GILT: New Graph AI Model Achieves Few-Shot Learning Without Language Models or Fine-Tuning
Researchers have introduced GILT (Graph In-context Learning Transformer), a framework for graph-based machine learning that requires neither Large Language Models nor per-graph tuning to generalize across diverse graph tasks. Current Graph Foundational Models either depend on text-based LLMs—limiting their use with numerical data—or require costly fine-tuning for each new graph, creating efficiency bottlenecks. GILT addresses both limitations simultaneously, achieving stronger few-shot performance with significantly less computational overhead, and has been accepted as an oral presentation at the GFM workshop at ICML 2026.
Graph Neural Networks (GNNs) are widely used for relational data but struggle to generalize to unseen graphs, motivating the development of Graph Foundational Models (GFMs). Existing GFMs face a core challenge: graph datasets are highly heterogeneous, each potentially having unique feature spaces, label sets, and topologies. LLM-based approaches are constrained to text-rich graphs and cannot easily handle numerical features, while structure-based pre-trained models require expensive per-graph fine-tuning before deployment. GILT introduces a unified token-based in-context learning (ICL) framework that reformulates node, edge, and graph-level classification tasks under a single architecture operating on generic numerical features. By dynamically inferring class semantics from context at inference time, GILT eliminates the need for task-specific tuning entirely. Experiments demonstrate that GILT outperforms both LLM-based and tuning-based baselines in few-shot settings while requiring substantially less time, suggesting a practical path toward truly generalizable graph learning. The work is publicly available and was accepted as an oral presentation at the GFM @ ICML 2026 Workshop.
What's missing
It is unclear how GILT scales with very large or highly sparse graphs, and whether the tuning-free property holds across all graph domains or primarily those with numerical features.
What different sources said
- arXiv cs.AICenter
GILT: An LLM-Free, Tuning-Free Graph Foundational Model for In-Context 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.