Welcome to the AI Platform Tech Community!
There are several exciting things happening in the AI space and across AI platform at Microsoft. Welcome to the AI platform tech community where we would like to share our learnings from our engagements working closely with partners and customers. You will see blog articles providing deep dives, getting started and 'tips and tricks' across various AI technology areas. This community includes members from the AI Platform customer engineering and product teams. Thanks!AI Discord - Community Roundtable #5 - Working With Open LLMs
AI Discord - Community Roundtable #5 - Working With Open LLMs Thanks for everyone who joined our live session with Sharmila and Facundo on working with LLMs. This was part of our series in the AI Community Discord where we host community roundtables every 2 weeks on a Generative AI topic that is voted on by the community. This week we discussed working with Open LLMs including Mistral and Llama2. Our speakers walked us through what they are seeing in terms of how teams are blending OpenAI models + Open LLMs, how to work with these modes in Azure AI Studio and we saw a demo of using Prompt Flow to evaluate different models on one prompt. Recording Here is the recording. (Apologies for audio problems at the start) Additional Resources Explore the model catalog in Azure AI Studio Develop a flow using Prompt Flow Tune prompts using variants
RAG / Vector Database best practices for CoPilot Studio
I have been working with Microsoft Gen AI LLM tools (Azure OpenAI Studio and CoPilot Studio) for building a custom 'agent' for answering questions about a set of company internal documents. It seems like RAG is the best approach and fine-tuning would be overkill. In support of RAG with a vector database I would like to understand best practices. It isn't clear to me if manually uploading files to the CoPilot within CoPilot studio does effective preprocessing of the documents (e.g. tables) and chunking or not. And, if it adds embeddings for words not in the pre-trained LLMs embeddings vocabulary. I am looking for the best practice on an ongoing basis for automating the updates (add,update,delete) to the RAG content on an ongoing basis for multiple additional customized LLMs with different sets of documents. It seems like that leveraging open source technology like "langchain" might be a way to achieve consistent results for LLMs that might be updated on a regular basis with RAG content. Is that advised? Or are there Microsoft tools that might be better for automating content updates? Also, there is a choice of what vector database to use, posgrestsql, Cosmos DB (Mongo), etc. Which are supported and recommended for Copilot studio. Thanks in advance for any guidance!Innovate with Azure AI Studio AMA: Unleashing Generative AI for Enterprise Solutions
Join us for an AMA on Azure AI Studio next week, December 14 at 9AM PT! Check out the link below for more details: Innovate with Azure AI Studio AMA: Unleashing Generative AI for Enterprise Solutions - Microsoft Community Hub
UHRS
Hi all, I'm Alex and new to this space, Im currently working on the uhrs platform, was looking at a way to contact the admin and found through. I have experience in Ai of self driven cars and human robots from a common vendor, Remotasks. Id be greatful if I get a welcome and a small introduction of this space and what you guys talk about...Thanks in advance .Speech driven Personal User Bot for a Blind User
Hello Theoretically, is it possible for a visually impaired User to develop and build a (PUB) on the Azure platform but resides on the Users device which is Windows driven? The PUB (named ALBOT) is solely speech driven and is the preferred IOT channel by the User? The current Windows Cortana feature whilst developing well is not at the level of interactivity for my needs. I have not yet registered for an Azure account - I shall do so when I am confident that my specs are well defined. Any signposting would be very much appreciated. Many thanks from ALBOT1115Attend this IT Showcase webinar to learn how data-driven decision making fuels Microsoft’s IT team
Timing: Wednesday, June 17, 2020 | 10 AM PT / 1 PM ET Register for this webinar at https://aka.ms/DataDrivenDecisionMaking. A lot of companies are moving toward digital transformation, including Microsoft. Jose Prieto, a senior data and applied scientist manager, believes that data is central to the decision-making process at Microsoft. Microsoft’s Core Services Engineering and Operations (CSEO) Data team is helping leaders drive data-driven decision making by giving them access to data-powered frameworks, products, and visualization solutions. These tools enable leaders to track business processes and systems and be more agile and informed in their decision making. “We are a big organization, and the challenges we face are faced by customers as well,” says Rahul