SemiFA: An Agentic Multi-Modal Framework for Autonomous Semiconductor Failure Analysis Report Generation

2026-04-14 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Robotics & Autonomous Systems · Depth: Expert, quick

Summary

SemiFA is an agentic multi-modal framework designed to autonomously generate structured semiconductor failure analysis (FA) reports from inspection images in under one minute. The system employs a four-agent LangGraph pipeline: a DefectDescriber utilizing DINOv2 and LLaVA-1.6 for defect classification and narration, a RootCauseAnalyzer that integrates SECS/GEM equipment telemetry with historical defect data from a Qdrant vector database, a SeverityClassifier for impact assessment, and a RecipeAdvisor for corrective process adjustments. A fifth node compiles the final PDF report. The framework was developed using SemiFA-930, a dataset of 930 annotated semiconductor defect images across nine classes. Its DINOv2-based classifier achieved 92.1% accuracy on 140 validation images, and the full pipeline generates reports in 48 seconds on an NVIDIA A100-SXM4-40 GB GPU. Multi-modal fusion, particularly equipment telemetry, improved root cause reasoning by +0.86 composite points.

Key takeaway

For semiconductor FA engineers seeking to reduce report generation time and improve accuracy, SemiFA demonstrates a viable path to automation. Its integration of SECS/GEM telemetry with visual data significantly enhances root cause analysis, potentially freeing up several hours of expert time per case. You should consider exploring multi-modal agentic frameworks to automate complex, data-intensive analytical tasks in your operations.

Key insights

SemiFA automates semiconductor failure analysis report generation using a multi-modal agentic framework integrating vision and equipment telemetry.

Principles

• Decompose complex tasks into specialized agents.
• Multi-modal fusion enhances reasoning accuracy.
• Equipment telemetry is critical for root cause analysis.

Method

SemiFA uses a LangGraph pipeline with agents for defect description (DINOv2, LLaVA-1.6), root cause analysis (SECS/GEM, Qdrant DB), severity classification, and recipe advising, culminating in automated PDF report generation.

In practice

• Integrate SECS/GEM data for enhanced FA.
• Utilize DINOv2 for defect classification.
• Employ LangGraph for agent orchestration.

Topics

• SemiFA
• Semiconductor Failure Analysis
• Agentic Multi-Modal Framework
• DINOv2
• LLaVA-1.6

Best for: Computer Vision Engineer, Research Scientist, AI Scientist, Machine Learning Engineer, AI Engineer

Related on AIssential

• SemiFA: An Agentic Multi-Modal Framework for Autonomous Semiconductor Failure Analysis Report Generation — SemiFA automates semiconductor failure analysis report generation using a multi-modal agentic framework and a new dataset.
• A Multi-AI-agent Framework Enabling End-to-end Finite Element Analysis for Solid Mechanics Problems — AbaqusAgent uses a multi-LLM-agent framework to automate end-to-end FEA from natural language, achieving 86% success across 50 problems.
• Toward AI VIS Co-Scientists: A General and End-to-End Agent Harness for Solving Complex Data Visualization Tasks — An AI agent harness automates end-to-end visual analysis application design, bridging data to interactive interfaces.
• MOSAIC: Modular Orchestration for Structured Agentic Intelligence and Composition — MOSAIC structures LLM-based agentic model selection through blueprints, improving automated data science performance and traceability.
• Hacking with the Automaton Constraint — Decomposing complex DFA constraints into layered, simpler automata simplifies implementation and debugging.
• Article: Building Hierarchical Agentic RAG Systems: Multi-Modal Reasoning with Autonomous Error Recovery — Hierarchical multi-agent RAG with autonomous error recovery significantly improves accuracy and reduces hallucinations across structured and unstructured enterprise data.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.