Improved Random Features for Dot Product Kernels

被引:0
|
作者
Wacker, Jonas [1 ]
Kanagawa, Motonobu [1 ]
Filippone, Maurizio [2 ]
机构
[1] EURECOM, Data Sci Dept, Biot, France
[2] KAUST, Stat Program, Thuwal, Saudi Arabia
关键词
Random features; randomized sketches; dot product kernels; polynomial kernels; large scale learning;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Dot product kernels, such as polynomial and exponential (softmax) kernels, are among the most widely used kernels in machine learning, as they enable modeling the interactions between input features, which is crucial in applications like computer vision, natural language processing, and recommender systems. We make several novel contributions for improving the efficiency of random feature approximations for dot product kernels, to make these kernels more useful in large scale learning. First, we present a generalization of existing random feature approximations for polynomial kernels, such as Rademacher and Gaussian sketches and TensorSRHT, using complex-valued random features. We show empirically that the use of complex features can significantly reduce the variances of these approximations. Second, we provide a theoretical analysis for understanding the factors affecting the efficiency of various random feature approximations, by deriving closed-form expressions for their variances. These variance formulas elucidate conditions under which certain approximations (e.g., TensorSRHT) achieve lower variances than others (e.g., Rademacher sketches), and conditions under which the use of complex features leads to lower variances than real features. Third, by using these variance formulas, which can be evaluated in practice, we develop a data-driven optimization approach to improve random feature approximations for general dot product kernels, which is also applicable to the Gaussian kernel. We describe the improvements brought by these contributions with extensive experiments on a variety of tasks and datasets.
引用
收藏
页码:1 / 75
页数:75
相关论文
共 50 条
  • [1] Improved Bounds on the Dot Product under Random Projection and Random Sign Projection
    Kaban, Ata
    KDD'15: PROCEEDINGS OF THE 21ST ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2015, : 487 - 496
  • [2] Regularization with dot-product kernels
    Smola, AJ
    Ovári, ZL
    Williamson, RC
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 13, 2001, 13 : 308 - 314
  • [3] Learning Kernels with Random Features
    Sinha, Aman
    Duchi, John
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 29 (NIPS 2016), 2016, 29
  • [4] Conditionally positive definite dot product kernels
    Menegatto, V. A.
    Oliveira, C. P.
    Peron, A. P.
    JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS, 2006, 321 (01) : 223 - 241
  • [5] On conditionally positive definite dot product kernels
    Menegatto, V. A.
    Oliveira, C. P.
    Peron, Ana P.
    ACTA MATHEMATICA SINICA-ENGLISH SERIES, 2008, 24 (07) : 1127 - 1138
  • [6] On Conditionally Positive Definite Dot Product Kernels
    V.A.MENEGATTO
    C.P.OLIVEIRA
    Ana P.PERON
    Acta Mathematica Sinica(English Series), 2008, 24 (07) : 1127 - 1138
  • [7] A Spectral Analysis of Dot-product Kernels
    Scetbon, Meyer
    Harchaoui, Zaid
    24TH INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS (AISTATS), 2021, 130
  • [8] On conditionally positive definite dot product kernels
    V. A. Menegatto
    C. P. Oliveira
    Ana P. Peron
    Acta Mathematica Sinica, English Series, 2008, 24 : 1127 - 1138
  • [9] Spherical Random Features for Polynomial Kernels
    Pennington, Jeffrey
    Yu, Felix X.
    Kumar, Sanjiv
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 28 (NIPS 2015), 2015, 28
  • [10] Taming graph kernels with random features
    Choromanski, Krzysztof
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 202, 2023, 202