Srivastava, a senior data and applied scientist manager in CSEO. “We hope that people will learn from our process.” Srivastava and Prieto, both senior data and applied scientist managers on the Data team in CSEO, will be sharing their insights in an https://aka.ms/DataDrivenDecisionMaking.” “We hope to inspire other enterprises and invite them to learn from our mistakes,” Prieto says. “There is more data than ever, and this roundtable will help you discern what data to focus on and how to use it.” They both hope that customers can learn from how Microsoft is evolving the way it uses data. “We are a big organization, and the challenges we face are faced by customers as well,” Srivastava says. “We hope that people will learn from our process.” [https://aka.ms/DataDrivenDecisionMaking.] [https://www.microsoft.com/en-us/itshowcase/metrics-that-matter-how-we-track-our-digital-transformation.] [https://www.microsoft.com/en-us/itshowcase/modern-data-platform-metrics-that-matter-the-most.] Prieto has worked with teams across Microsoft to identify top metrics that can help them measure engagement, adoption, retention, and task success. Srivastava has been focused on developing frameworks and metrics that enable an assessment of how employees use internal tools and applications and drive data-driven product development. They will share lessons learned from their experience advocating for data-driven decision making in CSEO and provide structure, frameworks, and methodologies on how to think about data. This means doing things like identifying success metrics before starting to build a product to ensure that you end up with the telemetry you need. “You have to think about what data you need to measure, which you can use to improve your processes and products,” Prieto says. Prieto and Srivastava will share some of the challenges they’ve made and how to overcome them, including the importance of bringing people along the decision-making process. “When someone successfully uses data, they become your biggest fan,” Srivastava says. “Starting small and making other people successful by incrementally building solutions is one way to gain leadership support.” https://aka.ms/DataDrivenDecisionMaking. https://www.microsoft.com/en-us/itshowcase/metrics-that-matter-how-we-track-our-digital-transformation. https://www.microsoft.com/en-us/itshowcase/modern-data-platform-metrics-that-matter-the-most. Take a look at the rest of the webinars hosted by IT Showcase: IT expert roundtable: Finance uses anomaly detection for royalty statements processing June 2, 2020 | 10 AM PT / 1 PM ET | https://info.microsoft.com/CO-AI-WBNR-FY20-06Jun-03-Financeusesanomalydetectionforroyaltystatementsprocessing-4861_LP01Registration-ForminBody.html Processing royalty payments at Microsoft requires a high level of accuracy and oversight. During this webinar, you’ll learn how our Finance team used anomaly detection and automation to replace time-consuming and costly manual processes. This new automated solution enhances our Sarbanes-Oxley Act (SOX) requirements and operational controls. User experience drives modern management at Microsoft June 16, 2020 | 10 AM PT / 1 PM ET | https://info.microsoft.com/CO-NOGEP-WBNR-FY20-06Jun-16-UserexperiencedrivesmodernmanagementatMicrosoft-4862_LP01Registration-ForminBody.html In this webinar, we discuss our modern management strategy with a specific look at our approach to delivering efficient devices and seamless access to apps, sites, services, documents, and data. IT expert roundtable: Digital transformation with Modern Data Foundations June 24, 2020 |10 AM PT / 1 PM ET | https://info.microsoft.com/CO-AI-WBNR-FY20-06Jun-24-DigitaltransformationwithModernData-4864_LP01Registration-ForminBody.html In this webinar, our experts will answer questions about how we built Modern Data Foundations to democratize data access at Microsoft and power a new breed of apps and services.
Home decor platform development
My customer is developing a home decor platform where users can upload images of their rooms, apply different decor items like furniture or wall colors, and visualize how they would look. The platform relies on AI tools and APIs for object recognition and placement but is facing issues with accurately identifying the floor and objects. The AI is currently rejecting the floor or placing items incorrectly. They are also working on improving search results based on uploaded images, which show similar but not exact matches, leading to inconsistencies in user experience. They are facing challenges in achieving accuracy and reducing processing time.Deep Fake - what do u think about it ?
Hi, what do u think about deepfake technology ? I found this article https://jelesnianski.com/artificial-intelligence/before-you-believe-how-to-recognize-a-deepfake-and-is-it-inherently-evil/ Do you think that AI is more of a threat or an opportunity for development?
