Play all

- Long-term chat with external sources

- Building a chatbot that queries Wikipedia

- Building a chatbot

- Searching a conversation for a specific word

- Searching for similar lines in a conversation

- Searching a conversation transcript for a specific word

- Sound dampening and the try-accept loop

- Cutting off a long conversation

- The impossibility of containing a superintelligence

- Using GPT3 to generate a list of Google search queries

- Creating a chatbot with gpt3

- Using GPT-3 to answer questions about Romulus Augustus

- The difficulty of asking the right questions

- Searching for follow-up questions to add to the conversation

- Generating a follow-up question based on a Wikipedia article

- Generating a search query to find an external article

- Cutting up a passage to answer a question

- Extracting information from a Wikipedia article

- The Last Emperor

- The question asked was

- Creating a chatbot with Auto Muse

- The need for a model that incorporates background information

Description:

Explore techniques for enhancing chatbot memory and information retrieval in this 54-minute video tutorial. Learn how to build a chatbot that queries Wikipedia, search conversation transcripts, generate Google search queries using GPT-3, and extract relevant information from external sources. Discover methods for creating follow-up questions, cutting long conversations, and addressing the challenges of incorporating background information into AI models. Gain insights into the complexities of developing more intelligent and context-aware chatbots, with practical examples and discussions on the limitations of current AI technologies.

Fixing "Goldfish Memory" with GPT-3 and External Sources of Information in a Chatbot

David Shapiro ~ AI

Add to list

#Computer Science #Artificial Intelligence #Natural Language Processing (NLP) #LLM (Large Language Model) #GPT-3 #ChatGPT

0:00 / 0:00