Researchers Propose Parameter-Efficient Adapter for Multimodal Learning with Tabular and Image Data
Researchers have proposed TI-Adapter, a parameter-efficient framework for adapting pretrained AI models that jointly process structured tabular data and images. The method freezes pretrained encoders and inserts lightweight adapter modules, avoiding the computational cost of full fine-tuning. Experiments across 20 datasets show it matches or outperforms full fine-tuning while using significantly fewer trainable parameters.
A preprint submitted to arXiv introduces the Tabular-Image Adapter (TI-Adapter), a modality-specific fine-tuning framework designed to make multimodal learning with tabular and image data more computationally efficient. The core challenge it addresses is that full fine-tuning of large pretrained encoders is expensive, yet keeping encoders entirely frozen can limit how well a model adapts to specific tasks. TI-Adapter resolves this by freezing the pretrained tabular encoder and inserting a learned adapter after the tabular embedding, while the image branch is adapted using both embedding-level and bottleneck-level adapters. The approach was evaluated on 20 tabular-image benchmark datasets, where it achieved competitive or superior predictive performance compared to full fine-tuning. Ablation studies highlighted that the placement of adapters is a critical design choice for balancing accuracy and efficiency. The work contributes to the growing field of parameter-efficient transfer learning, extending adapter-based methods beyond purely vision or language domains into structured multimodal settings.
What's missing
The preprint has not yet undergone peer review. The paper does not report wall-clock training time or memory benchmarks alongside parameter counts, making it difficult to fully assess practical efficiency gains. Baseline comparisons beyond full fine-tuning (e.g., prompt tuning or LoRA-style methods) are not mentioned in the abstract, leaving open questions about how TI-Adapter ranks among the broader landscape of parameter-efficient techniques.
What different sources said
- arXiv cs.LGCenter
Parameter-Efficient Adapter Tuning for Tabular-Image Multimodal 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.