Summary of Value-driven Mixed-precision Quantization For Patch-based Inference on Microcontrollers, by Wei Tao et al.
Value-Driven Mixed-Precision Quantization for Patch-Based Inference on Microcontrollers
by Wei Tao, Shenglin He, Kai Lu, Xiaoyang Qu, Guokuan Li, Jiguang Wan, Jianzong Wang, Jing Xiao
First submitted to arxiv on: 24 Jan 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed QuantMCU method utilizes value-driven mixed-precision quantization to address the challenges of deploying neural networks on microcontroller units (MCUs). This approach leverages a novel patch-based inference technique that reduces redundant computation overhead. The method first employs value-driven patch classification (VDPC) to maintain model accuracy, and then applies 8-bit quantization to feature maps with outlier values. For patches without outliers, it utilizes an iterative search process to determine the optimal bitwidth for each feature map. Experimental results show that QuantMCU reduces computation by 2.2x on average while maintaining comparable model accuracy. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary QuantMCU is a new way to make neural networks work better on tiny computers called microcontrollers. These computers have limited power and memory, making it hard to run complex AI models. The solution uses a special kind of compression that reduces the amount of calculations needed. This approach helps keep the model accurate while using less energy. It’s like finding the right balance between speed and accuracy. |
Keywords
* Artificial intelligence * Classification * Feature map * Inference * Precision * Quantization
