analiza aiqbench
EmoTrack advances depression severity estimation by jointly encoding structured clinical features with sentence-level semantics, addressing the limitations of purely LLM-based or isolated training approaches. The model demonstrates improved accuracy on the DAIC-WOZ benchmark and supports longitudinal assessment through optional memory integration for preceding sessions. This work highlights a practical fusion of clinical knowledge with deep learning representations to enhance interpretability in mental health applications.
EmoTrack: Clinical-Semantic Modeling for Text-Based Depression Severity Estimation
arXiv:2605.22286v2 Announce Type: replace Abstract: Text-based counseling provides a valuable source of information for assessing depression severity. We study prediction of the total score on the eight-item Patient Health Questionnaire (PHQ-8), a self-report measure of depression severity, from counseling transcripts. Clinical-based methods rely mainly on large language model (LLM) inference to obtain structured session-level assessments, but these assessments provide limited information about which utterances support each score. Training-based methods train predictors directly on sentences or their semantic embeddings and preserve local conversational detail, but must learn clinical structure from limited labeled transcripts. Integrating a small set of clinical feature scores with a long sequence of high-dimensional utterance representations is challenging, as simple fusion can overemphasize dialogue evidence and fail to link clinical features to supporting utterances. Combining structured clinical assessments with sentence-level semantics, we propose EmoTrack, which jointly encodes clinical feature scores and utterance representations to predict the total PHQ-8 score. For longitudinal assessment across successive sessions, the model can also incorporate a compressed representation of the preceding session as optional memory. On the real-world DAIC-WOZ benchmark, EmoTrack reduces mean absolute error from 2.8234 to 2.4708 relative to the strongest evaluated baseline. Across five settings covering single-session estimation, cross-dataset transfer without adaptation, and longitudinal assessment, it reduces normalized aggregate error by approximately 6.1% relative to the strongest baseline.
Abstract:Text-based counseling provides a valuable source of information for assessing depression severity. We study prediction of the total score on the eight-item Patient Health Questionnaire (PHQ-8), a self-report measure of depression severity, from counseling transcripts. Clinical-based methods rely mainly on large language model (LLM) inference to obtain structured session-level assessments, but these assessments provide limited information about which utterances support each score. Training-based methods train predictors directly on sentences or their semantic embeddings and preserve local conversational detail, but must learn clinical structure from limited labeled transcripts. Integrating a small set of clinical feature scores with a long sequence of high-dimensional utterance representations is challenging, as simple fusion can overemphasize dialogue evidence and fail to link clinical features to supporting utterances. Combining structured clinical assessments with sentence-level semantics, we propose EmoTrack, which jointly encodes clinical feature scores and utterance representations to predict the total PHQ-8 score. For longitudinal assessment across successive sessions, the model can also incorporate a compressed representation of the preceding session as optional memory. On the real-world DAIC-WOZ benchmark, EmoTrack reduces mean absolute error from 2.8234 to 2.4708 relative to the strongest evaluated baseline. Across five settings covering single-session estimation, cross-dataset transfer without adaptation, and longitudinal assessment, it reduces normalized aggregate error by approximately 6.1% relative to the strongest baseline.
Compared with v1, this version (V2) focuses on the EmoTrack method. The LongCounsel-8 dataset contribution is presented separately in arXiv:2609.03507
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2605.22286 [cs.LG]
(or arXiv:2605.22286v2 [cs.LG] for this version)
Text integral sub licența deschisă a publicației · 247 cuvinte
Subiecte conexe
Interpretable Symptom Vectors for Depression in a Large Language Model
arXiv · 3 sept. 2026
Three Pieces of Evidence that the NLA Reconstructor Does Not Read Semantics or Syntax
LessWrong · 9 oct. 2026
Interpretable Depression Detection from Social Media Text Using LLM-Derived Embeddings
arXiv · 27 iul. 2026
Characterizing Text Branch Sensitivity in Medical Vision-Language Segmentation via Evidence Decoupling
arXiv · 11 sept. 2026