Neural Networks as Explicit Word-Based Rules

2019/07/10 by Jindřich Libovický, Libovický, Jindřich · 1 voice
Computer Science · #68T50 #Computation and Language (cs.CL) #FOS: Computer and information sciences #I.2.7 #Machine Learning (cs.LG) #cs.CL #cs.LG

paper · pdf · doi:10.48550/arxiv.1907.04613

Abstract

Filters of convolutional networks used in computer vision are often visualized as image patches that maximize the response of the filter. We use the same approach to interpret weight matrices in simple architectures for natural language processing tasks. We interpret a convolutional network for sentiment classification as word-based rules. Using the rule, we recover the performance of the original model.

Discussions

• Neural Networks as Explicit Word-Based Rules [hn, 1 points, 0 comments]

Related

• Convolutional Neural Networks:
• Seeing in Words: Learning to Classify through Language Bottlenecks
• Learning Robust Representations of Text
• Molding CNNs for text: non-linear, non-consecutive convolutions
• Understanding Convolutional Neural Networks for Text Classification