In offline validation, the HMM timing model detected children's attention moments more accurately than fixed-interval and dwell-time triggers¶
Claim · All claims
Evidence · 1 study · causalr2·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? Across both proxy-labeling approaches (expert-annotated and future-gaze), the HMM achieved the highest F1 score, mainly due to substantially higher recall than the fixed 3s and 3s dwell baselines. → Zekun Wu 2026
Evidence¶
Zekun Wu 2026¶
Zekun Wu, Man Su, Huiyong Li, Tomohiro Nagashima, and Anna Maria Feit. (2026). Gaze-Informed Proactive AI Assistance for Children Exploring Picture Books. https://arxiv.org/abs/2607.00445 doi:10.48550/arxiv.2607.00445
q2 · i? · causal · r2
Offline validation on free-exploration gaze data from four children across nine images, replayed as 500 ms candidate intervals against two proxy labels (669/285 expert-annotated attention/no-attention intervals; 711/606 future-gaze instances). The HMM reached F1 0.784 (expert) and 0.651 (future-gaze); no effect size printed.
"Across both proxy-labeling approaches, the HMM achieved the highest F1 score, mainly due to substantially higher recall than the two baselines."
Discussion¶
Related Claims¶
- In offline validation, combined relevance-based secondary-AOI selection matched children's next attended AOI more often than random or single-distance selection — related
- Longer dwell time on AI suggestions is associated with better attention check performance but slightly lower task performance
- Speech detection showed preliminary success but remained challenging in noisy settings, particularly for speaker differentiation, children's voices, and diverse dialects
- LLM performance drops from offline to real-time proxy evaluation when clues arrive incrementally — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
hmm-timing-outperforms-fixed-and-dwell-triggers |
title |
In offline validation, the HMM timing model detected children's attention moments more accurately than fixed-interval and dwell-time triggers |
description |
In offline validation, the HMM timing model detected children's attention moments more accurately than fixed-interval and dwell-time triggers |
status |
review |
evidence_strength |
weak |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — zekun-wu-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.