Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Causal AI For AMS Circuit Design: Interpretable Parameter Effects Analysis

Mohyeu Hussain, David Koblah, Reiner Dizon-Paradis, Domenic Forte

Abstract

Analog-mixed-signal (AMS) circuits are highly non-linear and operate on continuous real-world signals, making them far more difficult to model with data-driven AI than digital blocks. To close the gap between structured design data (device dimensions, bias voltages, etc.) and real-world performance, we propose a causal-inference framework that first discovers a directed-acyclic graph (DAG) from SPICE simulation data and then quantifies parameter impact through Average Treatment Effect (ATE) estimation. The approach yields human-interpretable rankings of design knobs and explicit 'what-if' predictions, enabling designers to understand trade-offs in sizing and topology. We evaluate the pipeline on three operational-amplifier families (OTA, telescopic, and folded-cascode) implemented in TSMC 65nm and benchmark it against a baseline neural-network (NN) regressor. Across all circuits the causal model reproduces simulation-based ATEs with an average absolute error of less than 25%, whereas the neural network deviates by more than 80% and frequently predicts the wrong sign. These results demonstrate that causal AI provides both higher accuracy and explainability, paving the way for more efficient, trustworthy AMS design automation.

Causal AI For AMS Circuit Design: Interpretable Parameter Effects Analysis

Abstract

Analog-mixed-signal (AMS) circuits are highly non-linear and operate on continuous real-world signals, making them far more difficult to model with data-driven AI than digital blocks. To close the gap between structured design data (device dimensions, bias voltages, etc.) and real-world performance, we propose a causal-inference framework that first discovers a directed-acyclic graph (DAG) from SPICE simulation data and then quantifies parameter impact through Average Treatment Effect (ATE) estimation. The approach yields human-interpretable rankings of design knobs and explicit 'what-if' predictions, enabling designers to understand trade-offs in sizing and topology. We evaluate the pipeline on three operational-amplifier families (OTA, telescopic, and folded-cascode) implemented in TSMC 65nm and benchmark it against a baseline neural-network (NN) regressor. Across all circuits the causal model reproduces simulation-based ATEs with an average absolute error of less than 25%, whereas the neural network deviates by more than 80% and frequently predicts the wrong sign. These results demonstrate that causal AI provides both higher accuracy and explainability, paving the way for more efficient, trustworthy AMS design automation.

Paper Structure

This paper contains 11 sections, 3 equations, 8 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Explainable Causal Modeling for Analog Circuits
  4. Average Treatment Effect (ATE)
  5. Proposed Causal Analysis Methodology
  6. Results and Discussions
  7. Operational Transconductance Amplifier (OTA)
  8. Telescopic Op Amp
  9. Folded Cascode Op Amp
  10. Discussion
  11. Conclusion and Future Work

Figures (8)

  • Figure 1: Schematic of OTA with current mirror.
  • Figure 2: Overview of the proposed methodology. Preprocessed simulation data are used to train a causal model and estimate causal effects using YLearn (red, training stage). The learned effects are then applied to held‑out test data and evaluated in a Python pipeline (green, testing stage) to produce a ranked list of influential parameters.
  • Figure 3: Causal discovery graph of OTA.
  • Figure 4: Comparison of ATE values for OTA.
  • Figure 5: Schematic of a single-ended telescopic op amp.
  • ...and 3 more figures