Researchers have developed a Factorized Inverse Decision Model (FIDM) that analyzes verbalized cognitive tasks by decomposing task-execution likelihood into action and effort factors. This model uses a language model to generate structured task-execution traces from raw verbal transcripts, allowing for more nuanced inference than action-trajectory-only methods. When applied to a grocery-shopping dialog task involving 400 older adults for cognitive screening, FIDM demonstrated the ability to selectively estimate these factors and provided information complementary to existing clinical scores and language representations, showing consistent gains in binary classification. AI
IMPACT This model could improve the accuracy and depth of cognitive assessments by leveraging language models to analyze verbal behavior.
RANK_REASON The item describes a new scientific paper proposing a novel model for cognitive analysis. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
- Cognitive Screening and Cognitive Training in Seniors
- cognitive-status classification
- Factorized Inverse Decision Model
- Fashion Institute of Design & Merchandising
- grocery-shopping dialog task
- language model
- Older adults’ experiences of ageing, sex and HIV infection in rural Malawi
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
