Can Adversarial Weight Perturbations Inject Neural Backdoors.
Siddhant GargAdarsh KumarVibhor GoelYingyu LiangPublished in: CIKM (2020)
Keyphrases
- network architecture
- neural network
- neural fuzzy
- satisfiability problem
- learning rules
- neural computation
- hidden structure
- multi agent
- bio inspired
- nonlinear predictive control
- neural information processing
- artificial neural
- weight function
- neural model
- computational intelligence
- sat instances
- hebbian learning
- weighting scheme
- search engine
- learning algorithm
- weight assignment
- real world
- data sets