Selected article for: "feature learning and machine learning"
Author: Saba-Sadiya, Sari; Alhanai, Tuka; Ghassemi, Mohammad M
Title: Feature Imitating Networks
  • Cord-id: 99lhodj3
  • Document date: 2021_10_10
    • ID: 99lhodj3
    Snippet: In this paper, we introduce a novel approach to neural learning: the Feature-Imitating-Network (FIN). A FIN is a neural network with weights that are initialized to reliably approximate one or more closed-form statistical features, such as Shannon's entropy. In this paper, we demonstrate that FINs (and FIN ensembles) provide best-in-class performance for a variety of downstream signal processing and inference tasks, while using less data and requiring less fine-tuning compared to other networks
    Document: In this paper, we introduce a novel approach to neural learning: the Feature-Imitating-Network (FIN). A FIN is a neural network with weights that are initialized to reliably approximate one or more closed-form statistical features, such as Shannon's entropy. In this paper, we demonstrate that FINs (and FIN ensembles) provide best-in-class performance for a variety of downstream signal processing and inference tasks, while using less data and requiring less fine-tuning compared to other networks of similar (or even greater) representational power. We conclude that FINs can help bridge the gap between domain experts and machine learning practitioners by enabling researchers to harness insights from feature-engineering to enhance the performance of contemporary representation learning approaches.

    Search related documents:
    Co phrase search for related documents
    • accuracy speed and machine learning: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19
    • accuracy speed and machine vision: 1
    • accuracy variation and low accuracy: 1, 2, 3
    • accuracy variation and low performance: 1
    • accuracy variation and machine learning: 1
    • low accuracy and machine learning: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
    • low accuracy and machine vision: 1
    • low performance and machine learning: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18
    • low performance despite and machine learning: 1