Suggestions for the new version of SVA

[Mcpherson]

Thanks for your reply to my other post, I can't wait for the new version of SVA, and if you need any beta testers I would love to help out!

I just had a couple of suggestions for the AI in the new version:

1 One of the most basic things (and often overlooked) in most AI programs is that they need to distinguish YOU from I. For example:

You like classical music, I like rock music.

Some AI's will later respond with something like: Do you feel good that you like classical music? Completely confusing what it was told. The ability to understand the difference between YOU and I is very basic, but important in any AI.

2 I'm not sure how easy it would be to do, but the ability to keep a conversational string without constantly having to repeat the subject of the conversation in each sentence. For example:

Me: What do you think about cars?
AI: I don't know about cars, tell me about them.
Me: Cars are ....

Every following response would have to include the word "cars" for the AI to understand what you're talking about. That doesn't follow normal conversation, which would use "they" "those" and "them" instead of cars when responding to the AI. The AI would have to understand that the conversation is still about a subject until the subject is changed. So for example, the above conversation would instead be:

Me: What do you think about cars?
AI: I don't know about cars, tell me about them.
Me: They ....., and they also .....
AI: Cars can be fun to drive fast.
Me: However motorcycles are different.
AI: What's different about motorcycles?
Me: They .....

I hope you get the idea, and look forward to your thoughts.

Thanks,

Mcp

[Fantom]

I feel that is because the AIML KB was never designed for Digital Assistants. I am sure this will be changed when SIML comes to the playground as it is Digital Assistant oriented.

AIML is for simple Chatbots and most of the output is designed to fake intelligent responses and that is why it usually replaces one or the other word. It can clearly be seen in the example you have given in your post.