提出输入编码新方法，让二值神经网络更准更轻。
End-to-end fully-binarized network design: from Generic Learned Thermometer to Block Pruning
- 用可学习阈值替代传统编码，提升输入表示
- 结合块剪枝与知识蒸馏，模型小于1兆字节
- 适合传感器端持续运行的低功耗场景
现有二值神经网络研究多关注权重和激活的二值化，忽视原始输入数据的表示。本文提出通用学习型温度计（GLT）编码技术，通过学习非线性量化阈值改进输入表示，实现多阶段数据二值化，可替代传统采用自然二进制编码的模拟到数字转换（ADC）。此外，我们设计紧凑拓扑，结合轻量级分组卷积与块剪枝及知识蒸馏训练，进一步降低模型规模与计算复杂度。实验表明，GLT通过内在全局色调映射显著提升准确率（在STL-10和VWW数据集上验证）。当与块剪枝结合时，可实现小于1兆字节的全二值化模型，精度损失小，适用于传感器端始终在线推理场景。
原文摘要 · Abstract (English)
Existing works on Binary Neural Network (BNN) mainly focus on model's weights and activations while discarding considerations on the input raw data. This article introduces Generic Learned Thermometer (GLT), an encoding technique to improve input data representation for BNN, relying on learning non linear quantization thresholds. This technique consists in multiple data binarizations which can advantageously replace a conventional Analog to Digital Conversion (ADC) that uses natural binary coding. Additionally, we jointly propose a compact topology with light-weight grouped convolutions being trained thanks to block pruning and Knowledge Distillation (KD), aiming at reducing furthermore the model size so as its computational complexity. We show that GLT brings versatility to the BNN by intrinsically performing global tone mapping, enabling significant accuracy gains in practice (demonstrated by simulations on the STL-10 and VWW datasets). Moreover, when combining GLT with our proposed block-pruning technique, we successfully achieve lightweight (under 1Mb), fully-binarized models with limited accuracy degradation while being suitable for in-sensor always-on inference use cases.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。