posted 2009-10-20T01:54:51Z
eugene kleinberg's stochastic discrimination has always fascinated me, as being a "provably effective" technique for pattern recognition. but when i read donoho's paper on high dimensional phase transitions in modelling, compressive sensing, and experimental design, it makes me wonder what the connection may be, if any; for instance, does one see a similar kind of high-dimensional limit observed by donoho in stochastic discrimination?