Explainable Image Similarity: Integrating Siamese Networks and Grad-CAM.

With the proliferation of image-based applications in various domains, the need for accurate and interpretable image similarity measures has become increasingly critical. Existing image similarity models often lack transparency, making it challenging to understand the reasons why two images are considered similar. In this paper, we propose the concept of explainable image similarity, where the goal is the development of an approach, which is capable of providing similarity scores along with visual factual and counterfactual explanations. Along this line, we present a new framework, which integrates Siamese Networks and Grad-CAM for providing explainable image similarity and discuss the potential benefits and challenges of adopting this approach. In addition, we provide a comprehensive discussion about factual and counterfactual explanations provided by the proposed framework for assisting decision making. The proposed approach has the potential to enhance the interpretability, trustworthiness and user acceptance of image-based systems in real-world image similarity applications.

Special Issue on Machine Learning and AI for Sensors.

This article summarizes the works published under the "Machine Learning and AI for Sensors" (https://www [...].

A Convolutional Autoencoder Topology for Classification in High-Dimensional Noisy Image Datasets.

Deep convolutional neural networks have shown remarkable performance in the image classification domain. However, Deep Learning models are vulnerable to noise and redundant information encapsulated into the high-dimensional raw input images, leading to unstable and unreliable predictions. Autoencoders constitute an unsupervised dimensionality reduction technique, proven to filter out noise and redundant information and create robust and stable feature representations. In this work, in order to resolve the problem of DL models' vulnerability, we propose a convolutional autoencoder topological model for compressing and filtering out noise and redundant information from initial high dimensionality input images and then feeding this compressed output into convolutional neural networks. Our results reveal the efficiency of the proposed approach, leading to a significant performance improvement compared to Deep Learning models trained with the initial raw images.

Smoothing and stationarity enforcement framework for deep learning time-series forecasting.

Time-series analysis and forecasting problems are generally considered as some of the most challenging and complicated problems in data mining. In this work, we propose a new complete framework for enhancing deep learning time-series models, which is based on a data preprocessing methodology. The proposed framework focuses on conducting a sequence of transformations on the original low-quality time-series data for generating high-quality time-series data, "suitable" for efficiently training and fitting a deep learning model. These transformations are performed in two successive stages: The first stage is based on the smoothing technique for the development of a new de-noised version of the original series in which every value contains dynamic knowledge of the all previous values. The second stage of transformations is performed on the smoothed series and it is based on differencing the series in order to be stationary and be considerably easier fitted and analyzed by a deep learning model. A number of experiments were performed utilizing time-series datasets from the cryptocurrency market, energy sector and financial stock market application domains on both regression and classification problems. The comprehensive numerical experiments and statistical analysis provide empirical evidence that the proposed framework considerably improves the forecasting performance of a deep learning model.

Automatic classification of solitary pulmonary nodules in PET/CT imaging employing transfer learning techniques.

Early and automatic diagnosis of Solitary Pulmonary Nodules (SPN) in Computed Tomography (CT) chest scans can provide early treatment for patients with lung cancer, as well as doctor liberation from time-consuming procedures. The purpose of this study is the automatic and reliable characterization of SPNs in CT scans extracted from Positron Emission Tomography and Computer Tomography (PET/CT) system. To achieve the aforementioned task, Deep Learning with Convolutional Neural Networks (CNN) is applied. The strategy of training specific CNN architectures from scratch and the strategy of transfer learning, by utilizing state-of-the-art pre-trained CNNs, are compared and evaluated. To enhance the training sets, data augmentation is performed. The publicly available database of CT scans, named as Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI), is also utilized to further expand the training set and is added to the PET/CT dataset. The results highlight the effectiveness of transfer learning and data augmentation for the classification task of small datasets. The best accuracy obtained on the PET/CT dataset reached 94%, utilizing a modification proposal of a state-of-the-art CNN, called VGG16, and enhancing the training set with LIDC-IDRI dataset. Besides, the proposed modification outperforms in terms of sensitivity several similar researches, which exploit the benefits of transfer learning. Overview of the experiment setup. The two datasets containing nodule representations are combined to evaluate the effectiveness of transfer learning over the traditional approach of training Convolutional Neural Networks from scratch.

Explainable Machine Learning Framework for Image Classification Problems: Case Study on Glioma Cancer Prediction.

Image classification is a very popular machine learning domain in which deep convolutional neural networks have mainly emerged on such applications. These networks manage to achieve remarkable performance in terms of prediction accuracy but they are considered as black box models since they lack the ability to interpret their inner working mechanism and explain the main reasoning of their predictions. There is a variety of real world tasks, such as medical applications, in which interpretability and explainability play a significant role. Making decisions on critical issues such as cancer prediction utilizing black box models in order to achieve high prediction accuracy but without provision for any sort of explanation for its prediction, accuracy cannot be considered as sufficient and ethnically acceptable. Reasoning and explanation is essential in order to trust these models and support such critical predictions. Nevertheless, the definition and the validation of the quality of a prediction model's explanation can be considered in general extremely subjective and unclear. In this work, an accurate and interpretable machine learning framework is proposed, for image classification problems able to make high quality explanations. For this task, it is developed a feature extraction and explanation extraction framework, proposing also three basic general conditions which validate the quality of any model's prediction explanation for any application domain. The feature extraction framework will extract and create transparent and meaningful high level features for images, while the explanation extraction framework will be responsible for creating good explanations relying on these extracted features and the prediction model's inner function with respect to the proposed conditions. As a case study application, brain tumor magnetic resonance images were utilized for predicting glioma cancer. Our results demonstrate the efficiency of the proposed model since it managed to achieve sufficient prediction accuracy being also interpretable and explainable in simple human terms.