Automated Pronunciation Evaluation System Developed for Korean Toddler Speech
Researchers have developed an end-to-end automated system for evaluating pronunciation in Korean toddlers aged 2–5, combining neural speaker diarization with self-supervised learning models. The system addresses a gap in tools for Korean pediatric speech assessment, where speech sound disorders account for roughly 44% of communication disorder cases. If validated at scale, such tools could enable earlier, more accessible screening for speech disorders in young children.
A team of researchers has proposed an automated pronunciation evaluation pipeline for Korean toddler speech, to be presented at IEEE ICTs4ehealth in June 2026. The system integrates neural speaker diarization—used to separate child speech from caregiver speech in recordings—with self-supervised learning (SSL) models for pronunciation scoring. A newly created IRB-approved corpus of 53 recordings from children aged 2–5 was annotated by three independent reviewers, producing 1,190 consonant and 748 vowel binary correctness labels. Among three diarization models tested, NeMo SortFormer performed best, achieving 88.69% speaker count accuracy and a 33.04% diarization error rate, partly by handling the acoustic similarity between toddler voices and young female caregivers using exaggerated speech styles. For pronunciation scoring, a cross-model ensemble routing consonant prediction to HuBERT-large and vowel prediction to WavLM-large achieved balanced accuracies of 0.720 and 0.845, respectively, with a combined mean of 0.782. The work represents a step toward scalable, automated clinical tools for early speech disorder detection in Korean-speaking children.
What's missing
The corpus is notably small (53 recordings), and inter-annotator agreement statistics for the three reviewers are not reported in the abstract, leaving the reliability of the ground-truth labels unclear.
What different sources said
- arXiv cs.AICenter
Automated Pronunciation Evaluation for Korean Toddler Speech using Speech Diarization and Self-Supervised Learning
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.