mlsanityData to 14 Oct 2026

Evolutionary fine tuning of quantized convolution-based deep learning models

arXiv 2605.05228 ↗First posted 19 Apr 2026cs.LG cs.AI cs.NEcited by 0

Not on the map. Why?

Abstract (arXiv)

Deep learning models are the most efficient models in many machine learning tasks. The main disadvantage when using them in IoT, mobile devices, independent autonomous or real-time systems is their complexity and memory size. Therefore, much research has concentrated on compression techniques of deep learning architectures. One of the most popular technique is quantization. In most of the works, the quantization is done based on the nearest neighbour quantization technique. This work focuses on improving the quantization efficiency in pretrained and quantized models. This approach has the potential to improve the final accuracy of quantized models. The main postulate of the work is that final quantization states of the network based on nearest neighbour rounding does not guarantee optimal accuracy. In the presented work, the evolution strategy is used as an optimization approach. The evolution in each iteration changes the values of the small percentage of weights. It shifts theirs values to different quantization states. The work shows that proposed evolution with an appropriate set of operators and parameters can fast improve the accuracy of the quantized models. The results are presented for popular architectures such as VGG and Resnet for image classification and detection. Additionally, simulations were carried out for the autoencoder architecture.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 11 of its 22 references that are in the map. First posted on arXiv 19 Apr 2026.

Builds on

11 of 22 references.

References · 11

  1. 2015Learning both Weights and Connections for Efficient Neural Networkscited by 2,031
  2. 2020Up or Down? Adaptive Rounding for Post-Training Quantizationcited by 621
  3. 2017Evolving Deep Neural Networkscited by 95
  4. 2017WRPN: Wide Reduced-Precision Networkscited by 51
  5. 2016Faster CNNs with Direct Sparse Convolutions and Guided Pruningcited by 39
  6. 2016Hardware-oriented Approximation of Convolutional Neural Networkscited by 20
  7. 2018Training wide residual networks for deployment using a single bit for each weightcited by 10
  8. 2021Ensemble neuroevolution based approach for multivariate time series anomaly detectioncited by 8
  9. 2016Ristretto: Hardware-Oriented Approximation of Convolutional Neural Networkscited by 6
  10. 2021MONCAE: Multi-Objective Neuroevolution of Convolutional Autoencoderscited by 4
  11. 2021Speedup deep learning models on GPU by taking advantage of efficient unstructured pruning and bit-width reductioncited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Quantized Evolution Strategies: High-precision Fine-tuning of Quantized LLMs at Low-precision Cost
  2. 2023Bag of Tricks with Quantized Convolutional Neural Networks for image classification
  3. 2021Training Quantized Deep Neural Networks via Cooperative Coevolution
  4. 2022A Comprehensive Survey on Model Quantization for Deep Neural Networks in Image Classification
  5. 2024Differentiable Search for Finding Optimal Quantization Strategy
  6. 2024A Data-Free Analytical Quantization Scheme for Deep Learning Models
  7. 2024QGen: On the Ability to Generalize in Quantization Aware Training
  8. 2022QFT: Post-training quantization via fast joint finetuning of all degrees of freedom