An automated deep learning pipeline based on advanced optimisations for leveraging spectral classification modelling

Abstract

In deep learning (DL) modelling for spectral data, a major challenge is related to the choice of DL network architecture and the selection of the best hyperparameters. Often, slight changes to the neural architecture or its hyperparameter can have a direct influence on the model's performance, making its robustness questionable. To deal with it, this study presents an automated deep learning modelling based on advanced optimisation techniques involving Hyperband and Bayesian optimisation, to automatically find optimal neural architecture and its hyperparameters to reach robust DL models. The optimisation requires a base neural architecture to be initialized, however, later it automatically adjusts the neural architecture and the hyperparameters to reach the optimal model. Furthermore, to support the interpretation of the DL models, a wavelength weighing schema based on gradient-weighted class activation mapping (Grad-CAM) was implemented. The potential of the approach was showed on a real case of wheat variety classification with near-infrared spectral data. The performance of the classification was compared with that previously reported on the same dataset with different DL and chemometric approaches. The results showed that with the proposed approach a classification accuracy of 94.9% was reached, which was better than the best reported accuracy on the same data set i.e., 93%. Furthermore, the better performance was obtained with a simpler neural architecture compared to what was used in earlier studies. The automated deep learning based on advanced optimisation can support DL modelling of spectral data.

Tree and mycorrhizal fungal diversity drive intraspecific and intraindividual trait variation in temperate forests: Evidence from a tree diversity experiment

“…As deep learning has recently emerged as a promising tool in trait‐based ecology (Perry et al., 2022; Vasseur et al., 2022), we used a convolutional neural network (CNN) approach for leaf trait prediction based on the spectral data. First, input spectra were augmented from 2501 to 12,906 features by using transformations based on a combination of standard normal variates and Savitzky–Golay derivatives (Figure S3; Passos & Mishra, 2021). Samples within the calibration set were split into a training and a test set which accounted for a proportion of 70% and 30%, respectively.…”

Automatic Neural Network Hyperparameter Optimization for Extrapolation: Lessons Learned from Visible and Near-Infrared Spectroscopy of Mango Fruit

“…In the field of automated machine learning (Auto-ML), the state-of-the-art for automatically searching through a hyperparameter search space is Bayesian optimization. Bayesian optimization (BO) for HPO [22,23,24] chooses which hyperparameter configuration to try next. The first trial configuration is sampled randomly.…”

Intelligent Discrete Deep Learning Based Classification Methodology in Chemometrics

“…Dilmi and Ladjal have applied a combined method of long short-term memory recurrent neural networks (LSTM RNNs) and independent component analysis (ICA) techniques for water quality classification. Passos and Mishra et al have introduced an automated deep learning pipeline based on advanced optimization of leveraging spectral classification modeling. Cancilla et al have used machine learning models to study water quality.…”