"Fred told the waiter he wanted some chips" .
By reading this one sentence, it is amazing that us, human being, know many related (highly possible) information, not written in the sentence at all.
We know Fred was in a restaurant. We know Fred was a customer dining here. We know that the waiter and Fred were only few feet apart. We know Fred wanted a potato chips, not some wood chips. We know that within several minutes, the chips should be ready for Fred. And many more.
How do we all know that? This is the power of commonsense.
This article explore how much GPT-3 is capable on commonsense in 8 fundamental reasoning dimensions.
See also the article's appendix for more intensive experiments on causality and counterfactual inferences.
GPT-3 commonsense score summary
|
From GPT-3 to AGI : 8 Challenging Prospects
In this up-coming article, we will explore the gaps of GPT-3, one of the best AI model we have in 2021, toward Artificial General Intelligence.
The gaps include: Commonsense and Mental Model, System 2 Integration, Life-long Learning, Vision & Reinforcement Integration, Long-range Reasoning / Long-term goal planning and
Precise Creative Thinking
Human Thinking Model according to Cognitive Science. How far to go from GPT-3 ?
|