[1911.01547] On the Measure of Intelligencecontact arXivarXiv Twitter

To make deliberate progress towards more intelligent and more human-like artificial systems, we need to be following an appropriate feedback signal: we need to be able to define and evaluate intelligence in a way that enables comparisons between two systems, as well as comparisons with humans. Over the past hundred years, there has been an abundance of attempts to define and measure intelligence, across both the fields of psychology and AI. We summarize and critically assess these definitions and evaluation approaches, while making apparent the two historical conceptions of intelligence that have implicitly guided them. We note that in practice, the contemporary AI community still gravitates towards benchmarking intelligence by comparing the skill exhibited by AIs and humans at specific tasks such as board games and video games. We argue that solely measuring skill at any given task falls short of measuring intelligence, because skill is heavily modulated by prior knowledge and experie

8 mentions: @hardmaru@fchollet@me_datapoint@jesusrenero@Eloy_GJ@Smerity@___walter_
Date: 2020/02/13 21:51

Referring Tweets

@fchollet Also, if you're curious why ARC looks the way it does, where it comes from, and what it means to work on solving it, you should check out this paper: t.co/djNAIUI3J4
@hardmaru Excited to see a @Kaggle competition based on the Abstraction and Reasoning Corpus by @fchollet in “On the Measure of Intelligence” (t.co/m4Q0Y5EKAy) t.co/ejb2afTQ4k t.co/cmAJBPOGh2
@jesusrenero @j_murillo_arias François Chollet reached a similar line of thought in its recent paper (t.co/PyLfaV8YSq) and Abstract Reasoning Challenge.
@me_datapoint Two very knowledgeable articles about current shortcomings in #AI/#ML and promising ways towards more general, robust, and flexible #AI: 1."On the measure of intelligence" @fchollet: t.co/0q741iqs6v 2."The Next Decade in AI" @GaryMarcus: t.co/x17ORzPnPn
@Eloy_GJ @canizalesdiaz @fchollet Abstraction and Reasoning Corpus (ARC), François talks about this in his recent publication ‘On The Measure Of Intelligence’ t.co/da6QkfAw3T
@Smerity @dileeplearning @fchollet Overall I think many of the concepts and discussions from @fchollet's "On the Measure of Intelligence" come through as a successful + interesting environment. Framing it as a competition + accessible dataset means finding what's easy vs hard faster =] t.co/BCotDxlBDw
@___walter_ This Kaggle competition based on the Abstraction and Reasoning Corpus by @fchollet check (“On the Measure of Intelligence” t.co/bhOrWhMgcn) is possibly the most interesting thing I saw in a long, long time. Incredibly challenging and inspiring. t.co/kCALZxGfAp