New Reinforcement Learning Method Improves Training Efficiency by Gradually Transferring Control from Baseline Policies
Researchers have proposed sorted Group Policy Optimization (sGPO), a method that reduces wasted compute in reinforcement learning with verifiable rewards (RLVR) training by using a cheap inference pass to assess query difficulty before allocating training resources. Standard RLVR wastes compute on queries that are either too easy or too hard for the current model, generating little to no useful learning signal. sGPO addresses this inefficiency by dynamically filtering and scheduling training data, achieving comparable or better performance at roughly one-third the total compute cost.
Standard RLVR training assigns a fixed rollout budget to every query regardless of difficulty, leading to two failure modes: easy queries yield near-zero advantage because the model already solves them, while unsolvable queries yield no signal because the model never does. sGPO addresses this by running a single offline profiling pass under the initial policy to estimate an empirical success rate per query, which serves as a model-aware proxy for difficulty. This success rate then drives three simultaneous mechanisms: filtering out trivial and unsolvable queries, setting adaptive rollout group sizes inversely proportional to success rate, and constructing an easy-to-hard curriculum. The result is that training compute is concentrated on queries in the productive middle range where the model can learn. The authors report that sGPO matches or exceeds baseline RLVR performance while reducing total training compute—including the upfront inference profiling cost—by a factor of three.
What's missing
The work has not yet undergone peer review.
What different sources said
- arXiv cs.AICenter
sGPO: Trading Inference FLOPs for Training Efficiency in RLVR
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.