Falcon-X: New Time Series Foundation Model Improves Multivariate Forecasting
Researchers have introduced CITRAS-FM, a 7-million-parameter time series foundation model capable of zero-shot forecasting that incorporates external covariates and runs in under 0.1 seconds on a CPU. Unlike larger existing models, it uses a patch-based decoder-only Transformer with a novel Shifted Attention mechanism to leverage known future covariates during forecasting. The work addresses a gap in deployable, covariate-aware forecasting models suitable for resource-constrained environments.
CITRAS-FM is a lightweight time series foundation model (TSFM) with only 7 million parameters, designed to perform zero-shot forecasting across univariate, multivariate, and covariate-informed settings without requiring task-specific fine-tuning. The model is built on a patch-based, decoder-only Transformer architecture and introduces a technique called Shifted Attention within its cross-variate module to effectively utilize covariates—external variables known throughout the forecast horizon—that influence the target series. A key challenge addressed by the work is the scarcity of covariate-rich training data; to overcome this, the authors propose CovSynth, a method that synthesizes realistic covariates from decomposed components of target time series, enabling covariate-aware pretraining. Evaluated on fev-bench across 100 tasks spanning diverse forecasting settings, CITRAS-FM achieves state-of-the-art zero-shot accuracy among sub-10M parameter TSFMs while delivering sub-0.1-second inference on standard CPUs. The model has been accepted to EUSIPCO 2026, positioning it as a practical solution for real-time forecasting applications where computational resources are limited.
What's missing
The study does not report comparisons against larger TSFMs (beyond sub-10M models) in terms of accuracy trade-offs, leaving open how much performance is sacrificed relative to full-scale models. The generalizability of CovSynth-generated covariates to real-world covariate distributions beyond the fev-bench benchmark is not fully characterized. Long-term robustness and performance on highly non-stationary or domain-specific time series (e.g., medical or financial) remain open questions.
What different sources said
- arXiv cs.LGCenter
CITRAS: Covariate-Informed Transformer for Time Series Forecasting
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.