Tutorial #3: Symbolic AI - Symbols, Operations, Expressions, LLM-based functions!
10It is my privilege to introduce Marius and Leo, creators of Symbolic AI. I also showcase some of the cool features you can do with Symbolic AI!
Symbolic AI defines everything as a symbol, with some symbols to store information, and some symbols are like expressions used to process the symbols to get information. At its core, symbols are manipulated using Large Language Model-defined operations.
Each symbol is basically stored in text form. There are basic operations already defined that enable you to do translation (translate), classification (choice), compose text (compose), do question answer (query). If you want to customize your own, you can use expressions to do any text-based instruction to the symbol. Furthermore, if you want even more flexibility, you can simply use an LLM-based function, just like in Wolfram-Alpha!
Really cool stuff, next time round we will demonstrate their Chatbot and AutoGPT-like systems using Symbolic AI!
What's more, there are also various other engines (other than OpenAI) you can interface with to process symbols or to generate symbols. For instance, you can use WolframAlpha (mathematics and symbol inference) to process symbols. You can also use Whispr (voice to text), Pinecone (database) and many more to generate symbols!
~~~~~~~~~~~~~~~~
Concept Video: https://www.youtube.com/watch?v=RW_7JdXvbRA
Symbolic AI Github: https://github.com/Xpitfire/symbolicai
Notebook used for demonstration: https://github.com/tanchongmin/SymbolicAITutorial/blob/main/Basics.ipynb
Wolfram Alpha LLM Functions: https://writings.stephenwolfram.com/2023/05/the-new-world-of-llm-functions-integrating-llm-technology-into-the-wolfram-language/
~~~~~~~~~~~~~~~~
0:00 Introduction
0:40 Install Symbolic AI
1:44 Set your OpenAI LLM model
2:32 Symbol: Basic unit of information storage in text
5:10 Expressions: Perform text-based operations on symbols
6:51 Compose and Query functions for processing text
8:32 LLM-based functions!
12:20 Cost of using model (1 cent over the last few operations!)
12:39 Summary
13:16 Leo’s input
13:50 Code Walkthrough (brief)
18:04 Conclusion - Using GPT4 API has better performance!
19:45 Teaser - Advanced features of Symbolic AI - Sequential Processing, Chatbot
~~~~~~~~~~~~~~~~
AI and ML enthusiast. Likes to think about the essences behind breakthroughs of AI and explain it in a simple and relatable way. Also, I am an avid game creator.
Discord: https://discord.gg/bzp87AHJy5
LinkedIn: https://www.linkedin.com/in/chong-min-tan-94652288/
Online AI blog: https://delvingintotech.wordpress.com/
Twitter: https://twitter.com/johntanchongmin
Try out my games here: https://simmer.io/@chongmin