Hardware Approximate Techniques for Deep Neural Network Accelerators: A Survey

Abstract

Deep Neural Networks (DNNs) are very popular because of their high performance in various cognitive tasks in Machine Learning (ML). Recent advancements in DNNs have brought beyond human accuracy in many tasks, but at the cost of high computational complexity. To enable efficient execution of DNN inference, more and more research works, therefore, exploit the inherent error resilience of DNNs and employ Approximate Computing (AC) principles to address the elevated energy demands of DNN accelerators. This article provides a comprehensive survey and analysis of hardware approximation techniques for DNN accelerators. First, we analyze the state of the art and by identifying approximation families, we cluster the respective works with respect to the approximation type. Next, we analyze the complexity of the performed evaluations (with respect to the dataset and DNN size) to assess the efficiency, the potential, and limitations of approximate DNN accelerators. Moreover, a broad discussion is provided, regarding error metrics that are more suitable for designing approximate units for DNN accelerators as well as accuracy recovery approaches that are tailored to DNN inference. Finally, we present how Approximate Computing for DNN accelerators can go beyond energy efficiency and address reliability and security issues, as well.

A Survey of FPGA-Based Vision Systems for Autonomous Cars

“…Another aggressive resource reduction technique impacting accuracy is the use of approximate computing units to perform the required computation [282].…”

Energy-Efficient DNN Inference on Approximate Accelerators Through Formal Property Exploration

“…Approximate computing has been heavily utilized on DNN inference [8]. Many works present mapping methodologies that balance out the computation accuracy-power consumption trade-off, and recent research has focused on the design and utilization of approximate multipliers on DNN inference.…”

“…A great amount of DNN operations can tolerate some degree of approximation [6], [7], and since the majority of DNN inference is spent on convolution and matrix multiplication operations, the design of approximate MAC units has attracted significant interest. Particularly, the majority of research has been focused on the design of approximate multipliers [8], as they are the most complex components of the MAC units and dominate energy consumption inside the unit. However, such multipliers are not application specific.…”

Exposing Reliability Degradation and Mitigation in Approximate DNNs Under Permanent Faults

“…Consequently, their deployment as DNN accelerators in smart nanoscale applications becomes very challenging [7] e.g., complex DNN analytics in resource-constrained edge devices, where safety and energy are critical considerations, can lead to an unexpected energy outage which can jeopardize human lives. This problem can be solved by leveraging approximate computing [8], [9] -an inexact computing method that exploits the inherent error resilience of onboard applications for energy efficiency in DNN accelerators. However, approximate hardware acceleration is deemed to be inherently less reliable [10].…”

“…Approximate computing-based deep neural network (AxDNN) accelerators are designed by incorporating inexact arithmetic units [11] [12], computation skipping [13], memory skipping [14], etc. Their fabrication at the nanoscale follows a sophisticated manufacturing process whose imperfections may result in manufacturing defects, such as process variations and permanent faults (stuck-at faults) [9]. As discussed in this paper, the permanent faults affect the compute units of AxDNN accelerators in every execution cycle and their presence as unmasked faults leads to serious failures in the whole system.…”