OmniQ platform verifies 59 quantum papers for AI scientists

Nanyang Technological University researchers have subjected 59 quantum computing papers to rigorous verification using OmniQ, a new virtual research environment that transforms published studies into executable tasks. The systematic evaluation of five AI systems powered by large language models and six agent harnesses revealed a fundamental barrier to autonomous scientific discovery: current AI often fails to reproduce published results due to a mismatch between scientific communication and AI requirements. Human expert intervention consistently bridged this gap, demonstrating that missing information can be converted into reusable research experience, and leading the team to develop Residual Agents (RA), a self-improving AI scientist. The work establishes a new paradigm where verified research becomes a reusable capability supporting autonomous discovery. This expert-audited virtual research environment doesn’t simply assess calculations; it converts published studies into executable workflows, an important step toward fully autonomous scientific discovery in a field where both…