aiqbench analysis
WELD introduces a critical new benchmark for affective computing by providing the first long-duration, naturalistic workplace dataset that supports both longitudinal individual analysis and team-level relational modeling. This resource addresses significant gaps in existing corpora regarding temporal depth and ecological validity while highlighting specific challenges like demographic bias in facial expression recognition and the limitations of standard survival metrics.
WELD: The First Naturalistic Long-Period Small-Team Workplace Emotion Dataset for Ubiquitous Affective Computing
arXiv:2510.15221v3 Announce Type: replace Abstract: Affective computing has matured rapidly in laboratory settings, yet no prior dataset combines (i) months-to-years of duration, (ii) a naturalistic workplace context, (iii) a stable small-team social structure, and (iv) a fully passive sensing protocol that survives institutional review. We introduce WELD, the first dataset to satisfy all four. WELD comprises 733,780 per-frame seven-class facial-expression probability vectors from 49 employees of a Chinese software company over 30.1 months (Nov 2021 - May 2024) -- the longest naturalistic in-the-wild emotion corpus and the only multi-year corpus supporting both within-individual longitudinal and within-team relational analyses on the same subjects. Data are released under a four-tier access model with only aggregated probabilities publicly downloadable. We validate the corpus by replicating three established phenomena (+43.1% weekend valence boost; 13:00-trough diurnal cycle; Shanghai 2022 lockdown effect d=-0.40), and report four novel findings: (1) variance decomposition attributes 19.3% of daily-valence variance to between-person differences and 29.8% to month seasonality -- a quantitative ceiling for future predictive models; (2) Hidden Markov decomposition reveals six emotional regimes with asymmetric negative-state dwell times (16-18 d vs 3 d); (3) leave-one-person-out turnover prediction reaches AUC=0.79 yet a Cox concordance index of only 0.52, exposing a metric-trap when AUC is reported without survival-aware baselines; (4) the corpus reveals systematic over-prediction of "angry" by an off-the-shelf FER model on neutral Asian faces (0.194 vs ~0.05 Western priors), making WELD valuable for FER fairness audits. A complex-systems analysis of the corpus appears as a companion preprint (arXiv:2510.16046).
Abstract:Affective computing has matured rapidly in laboratory settings, yet no prior dataset combines (i) months-to-years of duration, (ii) a naturalistic workplace context, (iii) a stable small-team social structure, and (iv) a fully passive sensing protocol that survives institutional review. We introduce WELD, the first dataset to satisfy all four. WELD comprises 733,780 per-frame seven-class facial-expression probability vectors from 49 employees of a Chinese software company over 30.1 months (Nov 2021 - May 2024) -- the longest naturalistic in-the-wild emotion corpus and the only multi-year corpus supporting both within-individual longitudinal and within-team relational analyses on the same subjects. Data are released under a four-tier access model with only aggregated probabilities publicly downloadable. We validate the corpus by replicating three established phenomena (+43.1% weekend valence boost; 13:00-trough diurnal cycle; Shanghai 2022 lockdown effect d=-0.40), and report four novel findings: (1) variance decomposition attributes 19.3% of daily-valence variance to between-person differences and 29.8% to month seasonality -- a quantitative ceiling for future predictive models; (2) Hidden Markov decomposition reveals six emotional regimes with asymmetric negative-state dwell times (16-18 d vs 3 d); (3) leave-one-person-out turnover prediction reaches AUC=0.79 yet a Cox concordance index of only 0.52, exposing a metric-trap when AUC is reported without survival-aware baselines; (4) the corpus reveals systematic over-prediction of "angry" by an off-the-shelf FER model on neutral Asian faces (0.194 vs ~0.05 Western priors), making WELD valuable for FER fairness audits. A complex-systems analysis of the corpus appears as a companion preprint (arXiv:2510.16046).
WELD: 733,780 per-frame 7-class facial-expression probability records from 49 employees over 30.1 months (Nov 2021-May 2024). v2 corrects attrition metrics after removing leakage (binary AUC=0.79, survival C-index=0.52) and adds a FER fairness audit. 49 pp, 14 figs, 1 supp PDF
Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (cs.LG)
arXiv:2510.15221 [cs.AI]
(or arXiv:2510.15221v3 [cs.AI] for this version)
Full text under the publisher's open licence · 302 words
Related coverage
State of Pandemic Early Warning
LessWrong · Sep 24, 2026
How your everyday activity feeds Amazon AI training models
Proton Blog · Oct 5, 2026
Brief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident
METR · Aug 26, 2026
Game Design Creativity evals
LessWrong · Oct 12, 2026