MPC-Flow: Model Predictive Control Framework for Solving Inverse Problems with Flow-Based Generative Models
Researchers have proposed MPC-Flow, a model predictive control (MPC) framework that enables training-free guidance of flow-based generative models for solving inverse problems such as image inpainting, deblurring, and super-resolution. Existing approaches to conditional generation in flow models require computationally expensive differentiation through model trajectories or adjoint solves, limiting practical use. MPC-Flow addresses this bottleneck and scales to state-of-the-art architectures, including a quantised 32-billion-parameter FLUX model running on consumer hardware.
MPC-Flow, accepted for publication at ICML 2026, reformulates the problem of guided generation in flow-based models as a sequence of smaller optimal control sub-problems rather than a single, computationally prohibitive trajectory optimisation. The framework provides theoretical grounding linking its approach to the underlying optimal control objective and shows that different algorithmic choices within the framework yield a spectrum of guidance strategies, including variants that entirely avoid backpropagation through the generative model. The authors evaluate MPC-Flow on standard image restoration benchmarks covering both linear and non-linear inverse problems, reporting strong performance across tasks. A notable practical result is the demonstrated ability to guide FLUX.2, a 32-billion-parameter model, in a quantised setting on consumer-grade hardware without any task-specific training. The work addresses a key scalability gap in training-free conditional generation, potentially broadening access to powerful flow-based priors for scientific and engineering inverse problems.
What's missing
It is unclear how performance degrades as quantisation level increases on the large FLUX model. The scope of 'consumer hardware' tested is not specified, leaving practical hardware requirements ambiguous.
What different sources said
- arXiv cs.LGCenter
Solving Inverse Problems with Flow-based Models via Model Predictive Control
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.