Talk:Transduction

This page was supposed to be more of a disambiguation page between the physical, molecular bio, physiology sides of transduction. If it's going to become the de facto physics transduction page, the format of the topmatter should be adjusted. I personally would prefer the information be sorted out to its appropriate disambig page such as Transduction (genetics) or Transduction (physics). --Chinasaur 17:26, 30 Aug 2004 (UTC)

Hello. I've put transduction (machine learning) on the list. In that context, transduction means to make inferences about new data directly from previous data, without constructing a model to explain the previous data and extrapolating the model to the new data. (The contrast drawn here is with induction, which is constructing a model in this context.) I am certain that this use of "transduction" is older than machine learning, however, I don't have any references. The only reference I've come across in a web search is for Vladimir Vapnik, but I know I've seen the term used in the same way in an unpublished paper by Wray Buntine, and doubtless there are other refs. Can someone help me out with a reference to the original use of "transduction" in the sense described? Regards & happy editing, Wile E. Heresiarch 15:33, 1 Sep 2004 (UTC)

To me, your transduction (machine learning) looks pretty much like transduction (psychology) (the one from Developmental_psychology). --Rubix

Cybernetics
The VSM use of the term transduction is notable and well-documented. Engineering purists have removed this usage from the Transducer page; although VSM can be regarded as systems engineering, it presumably isn't the kind of engineering that the engineering purists have in mind. So I have added an extra link from this disambig page to the VSM page. --RichardVeryard (talk) 11:43, 3 March 2008 (UTC)