Fastest & easiest way to integrate AI using Windows AI APIs Wednesday, May 21 | 12:00 AM - 1:00 AM Central European Summer Time Duration 1 hour BRK224 On Demand Breakout In Seattle + Online Was recorded Speakers: Profile picture of Rahul Amlekar Rahul Amlekar | Microsoft Profile picture of Lei Xu Lei Xu | Microsoft

Save to favorites By the end of this session, you�ll be able to leverage on-device ML models and walk away with the knowledge on how to elevate your AI experience. We�ll show you how to fine-tune Phi Silica - our own SLM - with your custom data using a LoRA adapter. We will show you how to build innovative AI capabilities like Semantic Search and Knowledge Retrieval in your apps with semantic index, all while following responsible AI practices.

AI Summary Introduction to Key APIs: The session commenced with an emphasis on available APIs such as text summarization, rewrite, image description, and OCR for text recognition. These APIs are stable within the Windows App SDK, enabling developers to build production-level applications efficiently.

Advanced API Features: Two noteworthy experimental APIs were introduced - image object erasure and conversation summary. The training centered on the utilization of the Phi Silica Prompt API and the creation of Laura fine-tuning adapters, both currently in public preview.

Semantic Search and Knowledge Retrieval: These APIs, presently in private preview, enhance search capabilities by allowing semantic understanding and retrieval of data at runtime, critical for AI-driven interactive applications. Interested developers are encouraged to sign up for access to the SDKs.

Integration Demonstrations: Real-world application examples from partnerships, such as with Filmora AI Mate, were shown to illustrate the practical use of these advanced tools in augmenting digital experiences efficiently and effectively.

Developer Tools and Support: The session also highlighted tools developers could leverage, including the AI Dev Gallery for samples and the AI Toolkit for training custom models. Numerous resources were offered to aid developers in utilizing these tools effectively.

Collaborative Success Stories: The presentation concluded with testimonials from various partners who have successfully integrated Windows AI capabilities into their solutions, demonstrating the impactful advancements in local AI processing and cloud collaboration.

About the speakers Profile picture of Rahul Amlekar Rahul Amlekar Senior Product Manager Microsoft Product Manager for Windows Copilot Runtime APIs.

Save to favorites Profile picture of Lei Xu Lei Xu Product Manager Microsoft Lei Xu is a Principal Product Manager Lead at Microsoft, driving efforts to make Windows the best platform for building and running on-device AI. He is passionate about empowering developers to create fast, private, and personalized AI applications.