# Approximate FPGA-based LSTMs under Computation Time Constraints

arXiv: https://arxiv.org/abs/1801.02190 · HTML: https://mlsanity.com/paper/1801.02190/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Jan 2018.

## Built on by (most cited first, 2 of 2)

- [Design Space Exploration of Approximate Computing Techniques with a Reinforcement Learning Approach](https://mlsanity.com/paper/2312.17525/) (2023)
- [MOHAQ: Multi-Objective Hardware-Aware Quantization of Recurrent Neural Networks](https://mlsanity.com/paper/2108.01192/) (2021)

## Abstract (arXiv)

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.