FOGO: New Optimizer Addresses Gradient Interference in Standard and Continual Learning
Researchers introduced FOGO, an optimization algorithm that prevents dominant gradient directions from suppressing useful but rare update directions during neural network training. The method uses spectral orthogonalization and compact memory to detect and resolve gradient conflicts in both standard training and continual learning scenarios. This addresses a fundamental optimization problem that causes both short-term forgetting within training steps and long-term knowledge loss in continual learning settings.
FOGO (Forgetting-aware Orthogonalization Optimizer) is a new scalable optimizer designed to address gradient interference—a phenomenon where dominant mini-batch gradients suppress rare but useful update directions during training. The researchers argue that forgetting is not unique to continual learning but occurs during standard training as well, compounding into long-term knowledge loss when unaddressed. FOGO works by spectrally orthogonalizing momentum updates to prevent dominant directions from monopolizing optimization, while maintaining a compact codebook memory of past directions using random projection. The method resolves conflicts between current and stored update directions through lightweight orthogonal correction with minimal computational overhead and no additional data storage requirements. Experiments across class-imbalanced classification, continual visual learning with domain and class shifts, continual fine-tuning of vision-language models (LLaVA-7B), and language model pretraining (GPT-2) show consistent improvements in convergence and knowledge retention compared to existing optimizers like Adam and Muon.
What's missing
The paper does not discuss computational complexity analysis or wall-clock time comparisons with baseline optimizers, which would be relevant for practitioners evaluating adoption. Additionally, the specific hyperparameter sensitivity of FOGO and guidance for setting the codebook size are not detailed in the abstract.
What different sources said
- arXiv cs.LGCenter
FOGO: Forgetting-aware Orthogonalization Optimizer
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.