New Fine-Tuning Method Improves Chinese Scholarly Text Classification with Imbalanced Data
Researchers have proposed a novel large language model (LLM)-based oversampling technique designed to generate more diverse synthetic minority-class samples for imbalanced classification tasks. Existing methods like SMOTE struggle with categorical data and limited diversity, while prior LLM approaches improved data fidelity but still produced insufficiently varied samples. The new method outperforms eight state-of-the-art baselines across ten tabular datasets, suggesting meaningful practical gains for machine learning applications where class imbalance is a persistent challenge.
A research team has introduced a new oversampling framework that leverages fine-tuned large language models to address the longstanding problem of imbalanced classification in tabular data. Unlike traditional methods such as SMOTE, which require converting categorical variables into numerical vectors and risk information loss, the proposed approach operates directly on structured text representations. The method incorporates three key innovations: a sampling strategy conditioned on both minority class labels and features, a novel permutation strategy for fine-tuning pre-trained LLMs, and training on interpolated samples in addition to minority samples to further increase variability. The authors provide theoretical backing through an entropy-based analysis, formally demonstrating that the method encourages diversity in generated outputs. Experiments across ten tabular datasets show statistically significant improvements over eight competing baselines, with synthetic samples evaluated as both realistic and diverse. The work was submitted to arXiv in October 2025 and revised in June 2026, and has not yet undergone formal peer review.
What's missing
The study has not yet been peer-reviewed, as it is a preprint hosted on arXiv. The paper does not report computational cost or inference time comparisons against baselines, which are relevant for practical deployment. It is also unclear how the method performs on non-English or highly domain-specific tabular datasets, or whether gains hold under varying degrees of class imbalance ratios beyond those tested.
What different sources said
- arXiv cs.AICenter
Large Language Models for Imbalanced Classification: Diversity makes the difference
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.