E-Verify: A Paradigm Shift to Scalable Embedding-based Factuality Verification
Decomposition, embedding and interaction for scalable factuality verification.
RESEARCH
Papers and public research
Selected work involving the team, with original titles, publication status and accessible sources.
Vibe Calibration: Autonomous Bring-up of a 112-Qubit Superconducting Quantum Processor by a Skill-Orchestrating Language Agent
Executable Skills for language-agent orchestration of quantum calibration.
Cited in a joint MIT–OpenAI technical case study. See reference [4]
CrystalFormer-CSP: Thinking Fast and Slow for Crystal Structure Prediction
Crystal structure prediction conditioned on composition, combining generation and search.
Universal power-efficiency trade-off in battery charging
A finite-time thermodynamic study of charging power and efficiency.
Gated Residual Discrepancy for Efficient Bi-Encoder Factuality Verification
Residual discrepancy modeling and reusable evidence representations for bi-encoder verification.
BeyondAtoms: Benchmarking and Verifying Relational Factuality Beyond Atomic Facts
Evaluating relationships between atomic facts against supporting evidence.
Chromosome-scale genomes of wild and cultivated Morinda officinalis
Chromosome-scale genome data for wild and cultivated Morinda officinalis.
Includes work from different research collaborations. EMNLP 2026 acceptance status is based on team-provided records; downloads are not offered for entries without a public source.