New Analytic Bijections Improve Normalizing Flows with Better Smoothness and Interpretability
Researchers have introduced three families of analytic bijections for normalizing flows that are globally smooth, defined on all real numbers, and analytically invertible in closed form. The work addresses longstanding trade-offs in existing approaches—affine transforms lack expressivity, splines are only piecewise smooth, and residual flows require numerical inversion. The method, accepted at ICML 2026, also introduces a novel radial flow architecture that achieves comparable quality to coupling flows with 1,000 times fewer parameters on radially structured targets.
A team of researchers has developed three new families of analytic bijections designed to overcome fundamental limitations in normalizing flow models, a class of generative models used in machine learning. Unlike affine transformations, which are smooth but insufficiently expressive, or monotonic splines, which offer local control but are only piecewise smooth and bounded, the proposed bijections are globally smooth (C∞), defined across all of ℝ, and admit closed-form analytic inverses. In addition to serving as drop-in replacements in standard coupling flow architectures—where they match or exceed spline-based performance—the authors introduce radial flows, a novel architecture that transforms the radial coordinate of inputs while preserving angular direction. Radial flows demonstrate exceptional training stability, geometrically interpretable transformations, and dramatic parameter efficiency on targets with radial structure. The approach is validated on 1D and 2D benchmarks and applied to higher-dimensional physics problems, specifically φ⁴ lattice field theory, where the bijections outperform affine baselines and help address mode collapse. The paper, which includes new CIFAR-10 and tabular data results, has been accepted as a final version at ICML 2026.
What's missing
The paper does not fully characterize failure modes outside radially structured or physics-motivated targets. Scalability to very high-dimensional general-purpose generative modeling (e.g., image synthesis beyond CIFAR-10) remains an open question.
What different sources said
- arXiv cs.LGCenter
Analytic Bijections for Smooth and Interpretable Normalizing Flows
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.