Advanced RAG with Azure AI Search - top questions answered | BRK100

Поделиться
HTML-код
  • Опубликовано: 23 май 2024
  • During this session, we will answer common questions for vector search, RAG and building GenAI stacks, then show you how to solve them. From configuring a SKU to working with custom embeddings, we will demo how to implement it with Azure AI Search. As we walk through these tutorials, we will share lessons learned along the way, and how it influenced our product roadmap.
    To learn more, please check out these resources:
    * aka.ms/Build24CollectionsAI-A...
    * aka.ms/build-search
    𝗦𝗽𝗲𝗮𝗸𝗲𝗿𝘀:
    * Sumit Bhattacharyya
    * Mike Carter
    * Ronak Chokshi
    * Gia Mondragon
    * K P Rao
    * Farzad Sunavala
    * Vinod Valloppillil
    𝗦𝗲𝘀𝘀𝗶𝗼𝗻 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻:
    This video is one of many sessions delivered for the Microsoft Build 2024 event. View the full session schedule and learn more about Microsoft Build at build.microsoft.com
    BRK100 | English (US) | AI Development
    #MSBuild
  • НаукаНаука

Комментарии • 6

  • @muralim8164
    @muralim8164 26 дней назад +1

    Is it possible to combine structured data (like JSON from a database or API) with unstructured data (like documents and audio) using the RAG pattern? I want to help customer care agents quickly call or email customers to share extra documents based on their structured data (employment type, age, etc.) and existing documents shared by customer. The mandatory document list and rules for each customer type are in a our internal wiki/CRM website. Should we try it with Multitenant Azure search? I still not clear how can I integrate my Json data that can be used to query.

  • @Momentum_Option_Buyer
    @Momentum_Option_Buyer Месяц назад +1

    When is GA for integrated vectorization?

  • @imai-pg3cz
    @imai-pg3cz Месяц назад

    what file format gives the best results? pdf json files or office docs

  • @imai-pg3cz
    @imai-pg3cz Месяц назад

    what kind of file format did CHR to have their bot point to the files on the website?

  • @imai-pg3cz
    @imai-pg3cz Месяц назад +2

    when using RAG, do you have to update the data every time to be up-to-date? is there a way of automatically feeding the RAG with data?

    • @saurabhsensharma
      @saurabhsensharma 29 дней назад

      AI search offers automatic index refreshes down to every 5 mins that can perform “incremental” vectorisation automatically for changes in data sources on Azure Blob Storage. If you want more control on the refreshes example immediate or batch process (say end of day), you can leverage blob storage change notifications (immediate) or use Blob Change feed (for a consistent, ordered, and guaranteed log of changes useful for batch processing changes in your data source). Checkout this session from around the 30th min for more context and live demonstration of how to use change feed. ruclips.net/video/e2zpe5T_Jf0/видео.html
      Hope this helps :)