The Machine Learning FMEA Repository

Background

Repository for the Machine Learning Failure Mode and Effects Analysis (ML FMEA) Template. The ML FMEA is detailed within the SAE World Congress 2025 publication, "Introducing the ML FMEA" by Paul Schmitt, Bodo Seifert, Jerry Lopez, Mario Bijelic, Felix Heide, Krzysztof Pennar

The ML FMEA Template provided here is a practical tool designed to:

• Systematically identify, prioritize, and mitigate risks throughout the ML development pipeline.
• Promote risk transparency and communication between ML development teams and safety experts.
• Tailor known best practices to minimize potential failure points in ML applications.

This repository accompanies the methodology described in our paper, offering a populated ML FMEA template to help development teams manage and communicate risks effectively.

---

Features

• Step-by-step framework: Aligns ML development with PFMEA principles.
• Customizable template: Adaptable to various safety-critical applications.
• Built-in guidance: Includes examples of common failure causes and mitigations at each pipeline step.
• Holistic approach: Considers the ML pipeline as a value-add process, not just the resulting model.

---

How to Use

1. Download the Template:

   • Clone this repository using:

     git clone [repository_url]
     

   • Or download the template directly as a ZIP file.

2. Open the Template:

   • The template is provided in .xlsx format.
   • Ensure you have the required software (e.g., Microsoft Excel, Google Sheets) to open and edit the file.

3. Populate the Template:

   • Follow the pipeline steps outlined in the template.
   • Customize the failure modes, causes, and mitigations to suit your specific ML application.

4. Integrate into Your Workflow:

   • Use the template as part of your development lifecycle to manage and mitigate ML risks systematically.
   • Share the completed analysis with your team and safety experts to ensure transparency.

---

Files in the Repository

• ML_FMEA_Template.xlsx: The core FMEA template with pre-populated examples.
• Examples/: Example applications of the template in different industries (e.g., automotive, robotics).
• Docs/: Supporting documentation and FAQs on using the template.
• Contributions/: Community-contributed improvements and adaptations.

---

Requirements

• Software to open and edit spreadsheets (e.g., Microsoft Excel, Google Sheets).
• Basic understanding of ML pipelines and safety-critical processes.

---

Contributing

We welcome contributions to improve the ML FMEA template! To contribute:

1. Fork the repository.
2. Make your changes.
3. Submit a pull request with a clear description of the update.

---

License

The repository is licensed under the [chosen license, e.g., Apache 2.0/MIT/GNU GPL]. See the LICENSE file for details.

---

Contact

For questions or feedback, please reach out to pauls@massrobotics.org.