# Searching for Low-Bit Weights in Quantized Neural Networks

arXiv: https://arxiv.org/abs/2009.08695 · HTML: https://mlsanity.com/paper/2009.08695/

## Key facts

- Cited by 44 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Sep 2020.

## Built on by (most cited first, 30 of 44)

- [A Survey of Quantization Methods for Efficient Neural Network Inference](https://mlsanity.com/paper/2103.13630/) (2021)
- [Distilling Object Detectors via Decoupled Features](https://mlsanity.com/paper/2103.14475/) (2021)
- [Vision Transformer Pruning](https://mlsanity.com/paper/2104.08500/) (2021)
- [A Comprehensive Survey on Model Quantization for Deep Neural Networks in Image Classification](https://mlsanity.com/paper/2205.07877/) (2022)
- [Multi-Prize Lottery Ticket Hypothesis: Finding Accurate Binary Neural Networks by Pruning A Randomly Weighted Network](https://mlsanity.com/paper/2103.09377/) (2021)
- [ReCU: Reviving the Dead Weights in Binary Neural Networks](https://mlsanity.com/paper/2103.12369/) (2021)
- [A comprehensive review of Binary Neural Network](https://mlsanity.com/paper/2110.06804/) (2021)
- [Manifold Regularized Dynamic Network Pruning](https://mlsanity.com/paper/2103.05861/) (2021)
- [Learning Frequency Domain Approximation for Binary Neural Networks](https://mlsanity.com/paper/2103.00841/) (2021)
- [AdaBin: Improving Binary Neural Networks with Adaptive Binary Sets](https://mlsanity.com/paper/2208.08084/) (2022)
- [F8Net: Fixed-Point 8-bit Only Multiplication for Network Quantization](https://mlsanity.com/paper/2202.05239/) (2022)
- [Faster Than Lies: Real-time Deepfake Detection using Binary Neural Networks](https://mlsanity.com/paper/2406.04932/) (2024)
- [SiMaN: Sign-to-Magnitude Network Binarization](https://mlsanity.com/paper/2102.07981/) (2021)
- [GhostNets on Heterogeneous Devices via Cheap Operations](https://mlsanity.com/paper/2201.03297/) (2022)
- [Lipschitz Continuity Retained Binary Neural Network](https://mlsanity.com/paper/2207.06540/) (2022)
- [Automated Deep Learning: Neural Architecture Search Is Not the End](https://mlsanity.com/paper/2112.09245/) (2021)
- [APNN-TC: Accelerating Arbitrary Precision Neural Networks on Ampere GPU Tensor Cores](https://mlsanity.com/paper/2106.12169/) (2021)
- [Network Binarization via Contrastive Learning](https://mlsanity.com/paper/2207.02970/) (2022)
- [Recurrent Bilinear Optimization for Binary Neural Networks](https://mlsanity.com/paper/2209.01542/) (2022)
- [Low-bit Model Quantization for Deep Neural Networks: A Survey](https://mlsanity.com/paper/2505.05530/) (2025)
- [Equal Bits: Enforcing Equally Distributed Binary Network Weights](https://mlsanity.com/paper/2112.03406/) (2021)
- [BiPer: Binary Neural Networks using a Periodic Function](https://mlsanity.com/paper/2404.01278/) (2024)
- [GhostSR: Learning Ghost Features for Efficient Image Super-Resolution](https://mlsanity.com/paper/2101.08525/) (2021)
- [S3: Sign-Sparse-Shift Reparametrization for Effective Training of Low-bit Shift Networks](https://mlsanity.com/paper/2107.03453/) (2021)
- [Towards Accurate Binary Neural Networks via Modeling Contextual Dependencies](https://mlsanity.com/paper/2209.01404/) (2022)
- [XnODR and XnIDR: Two Accurate and Fast Fully Connected Layers For Convolutional Neural Networks](https://mlsanity.com/paper/2111.10854/) (2021)
- [Learning Versatile Convolution Filters for Efficient Visual Recognition](https://mlsanity.com/paper/2109.09310/) (2021)
- [Compacting Binary Neural Networks by Sparse Kernel Selection](https://mlsanity.com/paper/2303.14470/) (2023)
- [BOLD: Boolean Logic Deep Learning](https://mlsanity.com/paper/2405.16339/) (2024)
- [Efficient Mixed Precision Quantization in Graph Neural Networks](https://mlsanity.com/paper/2505.09361/) (2025)

## Abstract (arXiv)

Quantized neural networks with low-bit weights and activations are attractive for developing AI accelerators. However, the quantization functions used in most conventional quantization methods are non-differentiable, which increases the optimization difficulty of quantized networks. Compared with full-precision parameters (i.e., 32-bit floating numbers), low-bit values are selected from a much smaller set. For example, there are only 16 possibilities in 4-bit space. Thus, we present to regard the discrete weights in an arbitrary quantized neural network as searchable variables, and utilize a differential method to search them accurately. In particular, each weight is represented as a probability distribution over the discrete value set. The probabilities are optimized during training and the values with the highest probability are selected to establish the desired quantized network. Experimental results on benchmarks demonstrate that the proposed method is able to produce quantized neural networks with higher performance over the state-of-the-art methods on both image classification and super-resolution tasks.