RoMA: Robust Model Adaptation for Offline Model-based Optimization
This addresses the challenge of avoiding adversarial inputs in offline model-based optimization, which is crucial for applications like drug discovery and materials design, though it is an incremental improvement over existing methods.
The paper tackles the problem of maximizing a black-box objective function using a static dataset, where deep neural networks as proxy models often overestimate the true function at adversarially optimized inputs. The proposed RoMA framework uses robust pre-training and adaptation to leverage local smoothness, achieving state-of-the-art results by outperforming all methods in 4 out of 6 tasks and runner-up in the rest.
We consider the problem of searching an input maximizing a black-box objective function given a static dataset of input-output queries. A popular approach to solving this problem is maintaining a proxy model, e.g., a deep neural network (DNN), that approximates the true objective function. Here, the main challenge is how to avoid adversarially optimized inputs during the search, i.e., the inputs where the DNN highly overestimates the true objective function. To handle the issue, we propose a new framework, coined robust model adaptation (RoMA), based on gradient-based optimization of inputs over the DNN. Specifically, it consists of two steps: (a) a pre-training strategy to robustly train the proxy model and (b) a novel adaptation procedure of the proxy model to have robust estimates for a specific set of candidate solutions. At a high level, our scheme utilizes the local smoothness prior to overcome the brittleness of the DNN. Experiments under various tasks show the effectiveness of RoMA compared with previous methods, obtaining state-of-the-art results, e.g., RoMA outperforms all at 4 out of 6 tasks and achieves runner-up results at the remaining tasks.