KERNEL LEARNING APPARATUS USING TRANSFORMED CONVEX OPTIMIZATION PROBLEM

Title:

KERNEL LEARNING APPARATUS USING TRANSFORMED CONVEX OPTIMIZATION PROBLEM

Document Type and Number:

WIPO Patent Application WO/2019/186650

Kind Code:

A1

Abstract:

In a kernel learning apparatus, a data preprocessing circuitry preprocesses and represents each data example as a collection of feature representations that need to be interpreted. An explicit feature mapping circuit designs a kernel function with an explicit feature map to embed the feature representations of data into a nonlinear feature space and to produce the explicit feature map for the designed kernel function to train a predictive model. A convex problem formulating circuitry formulates a non-convex problem for training the predictive model into a convex optimization problem based on the explicit feature map. An optimal solution solving circuitry solves the convex optimization problem to obtain a globally optimal solution for training an interpretable predictive model.

More Like This:

JPH0766283	[Title of Invention] Optimal Control Air Conditioner
JP5198656	Nitride semiconductor device
WO/2019/042571	ASYNCHRONOUS GRADIENT AVERAGING DISTRIBUTED STOCHASTIC GRADIENT DESCENT

Inventors:

ZHANG HAO (JP)
NAKADAI SHINJI (JP)
FUKUMIZU KENJI (JP)

Application Number:

PCT/JP2018/012159

Publication Date:

October 03, 2019

Filing Date:

March 26, 2018

Export Citation:

Click for automatic bibliography generation Help

Assignee:

NEC CORP (JP)

International Classes:

G06N99/00

Other References:

VEDALDI, ANDRE A ET AL.: "Efficient Additive Kernels via Explicit Feature Maps", IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, vol. 34, no. 3, 23 January 2012 (2012-01-23), pages 480 - 492, XP011398496, Retrieved from the Internet [retrieved on 20180614], DOI: 10.1109/TPAMI.2011.153
MORI, KOHEI ET AL.: "VC-dimension reduction algorithms for hyperkernel SVM-type machines", IEICE TECHNICAL REPORT, vol. 110, 21 October 2010 (2010-10-21), pages 95 - 98
OKANOHARA, DAISUKE: "Kernel method is not slow, Applicable for large scale with Random Fourier Features", NIKKEI ROBOTICS, 10 June 2016 (2016-06-10), pages 36 - 38

Attorney, Agent or Firm:

IKEDA, Noriyasu et al. (JP)

Download PDF:

View/Download PDF PDF Help

Previous Patent: RECIPE CREATION METHOD, SEMICONDUCTOR DEVICE MANUFACTURING METHOD, PROCESSING DEVICE, AND PROGRAM

Next Patent: LASER GAS MANAGEMENT SYSTEM, ELECTRON DEVICE MANUFACTURING METHOD, AND EXCIMER LASER SYSTEM CONTROL ...