A friend of mine shared a link to an article about why children have a hard time learning to spell.  They argue that we are teaching spelling incorrectly.  We should be teaching how words are formed and the etymology not how to sound the word out to spell it.  They claim that about “12% of words in english are spelt the way they sound” [1].  No wonder I’m bad at spelling in english because that is usually how I spell things.  At least in Italian, words are spelt the way they sound!  Also, I was taught how to spell mainly as they described with spelling tests and memorization.  However, I did also learn definitions for each word.  However, I was really good at remembering the definitions, I guess I just didn’t think to associate the definitions with the spelling!  Although, I do sometimes do that.

So, maybe I should try learning to spell again using this method as I still struggle to spell and rely heavily on spell check.

[1] https://theconversation.com/why-some-kids-cant-spell-and-why-spelling-tests-wont-help-20497

Adaptive Data Analysis

Google Research recently wrote a piece The reusable holdout: Preserving validity in adaptive data analysis  (so I’m not go to write much).  It details the problem with statistics generated when the machine learning methods are adapted to the data through data analysis and repeated trials on the same hold out data.  This is a problem that is easily recognizable in machine learning competition leader boards like those on Kaggle (good article on that).  The solution they gave was to use their method detailed here and here (going to be published in Science) that allows us to reuse the hold out data set many times without loosing its validity.  So, that is awesome!  Hopefully the Kaggle competitions and data science research will be greatly improved by having more meaningful feedback.