2018/01/07 by Michalis Rizakis, Stylianos I. Venieris, Rizakis, Michalis +5 · 1 voice
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Hardware Architecture (cs.AR) #Machine Learning (cs.LG) #cs.AR #cs.CV #cs.LG
paper · pdf · doi:10.48550/arxiv.1801.02190
Accepted at the 14th International Symposium in Applied Reconfigurable Computing (ARC) 2018
arxiv created 2018/01/07 · arxiv published 2018/01/07 · arxiv updated 2018/01/10
Recurrent Neural Networks and in particular Long Short-Term Memory (LSTM) networks have demonstrated state-of-the-art accuracy in several emerging Artificial Intelligence tasks. However, the models are becoming increasingly demanding in terms of computational and memory load. Emerging latency-sensitive applications including mobile robots and autonomous vehicles often operate under stringent computation time constraints. In this paper, we address the challenge of deploying computationally demanding LSTMs at a constrained time budget by introducing an approximate computing scheme that combines iterative low-rank compression and pruning, along with a novel FPGA-based LSTM architecture. Combined in an end-to-end framework, the approximation method's parameters are optimised and the architecture is configured to address the problem of high-performance LSTM execution in time-constrained applications. Quantitative evaluation on a real-life image captioning application indicates that the proposed methods required up to 6.5x less time to achieve the same application-level accuracy compared to a baseline method, while achieving an average of 25x higher accuracy under the same computation time constraints.