PHD Discussions Logo

Ask, Learn and Accelerate in your PhD Research

Question Icon Post Your Answer

Question Icon

2 years ago in Artificial Intelligence By Mukesh

Are LLMs like ChatGPT moving away from the AI paradigm of AlphaGo?

Are LLMs like ChatGPT moving away from the AI paradigm of AlphaGo?Are large language models and systems like AlphaGo representing divergent branches of artificial intelligence?

All Answers (1 Answers In All)

By Shreya K Answered 8 months ago

They represent different paradigms but are converging. AlphaGo used Reinforcement Learning (RL) and search for mastery in a closed, rule-based world (Go). LLMs use supervised/self-supervised learning on massive text corpora for open-ended language tasks. While their training objectives differ, the boundary is blurring. LLMs now incorporate RL (e.g., RLHF for alignment), and AlphaGo-style search is used to augment LLM reasoning (e.g., Tree of Thoughts). The trajectory is toward hybrid systems that combine the generative knowledge of LLMs with the strategic, goal-oriented planning and search of AlphaGo-style AI.

Your Answer