# Understanding the Impact of Precision Quantization on the Accuracy and Energy of Neural Networks

arXiv: https://arxiv.org/abs/1612.03940 · HTML: https://mlsanity.com/paper/1612.03940/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Dec 2016.

## Built on by (most cited first, 18 of 18)

- [A Systematic Review of Green AI](https://mlsanity.com/paper/2301.11047/) (2023)
- [Green Edge AI: A Contemporary Survey](https://mlsanity.com/paper/2312.00333/) (2023)
- [MLPerf Power: Benchmarking the Energy Efficiency of Machine Learning Systems from Microwatts to Megawatts for Sustainable AI](https://mlsanity.com/paper/2410.12032/) (2024)
- [RecLight: A Recurrent Neural Network Accelerator with Integrated Silicon Photonics](https://mlsanity.com/paper/2209.00084/) (2022)
- [Resource-Efficient Deep Learning: A Survey on Model-, Arithmetic-, and Implementation-Level Techniques](https://mlsanity.com/paper/2112.15131/) (2021)
- [Low-Precision Hardware Architectures Meet Recommendation Model Inference at Scale](https://mlsanity.com/paper/2105.12676/) (2021)
- [How green is continual learning, really? Analyzing the energy consumption in continual training of vision foundation models](https://mlsanity.com/paper/2409.18664/) (2024)
- [PikeLPN: Mitigating Overlooked Inefficiencies of Low-Precision Neural Networks](https://mlsanity.com/paper/2404.00103/) (2024)
- [BitTrain: Sparse Bitmap Compression for Memory-Efficient Training on the Edge](https://mlsanity.com/paper/2110.15362/) (2021)
- [How Low Can We Go: Trading Memory for Error in Low-Precision Training](https://mlsanity.com/paper/2106.09686/) (2021)
- [A Mixture of Expert Approach for Low-Cost Customization of Deep Neural Networks](https://mlsanity.com/paper/1811.00056/) (2018)
- [FPGA-Accelerated SpeckleNN with SNL for Real-time X-ray Single-Particle Imaging](https://mlsanity.com/paper/2502.19734/) (2025)
- [Real-Time Cell Sorting with Scalable In Situ FPGA-Accelerated Deep Learning](https://mlsanity.com/paper/2503.12622/) (2025)
- [A Segmented Robot Grasping Perception Neural Network for Edge AI](https://mlsanity.com/paper/2507.13970/) (2025)
- [The Weight of a Bit: EMFI Sensitivity Analysis of Embedded Deep Learning Models](https://mlsanity.com/paper/2602.16309/) (2026)
- [Model Compression](https://mlsanity.com/paper/2105.10059/) (2021)
- [Optimizing the Energy Efficiency of Unreliable Memories for Quantized Kalman Filtering](https://mlsanity.com/paper/2109.01520/) (2021)
- [Fixflow: A Framework to Evaluate Fixed-point Arithmetic in Light-Weight CNN Inference](https://mlsanity.com/paper/2302.09564/) (2023)

## Abstract (arXiv)

Deep neural networks are gaining in popularity as they are used to generate state-of-the-art results for a variety of computer vision and machine learning applications. At the same time, these networks have grown in depth and complexity in order to solve harder problems. Given the limitations in power budgets dedicated to these networks, the importance of low-power, low-memory solutions has been stressed in recent years. While a large number of dedicated hardware using different precisions has recently been proposed, there exists no comprehensive study of different bit precisions and arithmetic in both inputs and network parameters. In this work, we address this issue and perform a study of different bit-precisions in neural networks (from floating-point to fixed-point, powers of two, and binary). In our evaluation, we consider and analyze the effect of precision scaling on both network accuracy and hardware metrics including memory footprint, power and energy consumption, and design area. We also investigate training-time methodologies to compensate for the reduction in accuracy due to limited bit precision and demonstrate that in most cases, precision scaling can deliver significant benefits in design metrics at the cost of very modest decreases in network accuracy. In addition, we propose that a small portion of the benefits achieved when using lower precisions can be forfeited to increase the network size and therefore the accuracy. We evaluate our experiments, using three well-recognized networks and datasets to show its generality. We investigate the trade-offs and highlight the benefits of using lower precisions in terms of energy and memory footprint.