| Title | Distributed Black-Box optimization via Error Correcting Codes |
| Publication Type | Conference Paper |
| Year of Publication | 2019 |
| Authors | Bartan, Burak, Pilanci, Mert |
| Conference Name | 2019 57th Annual Allerton Conference on Communication, Control, and Computing (Allerton) |
| Keywords | Black Box Security, black-box adversarial attacks, composability, convergence, Decoding, decoding step, deep convolutional neural networks, derivative-free optimization framework, distributed black-box optimization, encoding, error correcting codes, error correction codes, evolution strategies, Iterative methods, Linear programming, Metrics, neural nets, objective function, optimisation, Optimization methods, Perturbation methods, pubcrawl, resilience, Resiliency, stragglers, structured exploration methods, structured search directions, telecommunication security |
| Abstract | We introduce a novel distributed derivative-free optimization framework that is resilient to stragglers. The proposed method employs coded search directions at which the objective function is evaluated, and a decoding step to find the next iterate. Our framework can be seen as an extension of evolution strategies and structured exploration methods where structured search directions were utilized. As an application, we consider black-box adversarial attacks on deep convolutional neural networks. Our numerical experiments demonstrate a significant improvement in the computation times. |
| DOI | 10.1109/ALLERTON.2019.8919691 |
| Citation Key | bartan_distributed_2019 |
Distributed Black-Box optimization via Error Correcting Codes