knowledge
Discover an automated approach to assess teachers' geometric content knowledge using skills-based AI models for scalable and accurate evaluations.
Ordinary Least Squares as a Transformer Special Case
Discover how Ordinary Least Squares is mathematically a special case of Transformer models, revealing new insights into attention mechanisms and memory.
Sim-and-Real Co-Training for Generative Robot Policies
Explore sim-and-real co-training mechanisms that improve generative robot policies using structured alignment and action reweighting techniques.
SafeHarness: Secure Lifecycle Architecture for LLM Agents
SafeHarness integrates multi-layer security into LLM agent lifecycles, reducing attack risks and unsafe behaviors for safer AI deployments.