Imagine youre having friends over for lunch and plan to order a pepperoni pizza.
You recall Amy mentioning that Susie had stopped eating meat.
People take for granted the ability to deal with situations like these on a regular basis.
Its an extremely challenging effort.
40% off TNW Conference!
Naive, because people know such things despite not consciously working through physics equations.
Common sense also includes background knowledge of abstract notions, such as time, space and events.
This knowledge allows people to plan, estimate and organize without having to be too exact.
Even the latest models make absurd errors at times, suggesting thatsomething fundamental is missingin the AIs world model.
You are very thirsty.
So you
the highly touted AI text generator GPT-3supplied
drink it.
You are now dead.
A notable example is the four-yearMachine Common Senseprogram launched in 2019 by theU.S.
The project aims to build a computer system that can answer a wide range of commonsense questions.
Transformers to the rescue?
Commonsense question answering is an essential first step for building chatbots that can converse in a human-like way.
And how can we be sure an AI has common sense or not?
However, it is not always clear how cleanly these areas can be separated.
In ourrecent paper, experiments suggested that a clear answer to the first question can be problematic.
The annotators agreed on relatively concrete categories like time and space but disagreed on more abstract concepts.
Its already becoming painfully clear that even research in transformers is yielding diminishing returns.
Transformers are getting larger and morepower-hungry.
Arecent transformerdeveloped by Chinese search engine giant Baidu has several billion parameters.
It takes an enormous amount of data to effectively train.
Yet, it has so far proved unable to grasp the nuances of humancommon sense.
