Kunlun: New Architecture Establishes Scaling Laws for Large-Scale Recommendation Systems
Researchers from Meta have published a paper introducing Kunlun, a new scalable architecture designed to bring predictable scaling laws to large-scale recommendation systems. Unlike large language models, recommendation systems have historically lacked reliable scaling laws due to inefficient modules and poor resource allocation. The work matters because establishing such laws could allow engineers to more predictably allocate compute resources and improve ad recommendation performance at massive scale.
A team of Meta researchers has released a preprint on arXiv describing Kunlun, a unified architecture aimed at deriving scaling laws for recommendation systems comparable to those already established for large language models. The core problem the authors identify is poor scaling efficiency, attributed to low Model FLOPs Utilization (MFU) and suboptimal resource allocation in existing systems. To address this, Kunlun introduces several low-level optimizations — Generalized Dot-Product Attention (GDPA), Hierarchical Seed Pooling (HSP), and Sliding Window Attention — alongside higher-level innovations including Computation Skip (CompSkip) and Event-level Personalization. These changes reportedly increase MFU from 17% to 37% on NVIDIA B200 GPUs and double scaling efficiency compared to prior state-of-the-art methods. The architecture has already been deployed in major Meta Ads models, with the authors claiming significant production impact. The paper is 10 pages with 4 figures and has undergone three revisions since its initial submission in February 2026.
What's missing
The paper does not appear to disclose specific quantitative production metrics (e.g., click-through rate improvements or revenue impact) from the Meta Ads deployment, making independent assessment of real-world gains difficult. Additionally, as a preprint, the work has not yet undergone formal peer review, and the generalizability of the scaling laws to recommendation systems outside Meta's infrastructure remains an open question.
What different sources said
- arXiv cs.AICenter
Kunlun: Establishing Scaling Laws for Massive-Scale Recommendation Systems through Unified Architecture Design
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.