English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Neural networks are becoming a popular tool for solving many real-worldproblems such as object recognition and machine translation, thanks to itsexceptional performance as an end-to-end solution. However, neural networks arecomplex black-box models, which hinders humans from interpreting andconsequently trusting them in making critical decisions. Towards interpretingneural networks, several approaches have been proposed to extract simpledeterministic models from neural networks. The results are not encouraging(e.g., low accuracy and limited scalability), fundamentally due to the limitedexpressiveness of such simple models.  In this work, we propose an approach to extract probabilistic automata forinterpreting an important class of neural networks, i.e., recurrent neuralnetworks. Our work distinguishes itself from existing approaches in twoimportant ways. One is that probability is used to compensate for the loss ofexpressiveness. This is inspired by the observation that human reasoning isoften `probabilistic'. The other is that we adaptively identify the right levelof abstraction so that a simple model is extracted in a request-specific way.We conduct experiments on several real-world datasets using state-of-the-artarchitectures including GRU and LSTM. The result shows that our approachsignificantly improves existing approaches in terms of accuracy or scalability.Lastly, we demonstrate the usefulness of the extracted models through detectingadversarial texts.

Towards Interpreting Recurrent Neural Networks through Probabilistic Abstraction