SBIR - TOPIC 009 - AI/ML TOOL FOR VISUALIZING BEHAVIORAL AND SOCIAL SCIENCE RESEARCH

"The Hyperthesis team aims to develop an AI/ML-based literature search and visualization tool for Behavioral and Social Science Research (BSR), which mines existing paper and generates aggregate, interactive causal maps of the literature that combine a schematic overview of all the cause-effect relationships examined around a given topic with a detailed representation of each relationship (including its moderating and mediating factors, relationships sign and strength (for meta-analyses papers) and corresponding research papers). This tool will initially be populated with an index of all meta-analyses conducted around Alzheimer’s disease (AD/ADRD), helping users easily discern the impact of different risk/prevention factors on AD/ADRD, and the associations between AD/ADRD and other health and social outcomes (Phase I). In Phase II, the tool’s index will be extended across disciplines and research methodologies, such that it will automatically extract variables and relationships from, experiment-based, observational, and meta-analysis-based papers across BSR disciples, with high accuracy (i.e., precision and recall above .8). Once finalized (in Phase II), the envisioned search and visualization tool will make BSR findings faster to identify, comprehend, and leverage towards new scientific discoveries, thereby accelerating innovation and maximizing the societal value of behavioral research."