5

Is it true Azure Luis only support up to 500 intents per application? https://learn.microsoft.com/en-us/azure/cognitive-services/luis/luis-limits

My requirement is more than 1000 intents. How can I use Luis to do that?

KN B
  • 53
  • 3
  • 1
    A few questions. What do you need so many intents for? If you are assigning intents for Q&A, you could have a single Q&A intent and then let QnA Maker (or KB of your choice) find the specific question. You can even use prompts for multi-part or branching answers. If you really do need all of these intents, can you break them down or categorize them somehow? The dispatch model can have your main LUIS app send queries to the correct downstream model. So technically you could direct queries the 499 downstream apps, which each could have 499 intents. But it would help to understand the context. – billoverton May 13 '20 at 17:27
  • Let say I have list of intents like these: [A, B, C, (A,B), (B,C), (A,B,C), D, E, F, (D,E), ...]. User can talk about A only, or maybe (A, B) in one sentence. So I would like to pin point the exact intent user talk about before I can ask the next question That is why, it grows a lot. Any idea how can I solve it efficiently ? Or any other technology should solve it ? – KN B May 14 '20 at 05:22
  • Are you able to share some of these utterances and combinations? It might help us come up with a better solution if we can understand what types of questions/intents you are dealing with. – billoverton May 14 '20 at 19:30
  • I am still evaluating all the answers. TQ for the help. Very appreciate. – KN B May 21 '20 at 02:03

2 Answers2

1

To answer your question, yes it is true. But it is plenty. We have some of the biggest company using our platfomr to test their training data on LUIS/WATSON/DF... and it is extremely rare to pass the 500 and still get top performance. We typically advice anyway to fine tune your training data for 200 or 300 intents max and if you have more, look into a Model controller architecture with several slave (specific) models

So you might be sure you have 1000 intents, can you reduce it with Entity?

Benoit Alvarez
  • 919
  • 6
  • 3
  • Let say I have list of intents like these: [A, B, C, (A,B), (B,C), (A,B,C), D, E, F, (D,E), ...]. User can talk about A only, or maybe (A, B) in one sentence. So I would like to pin point the exact intent user talk about before I can ask the next question That is why, it grows a lot. Any idea how can I solve it efficiently ? Or any other technology should solve it ? – KN B May 14 '20 at 05:21
  • 1
    i see. I am not sure it is the best strategy. what we do when we have multiple subject in one question is that we split the question on "and", ", ", "Also", etc... and treat each question separately. If you don't, you will create massive overlap in your intent meaning and never be able to get a neat and clear definition for each intent – Benoit Alvarez May 14 '20 at 16:26
1

You should consider using Dispatch. It is a tool that was designed specifically for managing multiple LUIS models and/or QnA Maker knowledge bases that a bot needs to access.

You can find C#, Javascript, and Python samples on the BotBuilder-Samples repo, for reference, titled "14.nlp-with-dispatch".

In your case, this tool is provides a means for overcoming LUIS intent limitations by allowing you to create multiple models to draw from. Dispatch negotiates these models by creating a single LUIS app that then routes the requests to the appropriate model.

Hope of help!

Steven Kanberg
  • 6,078
  • 2
  • 16
  • 35