BEGIN:VCALENDAR
VERSION:2.0
PRODID:https://www.ost.ch/
METHOD:PUBLISH
BEGIN:VEVENT
UID:37d35929fc23cb30ace72afa91f720b9@fhs.ch
LOCATION:Oberseestrasse 10, 8640 Rapperswil
SUMMARY:Text Classification with Born's rule
DESCRIPTION:This talk presents a novel classification algorithm inspired by the notion of superposition of states in quantum physics and discusses its implementation in python. The implementation is compatible with the scikit-learn ecosystem. It supports both dense and sparse inputs and GPU-accelerated computing via cupy. Furthermore, the classifier can be embedded in a neural network architecture using pytorch. The network supports real and complex-valued inputs and outputs probabilities in the range [0, 1]. Classification performance, explainability, and computational efficiency are discussed. The method also applies to non-textual data and can be used in practice as a general-purpose classifier
DTSTART:20230307T161500
DTEND:20230307T164500
DTSTAMP:20230227T125433
END:VEVENT
END:VCALENDAR