SNN-MLIR: New Compiler Framework for Deploying Spiking Neural Networks Across Platforms
Researchers have released snn-mlir, an open-source MLIR dialect that compiles spiking neural network (SNN) models from the Neuromorphic Intermediate Representation (NIR) format into self-contained C11 code runnable on any CPU or embedded target. The tool addresses a gap left by NIR, which standardizes SNN model exchange across frameworks like SnnTorch, Lava, and Norse but provides no shared compilation path to actual deployment. By enabling portable, dependency-free deployment of SNNs on resource-constrained hardware, snn-mlir could lower barriers to neuromorphic computing in embedded and edge applications.
Spiking neural networks (SNNs) are trained across a fragmented ecosystem of frameworks, each with its own model format, a problem partially addressed by the Neuromorphic Intermediate Representation (NIR), which provides a common exchange format. However, NIR stops at description and does not provide a compilation pathway to running models on hardware. The newly presented snn-mlir fills this gap as an out-of-tree MLIR dialect that reads NIR files via a Python front end and emits dialect IR, automatically handling quantization rescaling across layers. A reference lowering pass then converts the dialect to standard linalg and arithmetic operations, ultimately producing self-contained, dependency-free C11 code. The dialect is type-polymorphic, supporting both floating-point (f32/f64) and quantized data within a single intermediate representation, serving both simulation and hardware-oriented deployment. The current implementation covers feedforward, fully-connected networks with a CPU backend, and the authors evaluate numerical fidelity, portability, and quantization cost. The project is released under the Apache-2.0 license with LLVM-exception and is publicly available on GitHub.
What's missing
The paper currently supports only feedforward, fully-connected network topologies with a CPU backend; it is unclear when or whether recurrent architectures, convolutional layers, or dedicated neuromorphic hardware backends (e.g., Intel Loihi, BrainScaleS) will be supported. Benchmark comparisons against existing SNN deployment toolchains in terms of runtime performance or code size are not described in the abstract.
What different sources said
- arXiv cs.LGCenter
Spiking Neural Network inference on FPGAs with hls4ml
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.