Projection Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Interpretable Decision Sets: A Joint Framework for Description and Prediction (2016, background): “Learning decision lists has been particularly well studied [5, 38, 21] and has been successfully used for building small and interpretable models [3, 24].”
The Use of Classifiers in Sequential Inference (2001, background): “ursive step: δt(s)=maxs′∈S δt−1(s′)P(s|s′,ot). In this model, the classiﬁers’ decisions are incorporated in the terms P(s|s′,o)and P1(s|o). To learn these classiﬁers we follow the projection approach [26] and separate P(s|s′,o)to many functions Ps′(s|o)according to the previous states s′. Hence as m”
Toward Attribute Efficient Learning of Decision Lists and Parities (2006, methodology): “A longstanding open problem in machine learning, posed first by Blum in 1990 [4,5,7,8] and again by Valiant in 1998 [33], is whether or not there exist attribute efficient algorithms for learning decision lists, which are essentially nested “if-then-else” statements (we give a precise definition in ”
Attribute-efficient learning of decision lists and linear threshold functions under unconcentrated distributions (2006, background): “This question was first asked by Blum in 1990 [3] and has since been re-posed numerous times [4, 5, 6, 29]; as we now briefly describe, a range of partial results have been obtained along different lines.”