Current browse context:
stat
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Multi-output Polynomial Networks and Factorization Machines
(Submitted on 22 May 2017 (v1), last revised 4 Nov 2017 (this version, v2))
Abstract: Factorization machines and polynomial networks are supervised polynomial models based on an efficient low-rank decomposition. We extend these models to the multi-output setting, i.e., for learning vector-valued functions, with application to multi-class or multi-task problems. We cast this as the problem of learning a 3-way tensor whose slices share a common basis and propose a convex formulation of that problem. We then develop an efficient conditional gradient algorithm and prove its global convergence, despite the fact that it involves a non-convex basis selection step. On classification tasks, we show that our algorithm achieves excellent accuracy with much sparser models than existing methods. On recommendation system tasks, we show how to combine our algorithm with a reduction from ordinal regression to multi-output classification and show that the resulting algorithm outperforms simple baselines in terms of ranking accuracy.
Submission history
From: Vlad Niculae [view email][v1] Mon, 22 May 2017 08:20:31 GMT (311kb,D)
[v2] Sat, 4 Nov 2017 18:28:53 GMT (937kb,D)
Link back to: arXiv, form interface, contact.