Chu, Minxing, Cheng, Quan, Li, Anhang, Atilla, Martins, Luo, Junyi, Jiang, Xinting, Qi, Ruichen, Saligane, Mehdi · Nov 19, 2026
arXiv cs.AI: ReLay recovers reusable layout programs from analog designs
Markdown · JSON · Original source
The authors of ReLay argue that analog layout generators—software tools for creating physical layouts of analog circuits—enable design reuse and migration, but cover only what their handcrafted libraries include. Knowledge in finished layouts remains implicit in their geometry. ReLay translates layouts into generator programs and back, the authors say, allowing existing designs to be resized parametrically and moved across PDKs (process design kits: technology-specific design rules and files for chip manufacturing processes). Repeated constructions in recovered programs become library functions, so the library can grow as layouts accumulate. The authors report testing ReLay on a of hand-drawn layouts spanning circuit families and open-source PDKs. In held-out layouts, learned functions account for 53–66% of layout-building calls. They also report that Frontier agents using ReLay had 64–77% lower cost per completed task than agents directly generating geometry, with comparable completion rates and electrical quality. These results suggest a route to reusing layout knowledge rather than rebuilding each generator library by hand, though the reported comparisons do not by themselves establish performance beyond the benchmark and tasks studied. The paper is a preprint under review.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
cs.AI · Article
Source details
Original source date: Nov 19, 2026 (2026-11-19).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-20T07:32:21.955Z. First observed: 2026-11-20T05:07:39.818Z. Retrieved: 2026-11-20T07:18:52.482Z.
Referenced by
Recorded citations in public readings. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The evidence is the publisher-page abstract; it does not provide the methodological detail needed to assess how the , held-out layouts, or task comparisons were constructed. The paper is identified as a preprint under review; the abstract’s reported results should be understood as the authors’ claims, not independently verified findings.
Open question
How well do ReLay’s reported reuse and task-cost results hold beyond the layouts and tasks described in the abstract?
Source-stated
The authors say ReLay recovers generator programs from existing layouts and supports cross-PDK portability and parametric resizing.
We present ReLay, a bidirectional layout-code framework that recovers generator programs from existing layouts, enabling cross-PDK portability and parametric resizing of existing designs.
Source-stated
The authors promote recurring constructions in recovered programs to library functions, allowing the library to evolve as layouts accumulate.
Constructions that recur across recovered programs are promoted to library functions, so the library evolves as layouts accumulate.
Source-stated
The authors report a benchmark comprising 1,050 hand-drawn layouts from 27 circuit families across three open-source PDKs.
1,050 hand-drawn layouts from 27 circuit families across three open-source PDKs
Source-stated
In held-out layouts, learned functions account for 53–66% of layout-building calls, according to the authors.
Learned functions account for 53-66% of layout-building calls in held-out layouts
Source-stated
The authors report Frontier LLM agents using ReLay had 64–77% lower API cost per completed task than direct geometry generation, with comparable completion rates and electrical quality.
Frontier LLM agents use ReLay to produce new layouts at 64-77% lower API cost per completed task than direct geometry generation, with comparable completion rates and electrical quality.
Source and contributor context
Chu, Minxing, Cheng, Quan, Li, Anhang, Atilla, Martins, Luo, Junyi, Jiang, Xinting, Qi, Ruichen, Saligane, Mehdi · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Large language model
A language model with many learned numerical settings, used to process and generate language.
API
An application programming interface: a defined way for software to request data or actions from another system.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Editorial provenance
Version 1. Publication ID: 3f1203f8-b8ab-4f93-8b74-6cd6d567e258.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-20T07:32:21.955Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 48e519faf1d509e4c92e419e0d8071ceeff0bfb12a364a7e5e11fad475a6c2fb. Source text hash: 4319701dcba7ddf2fa648d6487d1127d67d4c62b7461f75be8e2d0da94a1b23c.
Revision and source-check history
Version 1 · 2026-11-20T07:32:21.955Z
arXiv cs.AI: ReLay recovers reusable layout programs from analog designs
The authors of ReLay argue that analog layout generators—software tools for creating physical layouts of analog circuits—enable design reuse and migration, but cover only what their handcrafted libraries include. Knowledge in finished layouts remains implicit in their geometry. ReLay translates layouts into generator programs and back, the authors say, allowing existing designs to be resized parametrically and moved across PDKs (process design kits: technology-specific design rules and files for chip manufacturing processes). Repeated constructions in recovered programs become library functions, so the library can grow as layouts accumulate. The authors report testing ReLay on a of hand-drawn layouts spanning circuit families and open-source PDKs. In held-out layouts, learned functions account for 53–66% of layout-building calls. They also report that Frontier agents using ReLay had 64–77% lower cost per completed task than agents directly generating geometry, with comparable completion rates and electrical quality. These results suggest a route to reusing layout knowledge rather than rebuilding each generator library by hand, though the reported comparisons do not by themselves establish performance beyond the benchmark and tasks studied. The paper is a preprint under review.
