Training-Free Conformal Interval Emerges as Strong Baseline for Probabilistic Time-Series Forecasting
A new preprint on arXiv demonstrates that a parameter-free, training-free conformal prediction interval outperforms many established probabilistic forecasting methods across thousands of real-world time series. The method, called ConformalNaive, wraps a last-value point forecast in a split-conformal residual quantile and requires no model training whatsoever. The findings suggest that the field may be overstating the gains of learned forecasters by comparing them against insufficiently strong baselines.
Researchers have published a preprint arguing that a trivially simple conformal prediction interval — requiring no parameters and no training — constitutes a far stronger baseline than its near-total absence from the literature implies. Tested across 2,217 real time series drawn from nine public datasets including Monash, LOTSA, METR-LA, and others, the ConformalNaive method outperformed naive value-quantile baselines, the entire NPTS family (beating SeasonalNPTS on 64% of series), and the published Conformal Seasonal Pools (CSP) method on 71% of series (bootstrap 95% CI [69, 73]; paired Wilcoxon p ≈ 7.6e-135). It also matched simpler learned conformal predictors such as RCI and quantile regression within 2% relative Winkler score. Notably, ConformalNaive achieved 84–85% empirical coverage at a nominal 95% level on six benchmark datasets, compared to only 66% for the trained neural forecaster DeepNPTS, indicating better calibration. The paper identifies an important boundary condition: at multi-step seasonal horizons, the random-walk floor becomes the weakest method and seasonal pooling wins, motivating the authors' proposed ConformalNaive+, a one-line horizon-adaptive selector that recovers the better of two complementary floors. The authors conclude that a matching conformal naive floor must be treated as a mandatory baseline whenever a learned probabilistic forecaster claims performance gains.
What's missing
The study is a preprint and has not yet undergone peer review. The paper does not address computational latency or real-time deployment constraints, which may be relevant when comparing training-free methods to adaptive-online methods like SPCI, ACI, and AgACI that lead by 9–33% relative Winkler score.
What different sources said
- arXiv cs.LGCenter
Report the Floor: A Training-Free Conformal Interval Is a Mandatory Baseline for Probabilistic 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.