AttentionCap: Transformer Model Improves Capacitance Extraction for Advanced Chip Design
Researchers have proposed AttentionCap, a Transformer-based deep learning model for capacitance matrix extraction in integrated circuit design, accepted at DAC 2026. The model addresses limitations of existing MLP- and CNN-based approaches by supporting multiple metal-layer combinations and multiple process nodes within a single framework. It achieves significantly lower error rates and faster inference than prior methods, with strong transferability to new chip fabrication nodes using minimal additional data.
AttentionCap is a customized Transformer architecture designed to learn capacitance matrices for full-chip parasitic extraction, a critical step in electronic design automation (EDA). Unlike prior rule-based or CNN/MLP-based methods that are constrained to fixed metal-layer configurations and specific process nodes, AttentionCap introduces a Gram representation framework, a physics-aligned symmetric-attention output layer, and a normalized Laplacian loss function to better capture the structural properties of capacitance matrices. A process-node embedding further enables the model to generalize across multiple fabrication nodes simultaneously. Trained on synthetic data, the model achieves 0.67% self-capacitance error and 3.99% coupling-capacitance error on unseen real designs, representing 4.6× and 5.7× improvements over the CNN-Cap baseline, respectively, along with 192× faster inference. The pretrained model can be fine-tuned to an entirely new process node using only 5,000 samples and 4,000 training steps, demonstrating strong practical transferability. The work has been accepted at the 63rd ACM/IEEE Design Automation Conference (DAC 2026), and code and data are publicly available.
What's missing
The study relies on synthetic training data, and while results on unseen real designs are reported, the degree to which synthetic-to-real distribution shift may affect performance in full production EDA flows is not fully characterized. Additionally, comparisons are limited to a CNN-Cap baseline; broader benchmarking against other recent deep-learning extraction methods is not discussed.
What different sources said
- arXiv cs.LGCenter
AttentionCap: Transformer Based Capacitance Matrix Learning Toward Full-Chip Extraction
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.