A contrastive clustering model with a stick-breaking prior and a participant-identity loss sorts daily environment photos into distinct settings and broader environment types, then links clusters to smoking and mood outcomes.
Characterizing the obesogenic environment: the state of the evidence with directions for future research
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Infinite hierarchical contrastive clustering for personal digital envirotyping
A contrastive clustering model with a stick-breaking prior and a participant-identity loss sorts daily environment photos into distinct settings and broader environment types, then links clusters to smoking and mood outcomes.