Keysight Technologies, Inc. has unveiled its new Machine Learning Toolkit as part of the latest Keysight Device Modeling Software Suite, aiming to drastically reduce model development and extraction time. By cutting what previously took weeks into mere hours, the toolkit accelerates Process Design Kit (PDK) delivery and enhances Design Technology Co-Optimization (DTCO) applications.

The semiconductor industry is undergoing rapid transformation, fueled by advanced architectures like gate-all-around (GAA) transistors, wide-bandgap materials such as GaN and SiC, and heterogeneous integration strategies including chiplets and 3D stacking. While these innovations boost performance, they also create intricate challenges in modeling and parameter extraction. Traditional workflows depend on physics-based compact models and labor-intensive manual parameter extraction, requiring engineers to tweak hundreds of interconnected parameters under multiple operating conditions. This process can take weeks and often struggles to produce optimal results. With increasingly tight timelines, semiconductor teams now require faster, predictive, and automated AI/ML-driven modeling solutions.

AI Authority Trend: Keysight Launches AI Software Integrity Builder to Enhance AI Trust and Safety

Keysight’s Machine Learning Toolkit addresses these challenges by integrating advanced neural network architectures with ML-based optimization. The toolkit’s auto-extraction capabilities reduce parameter extraction steps from over 200 to fewer than 10, enabling faster PDK delivery, automating DTCO workflows, and accelerating time-to-market.

Key Features and Benefits Include:

  • Accelerated Parameter Extraction: The toolkit transforms hundreds of manual steps into 5–6 automated steps, optimizing over 80 parameters in a single run while capturing secondary effects, temperature variations, and dynamic behaviors. This significantly improves predictive accuracy across DC, RF, and large-signal domains.
  • Automated Workflow: Seamless integration with Keysight’s Device Modeling platform allows Python-based customization and robust automated modeling flows.
  • Scalable Across Technologies: Workflows adapt to FinFET, GAA, GaN, SiC, and bipolar devices, ensuring reusable and repeatable flows for multiple process nodes.
  • Enhanced DTCO Efficiency: Faster feedback between device and circuit design reduces PDK development cycles from weeks to days.

Nilesh Kamdar, General Manager of Keysight EDA, emphasized: “AI/ML is fundamentally transforming the traditional workflows and methodologies of compact modeling. With the new Machine Learning Toolkit, we empower our customers to deliver more predictive, higher-quality models in significantly less time accelerating PDK development and helping them keep pace with rapidly evolving semiconductor technologies.”

AI Authority Trend: Nokia and Keysight Achieve Breakthrough in AI-Ready Data Center Networking

By leveraging AI/ML-driven modeling, Keysight enables semiconductor companies to accelerate innovation, reduce development risk, and maintain a competitive edge in a fast-evolving market.

Additional Enhancements Across Keysight Solutions:

  • Device Modeling MQA 2026: Added rules for Aging Model QA for OMI and MOSRA.
  • Device Modeling WaferPro 2025: Introduced remote-control features for low-frequency noise testing with A-LFNA.
  • A-LFNA 2026: Added low-frequency noise stress test capability for seamless stress-to-noise testing.

AI Authority Trend: Keysight Unveils 3GPP AI Simulation Platform for AI-Enabled 6G Communications

To share your insights, please write to us at info@intentamplify.com