Study Compares Supervised Learning and In-Context Prompting for Turkish Idiomatic Expression Classification
Researchers evaluated supervised and large language model (LLM) approaches for classifying Turkish idiomatic light verb constructions (LVCs), finding that both methods have distinct strengths depending on prompting strategy. The study tested a fine-tuned Turkish BERT model against three instruction-tuned LLMs under zero-shot, one-shot, and few-shot prompting conditions on a controlled dataset of 147 examples. The findings highlight significant prompt sensitivity in LLMs and suggest that carefully constructed few-shot demonstrations can allow LLMs to match or surpass the supervised baseline.
Accepted to ACL Student Research Workshop 2026, this study addresses the challenge of detecting Turkish idiomatic light verb constructions (LVCs), which are difficult to classify because they share surface forms with literal verb-object combinations. The researchers constructed a controlled evaluation set of 147 examples with matched negatives, including out-of-domain random sentences and in-domain literal controls, to rigorously test classification performance. A supervised BERTurk model with a classifier head served as the baseline, compared against three instruction-tuned LLMs from different families. In zero-shot settings, LLMs performed well at rejecting negatives but showed very low recall for actual LVCs. One-shot prompting substantially improved LVC detection but introduced strong model-specific biases, causing some models to over- or under-predict idioms. Few-shot prompting improved calibration, with GPT-OSS-20B and Qwen 2.5-14B achieving robust overall performance. The results underscore that while the supervised baseline remains competitive, LLMs with well-designed demonstrations can match or exceed it for this metalinguistic classification task.
What's missing
The study's evaluation set is relatively small (N=147), which may limit the statistical reliability of performance comparisons across models. The paper does not report results on naturally occurring, non-controlled Turkish text, leaving open questions about real-world generalizability.
What different sources said
- arXiv cs.CLCenter
Supervision versus Demonstration-Based In-Context Learning for Multiword Expression Classification
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.