How to teach a chatbot to take into account the context of a particular dialogue?

Question

PHP

With phrases, it’s simple: what is most popular is what the bot should answer (well, according to my idea, the chatbot collects statistics and answers a certain question with the most popular answer, for example, the question: “How are you?” and the most popular answer: “Good”). But how to teach a chatbot to take into account the context of a particular dialogue? I can't even imagine such a database structure.
Me: My name is Vanya.
Me: What's my name?
Bot: Your name is Vanya. (and not the most popular answer to the question "What is my name?", for example, "I don't know")

Answer 1

In short, the task is extremely difficult, no one can solve it normally.
When a computer learns to be in the semantic context of a dialogue, 80% of AI tasks will be solved, in particular, machine translation.
It will be enough for such a bot to understand the text through one language backend and retell it through the second one. Alas, there are none.

Answer 2

Here, for example, is a description of the templating language used in the 2010 Loebner Prize -winning Suzette chat bot .
The bottom line is that as phrases are parsed, the state of the chat bot changes - it creates a set of variables on the fly, entering data into them. After that, the state of these variables is further analyzed. The templates are also organized in a tree structure. Those. after choosing a topic for conversation, further input processing occurs only in a certain subset of templates.

Answer 3

contact IBM, they succeeded , maybe they will share links or even let them play )

Answer 4

Google on the topic "dialogue systems", there are video lectures, in short, it's time for you to study machine learning and deep neural networks in particular, and this is just the beginning of the list. This is now the front line, there are no ready-made solutions yet.