Cognitive Search - Azure Search with AI | Azure Friday
Ғылым және технология
Jacob Jedryszek joins Scott Hanselman to talk about about using Cognitive Services with Azure Search with your mobile and web apps. Skip hiring search experts who know what an inverted index is. Don't worry about distributed systems expertise to scale your service to handle large amount of data. And forget about setting up, owning and managing the infrastructure. Let Azure Search do it all for you.
01:00 Creating a search index
02:37 AzSearch.js - Automagical UI and sample React controls
03:27 Searching the JFK Files
08:00 Adding search to Scott's blog
Cognitive Search - Azure Search with AI blog post
aka.ms/azfr/572/01
The JFK Files (microsoft/AzureSearch_JFK_Files repo)
aka.ms/azfr/572/02
AzSearch.js (jj09/AzSearch.js repo) - Automagical UI and sample React controls
aka.ms/azfr/572/03
Azure Search overview
aka.ms/azfr/572/04
Create a free account (Azure)
aka.ms/azfr/572/free
#microsoft #microsoftazure
Пікірлер: 14
Scott: I don't have a week to do that
5:55 oh soo there is a Cuban in here lol... I guess there are two now with me 😂
Need a follow up video on your quest to live in the Medina area
hey where is the Cognitive part!!!!
Awesome stuff!
This is very good. If our goal is to build a chat bot using questions and answers already in our websites. The data is unstructured other than being a question with a pairing answer. Is there a Cognitive Search design where we DO NOT needeto load our data anywhere including a Language Understanding Model. Still need LUIS or can we build sentiment and confidence within cognitive search?
If Azure Cognitive search does so much now also AI but then what "google" does and still the goto search engine for everyone in the world, what's your opinion on this?
The short links are incorrect. They are pointing to AKS links. Please fix. Tx.
How is the autosuggest working? Is it extracting data from blob or some data is fed for autosuggest model? If it is blob data then , it possibly will be giving word by word suggestion. For example, if I type "Aus" it suggests me "Australia", "Austin" etc I choose "Australia" and again I start type " has many sp" it suggests me "has many species of snakes", "has many spiders" etc and I choose "has many spiders". This is how it works?
@JustinAvery
4 жыл бұрын
I'd say initially it bases on the most popular string matches in the dataset. Over time I'm sure it weights the string matches with the likelyhood that people have chosen those options in the past.
@NishaSingh-qf2it
4 жыл бұрын
@@JustinAvery Alright! Thanks for reply. Will it do word-by-word suggestion or one sentence at a time? The example of "Aus" that I mentioned in the question?
@JustinAvery
4 жыл бұрын
@@NishaSingh-qf2it It should be doing a string match, so not word by word. For these kinds of suggestions, in most cases I've come across, the entire query is a single query and not broken up into chunks. When you do the full search it will search on the whole query for most relevant matches, but also match against individual words (although I'm sure this is something that can be tweaked in implementation).
@NishaSingh-qf2it
4 жыл бұрын
@@JustinAvery Thankyou so much for the information. Appreciate it! :)
Muzaffar hussain shakir