An intelligent virtual assistant or chatbot’s “dialogue goals” are predefined actions or goals that a virtual assistant can perform in response to user input during a conversation. Intents help the virtual assistant understand the user’s request or question and provide an appropriate response or action.
Intents can be defined and customized based on the specific needs of the virtual assistant and the types of interactions intended for it. Intent accuracy and efficiency are critical to ensuring a satisfactory user experience and achieving the goals of a virtual assistant.
In a banking environment, dialogue goals for an intelligent virtual assistant might include helping customers open accounts, providing information about loan interest rates, and helping customers transfer money.
What are dialogue intentions and how do you train them?
Dialog goals are a few key words that describe what the user wants the chatbot to do, usually a verb and a noun, such as “Find an ATM,” “Create an event,” and similar actions. The more quality data, the better.
The process of training the dialogue targets involves training a machine learning engine[MLE] as well as the Basic Meaning Engine.[FM]
Select Training
Step one
You start with the Electronics E-commerce Virtual Assistant, then under the Natural Language tab, click Training.
At the top of the screen you can see that there are several different ways we can train.
Select the Intentions tab to begin training
Then we can train our intents, our entities, bot synonyms, concepts and properties. More advanced topics will be discussed later.
Step two
We’ll start training intents by adding phrases—what the customer says or writes—to train our machine learning model. Here you can imagine trying to describe many different ways of asking the same question.
Step three
We will work on Track Order Intent.
Train track order
Once you’ve added all the expressions, you’ll train the machine learning model.
As a rule of thumb, with machine learning, the more quality data, the better. One caveat. Make sure you enter the various details that are most relevant to your topic.
Machine Learning Engine Learning
Miscellaneous data and synonyms
Another important point is Different Data, which means different ways of saying the same thing. In other words, synonyms or similar expressions. An example of how to train intentions is to follow an order.
The client can say: “I want to follow the order.” They could also say:
“I need help finding my order”
“What is the status of my shipment?”
“What is the status of my order?”
“I need help finding information about an order I recently placed.”
“Show me my order details”
“Track Order”
The above examples show many different ways that someone can say the same thing. These expressions in the dataset help train the ML model to recognize the many ways that someone can communicate a particular intent.
What is machine learning?
Machine learning is a subfield of artificial intelligence made up of a set of algorithms, features, and data sets that are constantly improving their experience. As the input grows, the machine on the AI ​​platform gets better at recognizing patterns and using it to make predictions.
Construction patterns
Within the underlying meaning engine, you can create patterns and rules. Fundamental is an approach to NLP (natural language processing) that focuses on understanding words themselves. Each user statement is broken down word by word into the intent (what the user is asking them to do) and the subjects (the necessary data needed to complete the task).
Track order intentions
Rules go hand in hand with properties. Again, the more variety you have in your dataset and representation of your properties, the better, so a virtual assistant will work best as someone to talk to.
Using data from testing and analysis
With both the Kore.ai test suite and Kore.ai analytics in your training suite, you can use them to train the model or make it smarter over time.
Build patterns within Intents
Add a few phrases to make sure the virtual assistant is trained to understand things related to order tracking, such as:
“I’d like to track my order” and recognize patterns. You can read more about this in the Kore.ai developer documentation.
In our example, “shipping status” represents a concept, so this pattern will use the concept “shipping status”. Click the Train button to train the ML model.
A workout can take a few seconds, maybe a few minutes. The more robust and more data you feed your ML models, the more time it takes. However, Kore.ai has made it very efficient and it never takes too long.
Learning Concepts
Track order intent training complete
Entities, synonyms and concepts
After the tutorial you should have some idea about order tracking and you can check it. Part of the concept definition is the understanding of entities, bot synonyms and concepts.
Subjects of training
Entities are things that can be extracted from a user’s speech, such as a phone number, payment method, or the person’s name on a credit or debit card. Entities are any information that may be useful to perform an action or collect the correct information for a given user. They help provide context to the virtual assistant so that it understands exactly what the user is trying to accomplish. Bot synonyms are similar.
Training Bot synonyms
With all the different words that can come up, what we want to do is find synonyms so we can talk about the same thing in different ways.
Training Bot synonyms
Concepts represent ideas such as “Hello” or “Goodbye” and can represent different ways of saying hello and goodbye. These are clusters of related words and synonymous terms that can be considered a group.
New concepts and synonyms
In our theoretical case, a new concept could be “Shipping Status”. Now that the concept instance is defined, the next step is to name the concept. Adding different synonyms for shipping status might include something like “Shipped,” “Processed,” “Arrival,” and perhaps “Delivery.”
Intent Patterns – Follow the order
With synonyms built, you can see all synonyms that can be used for this concept under “Delivery Status”.
Now you can test your newly created intelligence in your virtual assistant to modify the ML models and FM engine to optimize the experience and understanding.
This is a quick introduction to how to train dialog objects in terms of both machine learning and basic meaning. You can read more about Dialog Intents and Dialog Builder in the Kore.ai Developer Documentation and at https://kore.ai/platform/virtual-assistant/dialog-builder/ .
The Kore.ai website is full of information, videos, tutorials, white papers and more about the Kore.ai XO platform.
About Kore.ai
Kore.ai is a leader in conversational AI platforms and solutions, helping enterprises automate front- and back-office business interactions to deliver extraordinary experiences for their customers, agents and employees across voice and digital channels.