How Far Do Foundation Models Transfer to Infant Signals? A Cross-Dataset Transfer Audit with a Unified Need Ontology
2026-08-10 • Computation and Language
Computation and LanguageArtificial Intelligence
AI summaryⓘ
The authors studied several collections of baby crying sounds that are usually small and labeled differently, making it hard to compare or combine them. They found that testing on just one collection hides big differences in performance and even shows label conflicts between collections. By carefully checking and cleaning the data, they showed how combining these collections with consistent labels can improve learning, especially for small and noisy datasets. They also provide tools and a shared labeling system to help future research use multiple collections together more effectively.
infant cry corporalabel incompatibilitymacro-F1 scorecross-corpus transferdeduplication auditfrozen encodersfine-tuningdomain-adaptive pretrainingontology mappingjoint training
Authors
Wu Hangyu
Abstract
Public infant cry corpora are small, label-incompatible, and almost always evaluated one corpus at a time. We ask what this practice hides and what fixes it. Across four cry corpora screened by a multi-level leakage audit (byte-level and embedding-level deduplication plus a within-corpus train-test near-duplicate audit), we probe four frozen encoders and a handcrafted baseline under a unified five-class need ontology and shared task formulations. The audit exposes what single-corpus evaluation conceals: within-domain macro-F1 swings by 0.57-0.80 for the same encoder, cross-corpus transfer is negative on average (negative-transfer ratio 0.19-0.35, significant in 18 of 30 directed cells, BH-FDR), and 349 content-identical clip groups carry conflicting metadata labels across corpus distributions. The same audit, however, reveals a consistent way forward. Transfer into the noisiest corpus is consistently positive in effect size at matched training size and after near-duplicate removal, offering a practical recipe for small, noisy corpora. Frozen probes saturate at modest label budgets, while stabilized fine-tuning wins with full labels; domain-adaptive pretraining significantly beats stabilized fine-tuning at 5-10-shot (the 1-shot advantage is not robust to optimization-seed variance) but shows no significant advantage at 50-shot or beyond. In the tested binary, shared-label settings, ontology-mapped joint training wins in all four encoder-by-target combinations, whereas naively merging unmapped labels costs up to 37 F1 points. We release the ontology, mapping code, and audit pipeline, turning incompatible cry corpora into a usable joint-training resource.