LEAF: New Method Improves Speech-Aware Language Model Training Through Better Credit Assignment
Researchers have developed LEAF (Low-rank Exploration with Adaptive Forking), a new training method for speech-aware large language models that improves how credit is assigned to individual tokens during reinforcement learning. The method addresses limitations in current GRPO-style approaches by identifying shared prefixes in response batches and assigning advantages at the span level rather than uniformly across entire responses. LEAF demonstrates improved performance on speech question answering and translation tasks, with smaller models trained using the method outperforming larger baseline models.
LEAF is a retrospective tree-based reinforcement learning method designed to improve post-training of speech-aware large language models. The core innovation addresses a fundamental problem in current GRPO-style methods: they assign the same reward advantage to every token in a response, ignoring useful structural patterns. LEAF recovers this structure by sampling complete responses, identifying high-surprisal decision boundaries, grouping responses by shared prefixes, and assigning advantages at the span level using descendant rewards. The method operates without requiring online branching or additional decoding, making it computationally efficient. Empirical results show improvements over GRPO baselines on speech question answering and translation benchmarks, with notably smaller LEAF-trained models achieving performance comparable to or exceeding current state-of-the-art full-parameter baselines.
What different sources said
- arXiv cs.CLCenter
ParaBridge: Bridging Paralinguistic Perception and Dialogue Behavior in Speech Language Models
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.