Transfer Learning for Human & AI

Recent works onadversarial reprogramming have shown that it is possible to repurpose neuralnetworks for alternate tasks without modifying the network architecture or parameters . In this work, we broaden the scope of adversarialreprogramming beyond the data modality of the original task . We analyze thefeasibility of adversarially repurposing image classification neural networks for Natural Language Processing (NLP) and other sequence classification tasks . We demonstrate that by using highly efficient adversarialprograms, we can reprogram image classifiers to achieve competitive performance on a variety of text and sequence classification benchmarks without retraining the network . We use an efficient program that maps a sequence of discretetokens

Author(s) : Paarth Neekhara, Shehzeen Hussain, Jinglong Du, Shlomo Dubnov, Farinaz Koushanfar, Julian McAuley

Links : PDF - Abstract

Code :

https://github.com/oktantod/RoboND-DeepLearning-Project

Keywords : sequence - classification - reprogramming - image - tasks -

• Deep Learning A-Z
• Google Cloud Professional Data Engineer Specialization
• Complete Python Bootcamp
• How to Win a Data Science Competition on Coursera
• Khan Academy Multivariable Calculus
• AI Programming with Python
• Python for Everybody Specialization
• Machine Learning Nanodegree