Finding Patterns as the Path from Input to Output
Thomas D. Grant, Damon Wischik · 2020
The output of a machine learning system is nothing more than statements of the following form: “such and such a new case is likely to behave similarly to other similar cases that belong to the training dataset that was used to train this machine.” It has been described as “just curve fitting”, in the sense of drawing a curve through the datapoints in the training dataset, perhaps smoothing out some irregularities; the “learning” in machine learning is nothing more than tuning parameter values to make a well-fitting curve. This curve is then used to make predictions for new cases. The patterns found by machine learning are not laws of nature like Newton’s laws, and they are not stipulative rules like those laid down in statutes, they are simply fitted curves. Analogously, Holmes said that law emerges as people and their institutions find patterns in human experience. He wrote that a page of history is worth a volume of logic. When lawmakers ask for “the logic involved” in machine learning, or refer to “inferring rules from data”, they should really be asking for “a story about the training dataset”.