Intent classification (also known as intent detection, or intent recognition) is about retrieving the intent from a piece of text. This is especially useful in a discussion (i.e. chatbots and conversational AI), in order to understand what a person wants to achieve.
Performing intent classification with modern AI models like GPT-4, ChatGPT, GPT-J, GPT-NeoX, LLaMA 2, Dolphin, and ChatDolphin, gives very impressive results.
For example, imagine that someone says the following:
Hello
I spent some time on your documentation but I could not figure how to add a new credit card.
It is a problem because my current card is going to expire soon and I'm afraid that it will cause a service disruption.
How can I update my credit card?
Thanks in advance,
Looking forward to hearing from you,
John Doe
The intent behind this email is "update credit card".
GPT-J, GPT-NeoX, LLaMA 2, Dolphin, and ChatDolphin are the most advanced alternatives to GPT-4 and ChatGPT. These models are so big that they can adapt to many situations, and perfectly sounds like a human. For advanced use cases, it is possible to fine-tune GPT-J and Dolphin (train them with your own data), which is a great way to perform intent detection that is perfectly tailored to your company/product/industry.
In a conversation between an AI and a human, it can be very useful to understand what the person is truly looking for or asking for. Here are a couple of examples:
Support teams or sales teams receive tons of customer messages. In order to address them efficiently, it is very useful to first get the intent so one can instantly decide what to do and whether it's important or not.
It's sometimes hard for a user to find what he's looking for, especially if there're a lot of products or if the products are complex. In that case, building a chatbot that detects customers' intents and points them to the right product is a very good solution.
Building an inference API for intent classification with generative models is a necessary step as soon a you want to use intent classification in production. But building such an API is hard... First because you need to code the API (easy part) but also because you need to build a highly available, fast, and scalable infrastructure to serve your models behind the hood (hardest part). It is especially hard for machine learning models as they consume a lot of resources (memory, disk space, CPU, GPU...).
Such an API is interesting because it is completely decoupled from the rest of your stack (microservice architecture), so you can easily scale it independently, and you can access it using any programming language. Most machine learning frameworks are developed in Python, but it's likely that you want to access them from other languages like Javascript, Go, Ruby...
NLP Cloud proposes an intent classification API with generative models that gives you the opportunity to perform detection out of the box, with breathtaking results. If the base generative model is not enough, you can also fine-tune/train GPT-J or Dolphin on NLP Cloud and automatically deploy the new model to production with only one click.
For more details, see our documentation about intent classification with generative models here. For advanced usage, see the text generation API endpoint here. And easily test intent classification on our playground.