Hardware-Aware Neural Dropout Search for Reliable Uncertainty Prediction on FPGA
This work addresses the need for efficient and reliable uncertainty prediction in AI for critical decision-making, offering an automated solution to optimize dropout strategies and hardware deployment, though it is incremental in improving existing dropout-based methods.
The paper tackles the suboptimal performance of uniform dropout in Bayesian Neural Networks for uncertainty estimation by proposing a neural dropout search framework that automatically optimizes dropout configurations and hardware implementations on FPGA, achieving up to 33X higher energy efficiency compared to manually-designed GPU-based methods and outperforming state-of-the-art FPGA designs.
The increasing deployment of artificial intelligence (AI) for critical decision-making amplifies the necessity for trustworthy AI, where uncertainty estimation plays a pivotal role in ensuring trustworthiness. Dropout-based Bayesian Neural Networks (BayesNNs) are prominent in this field, offering reliable uncertainty estimates. Despite their effectiveness, existing dropout-based BayesNNs typically employ a uniform dropout design across different layers, leading to suboptimal performance. Moreover, as diverse applications require tailored dropout strategies for optimal performance, manually optimizing dropout configurations for various applications is both error-prone and labor-intensive. To address these challenges, this paper proposes a novel neural dropout search framework that automatically optimizes both the dropout-based BayesNNs and their hardware implementations on FPGA. We leverage one-shot supernet training with an evolutionary algorithm for efficient dropout optimization. A layer-wise dropout search space is introduced to enable the automatic design of dropout-based BayesNNs with heterogeneous dropout configurations. Extensive experiments demonstrate that our proposed framework can effectively find design configurations on the Pareto frontier. Compared to manually-designed dropout-based BayesNNs on GPU, our search approach produces FPGA designs that can achieve up to 33X higher energy efficiency. Compared to state-of-the-art FPGA designs of BayesNN, the solutions from our approach can achieve higher algorithmic performance and energy efficiency.