Papers
1
Total Citations
21
H-Index
1
About
Kirk Roberts is a leading figure in biomedical natural language processing and clinical informatics, with a focus on extracting structured information from unstructured medical text. His major contributions include advancing multimodal spatial role labeling, which integrates text and image data to understand spatial relationships in clinical contexts—a critical step for improving automated medical reasoning. His work on the CLEF 2017 mSpRL task overview (21 citations) established a benchmark for evaluating systems that combine linguistic and visual cues, driving progress in multimodal AI for healthcare. Beyond this, Roberts has shaped the field through his research on clinical concept extraction, temporal reasoning, and information retrieval from electronic health records, with his publications collectively garnering thousands of citations. He is also recognized for developing shared tasks and resources that enable reproducible research, such as the mSpRL dataset. As an associate professor at the University of Texas Health Science Center at Houston, Roberts continues to mentor the next generation of researchers, bridging the gap between natural language processing and real-world clinical decision support.
Research Focus
Key Achievements
Top Papers
- 1CLEF 2017: Multimodal Spatial Role Labeling (mSpRL) Task Overview21 citations · 2017