On Monday, September 9, 2019 at 12:32:28 PM UTC-5, Alan Grayson wrote: > > > > On Monday, September 9, 2019 at 4:06:33 AM UTC-6, John Clark wrote: >> >> Just 4 years ago 700 AI programs competed against each other and tried to >> pass a 8th-Grade multiple choice Science Test and win a $80,000 prize, but >> they all flunked, the best one only got 59.3% of the questions correct. But >> last Wednesday the Allen Institute unveiled a AI called "Aristo" that got >> 90.7% correct and then answered 83% of the 12th grade science test >> questions correctly. >> >> It seems to me that for a long time AI improvement was just >> creeping along but in the last few years things started to pick up speed. >> >> AI goes from F to A on the N.Y. Regents Science Exam >> <https://arxiv.org/pdf/1909.01958.pdf> >> >> John K ClarK >> > > Why do you think this has anything to do with intelligence and reasoning > ability? Maybe the programmers just expanded the memory and information of > those computers. AG >
https://www.geekwire.com/2019/allen-institutes-aristo-ai-program-finally-passes-8th-grade-science-test/ GeekWire: How is this approach different from IBM’s Watson <https://www.geekwire.com/2019/major-strategy-shift-ibm-watson-will-now-run-third-party-clouds-like-amazon-web-services/>? If Aristo were to compete against Watson, who would win? Clark: “The two systems were designed for very different kinds of questions. Watson was focused on encyclopedia-style ‘factoid’ questions where the answer was explicitly written down somewhere in text, typically many times. In contrast, Aristo answers science questions where the answer is not always written down somewhere, and may involve reasoning about a scenario, e.g.: - *“Otto pushed a toy car across a floor. The car traveled fast across the wood, but it slowed to a stop on the carpet. Which best explains what happened when the car reached the carpet? (A) Friction increased (B) Friction decreased…”* - *“City administrators can encourage energy conservation by (1) lowering parking fees (2) building larger parking lots (3) decreasing the cost of gasoline (4) lowering the cost of bus and subway fares.”* “Out of the box, Watson would likely struggle with science questions, and Aristo would struggle with the cryptic way that ‘Jeopardy’ questions were phrased. They’d each fail each other’s test. “Under the hood they are quite different too. In particular, Watson didn’t use deep learning (it was created before the deep learning technology) while Aristo makes heavy use of deep learning. Watson had many modules that tried different ways of looking for the answer. Aristo has a few (eight) modules that try a variety of methods of answering questions, including lookup, several reasoning methods and language modeling.” @philipthrift -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/d85e090f-9628-4148-b84a-e5e1dff7e169%40googlegroups.com.
