AutomationBench: Benchmarking AI Workflow Orchestration
Discover AutomationBench, a new benchmark evaluating AI agents' ability to orchestrate workflows across multiple apps using REST APIs effectively.
Error-Free AI Training for MedMNIST Biomedical Data
Discover how Artificial Special Intelligence enables error-free training on MedMNIST datasets, boosting accuracy in biomedical AI classification tasks.
Verified Patent Analysis Using Dependent Type Theory & AI
Discover a formally verified patent analysis framework combining dependent type theory with AI and Lean 4 for reliable, machine-checkable certificates.
How Adversarial Environments Exploit Agentic AI Flaws
Discover how adversarial environments deceive agentic AI, exposing trust gaps and vulnerabilities in tool-integrated systems.
Neuro-Symbolic Pipeline for Executable Narsese Reasoning
Explore a neuro-symbolic pipeline translating natural language to executable Narsese, enhancing AI reasoning with logic and uncertainty handling.