AutoSizer : Automatic Sizing of Analog and Mixed-Signal Circuits via Large Language Model (LLM) Agents

Xi Yu, Dmitrii Torbunov, Soumyajit Mandal, Yihui Ren · 2026

Read AutoSizer : Automatic Sizing of Analog and Mixed-Signal Circuits via Large Language Model (LLM) Agents by Xi Yu, Dmitrii Torbunov, Soumyajit Mandal, Yihui Ren (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-02849.txt.

Description

Sizing transistors for analog circuits is slow expert work, and Xi Yu and 3 co-authors try to automate it in a 2026 paper. Their system, AutoSizer, uses large language model agents inside a reflective loop. Language models reason well but are poor at precise number tuning, so the design splits the jobs.

An inner loop runs the sizing search. An outer loop studies progress and reshapes the search space using simulation feedback. The team also built a public benchmark of twenty-four circuits. AutoSizer reached better solutions faster, and more often, than traditional optimizers and earlier language-model agents across circuits of varying difficulty.

Catalog record

CREATORXi Yu, Dmitrii Torbunov, Soumyajit Mandal, Yihui Ren
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE85 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-02849

Browse the full library · About the Tower of Babel