A Simple Machine Learning Method for Commonsense Reasoning? A Short Commentary on Trinh & Le (2018)

Is there a'Simple' Machine Learning Method for Commonsense Reasoning? Menlo Park, CA This is a short Commentary on Trinh & Le (2018) ("A Simple Method for Commonsense Reasoning") that outlines three serious flaws in the cited paper and discusses why data-driven approaches cannot be considered as serious models for the commonsense reasoning needed in natural language understanding in general, and in reference resolution, in particular. A program is then asked the question "what was too small" as a followup to (1a), and the question "what was too big" as a followup to (1b). In a recent paper Trinh and Le (2018) - henceforth T&L - suggested that they have successfully formulated a „simple‟ machine learning method for performing commonsense reasoning, and in particular, the kind of reasoning that would be required in the process of language understanding. In simple terms, T&L suggest the following method for "learning" how to successfully resolve the reference "it" in sentences such as those in (1): generate two The Winograd Schema challenge was named after Terry Winograd, one of the pioneers of AI, who pointed out (Winograd, 1972) the need for using commonsense knowledge in resolving a reference such as „they‟ in sentences such as the following: The city councilmen refused the demonstrators a permit because they a.