Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore a comprehensive 30-minute conference presentation demonstrating practical AI application development using HPE Private Cloud AI through a basketball play classification system. Learn how to build, train, and deploy machine learning models that accurately recognize and categorize basketball plays including pick and roll, isolation, and fast break maneuvers. Discover the complete AI project lifecycle from initial problem identification—specifically detecting player fatigue for an NBA team—through data collection, model training, and deployment on private cloud infrastructure.
Examine the technical implementation process starting with the Slow Fast open-source video classification model and understand the challenges of creating labeled training datasets using manual extraction from YouTube videos with tools like QuickTime and Label Studio. Analyze the model's performance capabilities and limitations when working with small, manually-labeled datasets while maintaining functional accuracy for play recognition.
Understand the deployment strategy using HPE's Machine Learning Inferencing Service (MLIS) to create accessible model endpoints for streamlined video classification workflows. Explore advanced integration possibilities combining play classification with video language models (VLM) to enable natural language querying of video assets, such as searching for specific player actions in particular game situations and time periods.
Witness demonstrations of the platform's Retrieval-Augmented Generation (RAG) capabilities applied to NBA collective bargaining agreement analysis, showcasing how the infrastructure can provide rapid, valuable insights for sports organizations. Gain insights into the practical applications of private cloud AI infrastructure for complex, real-world video analysis and natural language processing tasks in professional sports environments.
