Arman Cohan
Assistant ProfessorCards
About
Research
Publications
2026
BPI26-028: Hippocrates-Karkinos-o1: An Enhanced Model for Oncology Decision Support and Guideline-Aware Reasoning With Self-Refinement and Guideline-Informed Retrieval.
Aguirre N, Wang B, Schaefer E, Huang J, Li X, Kolamuri S, Ying L, Li X, Ang S, Schneider E, Pai S, Cohan A, Ong C. BPI26-028: Hippocrates-Karkinos-o1: An Enhanced Model for Oncology Decision Support and Guideline-Aware Reasoning With Self-Refinement and Guideline-Informed Retrieval. Journal Of The National Comprehensive Cancer Network 2026, 24 PMID: 42061427, DOI: 10.6004/jnccn.2025.7253.Peer-Reviewed Original ResearchInvestigating Retrieval-Augmented Generation Systems on Unanswerable, Uncheatable, Realistic, Multi-hop Queries
Liu G, Li B, Cohan A, Walden W, Yang E. Investigating Retrieval-Augmented Generation Systems on Unanswerable, Uncheatable, Realistic, Multi-hop Queries. Lecture Notes In Computer Science 2026, 16484: 360-370. DOI: 10.1007/978-3-032-21300-6_26.Chapters
2025
From Scores to Steps: Diagnosing and Improving LLM Performance in Evidence-Based Medical Calculations
Wang B, Xia I, Zhang Y, Wang J, Ouyang F, Han S, Cohan A, Yu H, Yao Z. From Scores to Steps: Diagnosing and Improving LLM Performance in Evidence-Based Medical Calculations. 2025, 2025: 10809-10833. PMID: 41799784, PMCID: PMC12961587, DOI: 10.18653/v1/2025.emnlp-main.548.Peer-Reviewed Original ResearchReal-world medical applicationsEntity extractionCode executionLanguage modelHuman evaluationMedical benchmarksEvaluation pipelineMedical calculatorsArithmetic computationsGranular frameworkReasoning failuresError analysis frameworkAnalysis frameworkBenchmarksExpert judgmentMedical applicationsBenchmarking practicesAccuracyNumerical toleranceFrameworkPerformanceDatasetDecision-makingComputerExecutionRisks of AI scientists: prioritizing safeguarding over autonomy
Tang X, Jin Q, Zhu K, Yuan T, Zhang Y, Zhou W, Qu M, Zhao Y, Tang J, Zhang Z, Cohan A, Greenbaum D, Lu Z, Gerstein M. Risks of AI scientists: prioritizing safeguarding over autonomy. Nature Communications 2025, 16: 8317. PMID: 40968279, PMCID: PMC12446425, DOI: 10.1038/s41467-025-63913-1.Peer-Reviewed Reviews, Practice Guidelines, Standards, and Consensus StatementsA Large-Scale Study of Reranker Relevance Feedback at Inference
Reddy R, Dasigi P, Sultan A, Cohan A, Sil A, Ji H, Hajishirzi H. A Large-Scale Study of Reranker Relevance Feedback at Inference. 2025, 3010-3014. DOI: 10.1145/3726302.3730160.Peer-Reviewed Original ResearchMMVU: Measuring Expert-Level Multi-Discipline Video Understanding
Zhao Y, Zhang H, Xie L, Hu T, Gan G, Long Y, Hu Z, Chen W, Li C, Xu Z, Wang C, Shangguan Z, Liang Z, Liu Y, Zhao C, Cohan A. MMVU: Measuring Expert-Level Multi-Discipline Video Understanding. 2025, 00: 8475-8489. DOI: 10.1109/cvpr52734.2025.00793.Peer-Reviewed Original ResearchVideo understandingExpert levelRelevant domain knowledgeIn-depth error analysisDomain-specific knowledgeVideo benchmarksDomain knowledgeHuman expertsHuman expertiseVideoSpecialized domainsVisual perceptionBenchmarksIn-depth analysisHigh performanceError analysisData quality controlHigh qualityCase studyDatasetTest modelDataFoundation modelModelKnowledgeRouterRetriever: Routing over a Mixture of Expert Embedding Models
Lee H, Soldaini L, Cohan A, Seo M, Lo K. RouterRetriever: Routing over a Mixture of Expert Embedding Models. Proceedings Of The AAAI Conference On Artificial Intelligence 2025, 39: 11995-12003. DOI: 10.1609/aaai.v39i11.33306.Peer-Reviewed Original ResearchEmbedding modelRouting mechanismGeneral domain datasetsMulti-task trainingDomain-specific dataInformation retrieval methodsMulti-task modelDomain-specific expertsExpert retrievalInformation retrievalLanguage modelRouting techniquesRetrieval modelUnderperforming modelsRetrieval methodRetrievalSpecialized domainsDatasetGeneration researchExpertsQueryInformationLanguageTrainingEmbeddingmFollowIR: A Multilingual Benchmark for Instruction Following in Retrieval
Weller O, Chang B, Yang E, Yarmohammadi M, Barham S, MacAvaney S, Cohan A, Soldaini L, Van Durme B, Lawrie D. mFollowIR: A Multilingual Benchmark for Instruction Following in Retrieval. Lecture Notes In Computer Science 2025, 15573: 295-310. DOI: 10.1007/978-3-031-88711-6_19.ChaptersFollowIR: Evaluating and Teaching Information Retrieval Models to Follow Instructions
Weller O, Chang B, MacAvaney S, Lo K, Cohan A, Van Durme B, Lawrie D, Soldaini L. FollowIR: Evaluating and Teaching Information Retrieval Models to Follow Instructions. 2025, 11926-11942. DOI: 10.18653/v1/2025.naacl-long.597.Peer-Reviewed Original ResearchReIFE: Re-evaluating Instruction-Following Evaluation
Liu Y, Shi K, Fabbri A, Zhao Y, Wang P, Wu C, Joty S, Cohan A. ReIFE: Re-evaluating Instruction-Following Evaluation. 2025, 12247-12287. DOI: 10.18653/v1/2025.naacl-long.610.Peer-Reviewed Original Research
News
News
Get In Touch
Contacts
Email