Philip Tecklenburg and Oliver Kurz: Machine Learning at Mercedes-Benz скачать в хорошем качестве

Philip Tecklenburg and Oliver Kurz: Machine Learning at Mercedes-Benz

Oliver Kurz and Philip Tecklenburg of Mercedes-Benz discuss the implementation and challenges of using machine learning and computer vision at Scaled ML 2026. Chapters: 0:00 - Introduction and Quick Update from Mercedes-Benz Manufacturing 0:21 - Starting the AI Journey at Mercedes-Benz 1:38 - The Challenge of Implementing AI in Day-to-Day Plant Operations 2:11 - Opportunities: The Data from Digital Twin to Customer Feedback 3:25 - The Goal: Connecting All Data for Faster Development and Quality 4:17 - Challenges: The International Production Network and Regulations 4:45 - Europe’s High Data Protection and Strong Labor Force Challenges 5:26 - The US AI Ecosystem and High Workforce Agility 5:56 - China: Restrictions on US-Based AI 6:43 - Mercedes-Benz Plant in Alabama: Location and Overview 7:13 - Models Produced at the Alabama Plant (SUVs and Electric Vehicles) 7:31 - Celebrating 140 Years of the Automobile Patent 7:50 - The Importance of Innovation and Quality 10:09 - Using Computer Vision to Improve Manufacturing 10:24 - Mercedes-Benz as a Driver of Innovation in a Changing Market 10:51 - Manufacturing Shops and Their Requirements (Body, Paint, Assembly) 11:50 - Assembly Shop: Low Automation to Support High Variation and Individualization 12:46 - Four Layers of Quality Checks 12:53 - Layer 1: Self-Check and Machine Control Data (Adaptive Weld, DC Tools) 13:38 - Layer 2: Quality Gates (Visual Checks) and Computer Vision Potential 14:14 - Layer 3: Sample Checks and Measurement Data 14:36 - Layer 4: Field Feedback and Customer Experience 15:19 - Obstacles: Brownfield Implementation 15:30 - Hardware Implementation and Lighting Challenges 16:28 - Data Processing: Cloud Dependency and On-Site GPU Challenges 16:56 - Product Information Integration and Multiple IT Interfaces 17:47 - System Requirements: Visualization, Fault Feedback, and Running Existing Systems 18:28 - Use Case 1: MIG Weld Inspection in Body Shop 19:22 - Use Case 2: Underbody Fastener Inspection and NVH Topics 20:21 - Measurement System Analysis (MSA) for AI System Buy-Off 21:41 - Use Case 3: VIN Label Verification 22:18 - False Positives: Cost of Downtime ($600–$800 USD per second) 22:34 - False Negatives: Cost of Rework and Customer Experience 23:09 - Main Key Takeaways 24:08 - Computer Vision: A Collaboration with the Human Workforce 24:35 - Thank You