Original source
cs.AI · Article
Source details
Original source date: Nov 19, 2026 (2026-11-19).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-20T07:32:21.955Z. First observed: 2026-11-20T05:07:39.818Z. Retrieved: 2026-11-20T07:18:52.482Z.
Limits
The evidence is the publisher-page abstract; it does not provide the methodological detail needed to assess how the , held-out layouts, or task comparisons were constructed. The paper is identified as a preprint under review; the abstract’s reported results should be understood as the authors’ claims, not independently verified findings.
Open question
How well do ReLay’s reported reuse and task-cost results hold beyond the layouts and tasks described in the abstract?
Source-stated
The authors say ReLay recovers generator programs from existing layouts and supports cross-PDK portability and parametric resizing.
We present ReLay, a bidirectional layout-code framework that recovers generator programs from existing layouts, enabling cross-PDK portability and parametric resizing of existing designs.
Source-stated
The authors promote recurring constructions in recovered programs to library functions, allowing the library to evolve as layouts accumulate.
Constructions that recur across recovered programs are promoted to library functions, so the library evolves as layouts accumulate.
Source-stated
The authors report a benchmark comprising 1,050 hand-drawn layouts from 27 circuit families across three open-source PDKs.
1,050 hand-drawn layouts from 27 circuit families across three open-source PDKs
Source-stated
In held-out layouts, learned functions account for 53–66% of layout-building calls, according to the authors.
Learned functions account for 53-66% of layout-building calls in held-out layouts
Source-stated
The authors report Frontier LLM agents using ReLay had 64–77% lower API cost per completed task than direct geometry generation, with comparable completion rates and electrical quality.
Frontier LLM agents use ReLay to produce new layouts at 64-77% lower API cost per completed task than direct geometry generation, with comparable completion rates and electrical quality.
Source and contributor context
Chu, Minxing, Cheng, Quan, Li, Anhang, Atilla, Martins, Luo, Junyi, Jiang, Xinting, Qi, Ruichen, Saligane, Mehdi · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Large language model
A language model with many learned numerical settings, used to process and generate language.
API
An application programming interface: a defined way for software to request data or actions from another system.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Editorial provenance
Version 1. Publication ID: 3f1203f8-b8ab-4f93-8b74-6cd6d567e258.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-20T07:32:21.955Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 48e519faf1d509e4c92e419e0d8071ceeff0bfb12a364a7e5e11fad475a6c2fb. Source text hash: 4319701dcba7ddf2fa648d6487d1127d67d4c62b7461f75be8e2d0da94a1b23c.
passed · 2026-11-20T07:32:21.955Z · anthropic/claude-opus-5.5
Publication: 3f1203f8-b8ab-4f93-8b74-6cd6d567e258
attribution — pass: The reading attributes the framework, benchmark and results to the paper's authors and keeps them as reported claims. The paper is listed under cs.AR with a cs.AI cross-listing, so the 'arXiv cs.AI' label matches the supplied publisher field. It also notes the paper is a preprint under review.
Authors:Minxing Chu, Quan Cheng, Anhang Li, Martins Atilla, Junyi Luo, Xinting Jiang, Ruichen Qi, Mehdi Saligane
summary — pass: The reading leads with the authors' argument and explains analog layout generators and PDKs in plain language. It describes the two-way translation, the growing library, and the reported results. The closing interpretation is hedged and does not overstate. Optional repair: briefly gloss 'held-out' using the source's own phrase 'beyond the learning corpus'. The question raises a real open issue that follows from the abstract.
Learned functions account for 53-66% of layout-building calls in held-out layouts, demonstrating reuse beyond the learning corpus.
date — pass: The originalPublished date of 2026-11-19 matches the arXiv v1 submission date. The feed date of 2026-11-20 is later and was correctly not used.
[Submitted on 19 Nov 2026]
support — pass: Each claim's excerpt appears verbatim in the abstract, and every claim stays within it: the 1,050 layouts, 27 families and three PDKs, the 53–66% reuse figure, and the 64–77% API cost reduction with comparable completion rates and electrical quality. The suggested benefit is framed conditionally, and the limitations correctly note that the evidence is an abstract only.
Frontier LLM agents use ReLay to produce new layouts at 64-77% lower API cost per completed task than direct geometry generation, with comparable completion rates and electrical quality.
media — pass: The selected media is the verified arXiv abstract page, the direct original source. No better source media was found.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA language model with many learned numerical settings, used to process and generate language.
Definition referenceAn application programming interface: a defined way for software to request data or actions from another system.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivarXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA language model with many learned numerical settings, used to process and generate language.
Definition referenceAn application programming interface: a defined way for software to request data or actions from another system.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivarXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition reference