Transfer Learning

Deep learning techniques have become very popular among Artificial Intelligence (AI) techniques in many areas of life. Among many types of deep learning techniques, Convolutional Neural Networks (CNN) can be useful in image classification applications. In this work, a hybridized approach has been followed to classify lung nodule as benign or malignant. This will help in early detection of lung cancer and help in the life expectancy of lung cancer patients thereby reducing the mortality rate by this deadly disease scourging the world. The hybridization has been carried out between handcrafted features and deep features. The machine learning algorithms such as SVM and Logistic Regression have been used to classify the nodules based on the features. The dimensionality reduction technique, Principle Component Analysis (PCA) has been introduced to improve the performance of hybridized features with SVM. The experiments have been carried out with 14 different methods. It has been found that GLCM + VGG19 + PCA + SVM outperformed all other models with an accuracy of 94.93%, sensitivity of 90.9%, specificity of 97.36% and precision of 95.44%. The F1 score was found to be 0.93 and the AUC was 0.9843. The False Positive Rate was found to be 2.637% and False Negative Rate was 9.09%.

Cancer is an abnormal cell growth that occurs uncontrollably within the human body and has the potential to spread to other organs. One of the primary causes of mortality and morbidity for people is cancer, particularly lung cancer. Lung cancer is one of the non-communicable diseases (NCDs), causing 71% of all deaths globally, and is the second most common cancer diagnosed worldwide. The effectiveness of treatment and the survival rate of cancer patients can be significantly increased by early and exact cancer detection. An important factor in specifying the type of cancer is the histopathological diagnosis. In this study, we present a Simple Convolutional Neural Network (CNN) and EfficientNetB3 architecture that is both straightforward and efficient for accurately classifying lung cancer from medical images. EfficientnetB3 emerged as the best-performing classifier, acquiring a trustworthy level of precision, recall, and F1 score, with a remarkable accuracy of 100%, and superior performance demonstrates EfficientnetB3’s better capacity for an accurate lung cancer detection system. Nonetheless, the accuracy ratings of 85% obtained by Simple CNN also demonstrated useful categorization. CNN models had significantly lower accuracy scores than the EfficientnetB3 model, but these determinations indicate how acceptable the classifiers are for lung cancer detection. The novelty of our research is that less work is done on histopathological images. However, the accuracy of the previous work is not very high. In this research, our model outperformed the previous result. The results are advantageous for developing systems that effectively detect lung cancer and provide crucial information about the classifier’s efficiency.

With the advent of user-generated text information on the Internet, text sentiment analysis plays an essential role in online business transactions. The expression of feelings and opinions depends on the domains, which have different distributions. In addition, each of these domains or so-called product groups has its vocabulary and peculiarities that make analysis difficult. Therefore, different methods and approaches have been developed in this area. However, most of the analysis involved a single-domain and few studies on cross-domain mood classification using deep neural networks have been performed. The aim of this study was therefore to examine the accuracy and transferability of deep learning frameworks for the cross-domain sentiment analysis of customer ratings for different product groups as well as the cross-domain sentiment classification in five categories “very positive”, “positive”, “neutral”, “negative” and “very negative”. Labels were extracted and weighted using the Long Short-Term Memory (LSTM) Recurrent Neural Network. In this study, the RNN LSTM network was used to implement a deep transfer learning framework because of its significant results in sentiment analysis. In addition, two different methods of text representation, BOW and CBOW were used. Based on the results, using deep learning models and transferring weights from the source domain to the target domain can be effective in cross-domain sentiment analysis.

Satellite imagery provides valuable data to address the growing demand for agricultural production. However, analyzing such vast amounts of data requires advanced artificial intelligence methods, such as deep learning. The primary challenge lies in the scarcity of labeled training data, as its preparation is both costly and time-consuming. To address this issue, this study integrates remote sensing data, deep neural networks, and transfer learning techniques to estimate the cultivated area of strategic crops in Iran. Given the diverse climates and topographies across Iran’s provinces, in addition to Sentinel-1 and Sentinel-2 satellite data, MODIS sensor and SRTM elevation data were also utilized. To compensate for data limitations, transfer learning was employed to enhance model performance in data-deficient regions (Kermanshah and Markazi). This approach resulted in an approximate 10% improvement in Cohen’s Kappa coefficient. Furthermore, the study investigated the minimum data required for fine-tuning the models. The results demonstrated that even with a reduction of over 60% in the target province's training data, transfer learning still achieved model performance comparable to scenarios where it was not applied.

Relation extraction between drugs plays a crucial role in identifying drug–drug interactions and predicting side effects. The advancement of machine learning methods in relation extraction, along with the development of large medical text databases, has enabled the low-cost extraction of such relations compared to other approaches that typically require expert knowledge. However, to the best of our knowledge, there are limited datasets specifically designed for drug–drug relation extraction currently available. Therefore, employing transfer learning becomes necessary to apply machine learning methods in this domain. In this study, we propose DREAM, a method that first employs a trained relation extraction model to discover relations between entities and then applies this model to a corpus of medical texts to construct an ontology of drug relationships. The main contribution of this study is adapting the ACORD model to the medical domain using transfer learning, enabling the extraction of domain-specific drug–drug relations and the construction of an ontology. The extracted relations are subsequently validated using a large language model. Quantitative results indicate that the LLM agreed with 71 % of the relations extracted from a subset of PubMed abstracts. Furthermore, our qualitative analysis indicates that this approach can uncover ambiguities in the medical domain, highlighting the challenges inherent in relation extraction in this field.