Search
73 articles for “contrastive learning”
-
Self-Supervised Learning for Early Disease Detection in Low-Label Medical Image Datasets
Abstract: Getting reliable labeled data for medical image analysis is genuinely difficult. Radiologists, pathologists, and other specialist clinicians are expensive, slow, and not universally available — particularly in low-resource clinical environments across developing nations. Unlike natural image benchmarks such as ImageNet, which provide millions of meticulously labeled training examples, most publicly accessible medical imaging repositories offer only a few hundred to a few thousand annotated scans. Supervised deep learning models trained …
Published in Journal of Image Processing & Pattern Recognition Progress · Vol. 13, Issue 3, 2026 Read article
-
Continuous Learning in Language Models: A Survey of Streaming Data Processing Techniques
Abstract: The integration of continual learning with Large Language Models (LLMs) and Natural Language Processing (NLP) represents a transformative step toward creating adaptive, intelligent systems capable of functioning effectively in ever-changing environments. Traditional LLMs are typically trained on large, pre-collected datasets, which limits their ability to evolve as new information emerges. Continual learning, in contrast, enables models to acquire new knowledge incrementally without the need for complete retraining, thereby supporting long-term …
Published in Recent Trends in Programming languages · Vol. 12, Issue 3, 2025 · pp. 23–34 Read article
-
A Detailed Review on Traditional v/s Modern Education: Navigating the Path to Learning
Abstract: This abstract compares traditional and modern education methodologies, analyzing their respective approaches, strengths, and weaknesses. Traditional education, characterized by face-to-face instruction, textbooks, and standardized testing, emphasizes rote learning and adherence to established curriculum. In contrast, modern education integrates technology, interactive learning tools, and personalized approaches to foster critical thinking, creativity, and adaptability. While traditional education provides a structured framework and cultural continuity, modern education addresses diverse learning styles and prepares …
Published in International Journal of Education Sciences · Vol. 2, Issue 2, 2025 · pp. 67–74 Read article
-
Genomic Selection for Grain Yield in Wheat Using Machine Learning on DArT Molecular Markers: A Comparative Evaluation Across Multi-Environment Trials
Abstract: Genomic selection (GS) predicts complex quantitative traits directly from genome-wide molecular markers, bypassing the need for extensive phenotypic trials and accelerating plant breeding cycles. We conducted a comparative evaluation of seven regression approaches — ridge regression (the machine-learning equivalent of RR-BLUP), Lasso, Elastic Net, Partial Least Squares, linear Support Vector Regression, Random Forest, and Gradient Boosting — for predicting grain yield from 1,279 Diversity Array Technology (DArT) molecular markers genotyped …
Published in Research and Reviews : Journal of Computational Biology · Vol. 15, Issue 2, 2026 Read article
-
AI-Based Intelligent Traffic Signal Management System: A Review
Abstract: Traffic congestion is a growing problem in urban areas worldwide, leading to economic losses, increased pollution, and commuter frustration. Traditional traffic management systems rely on fixed timing cycles and lack adaptability to real-time traffic conditions. Intelligent traffic light control systems based on artificial intelligence (AI) have become a viable substitute for traditional techniques. These systems are able to evaluate large volumes of traffic data in real time, identify patterns, and …
Published in International Journal of Electronics Automation · Vol. 3, Issue 2, 2025 · pp. 1–5 Read article
-
Machine Learning Driven Mobile Price Prediction Using Feature Selection and Parameter Optimization
Abstract: Machine learning calculations are utilized in many fields like money, training, industry, medication, and online business. Machine learning calculations show execution contrasts relying upon the dataset and handling steps. Picking the right calculation, preprocessing and post-handling techniques have incredible significance in accomplishing great outcomes. The Random Forest classifier, K-nearest neighbor classifier, and support vector machine methods are evaluated to forecast mobile phone price categories. The “prediction” dataset which is taken …
Published in Current Trends in Information Technology · Vol. 14, Issue 3, 2024 · pp. 18–25 Read article
-
A Comparative Study on Various Machine Learning Techniques for the Prediction of Cardiac Ailment
Abstract: Over the last couple of decades, cardiovascular complexities have become the leading source of death in impoverished regions. With heart attack rates on the acceleration at a youthful age, it is necessary to put in place a process to recognize the symptoms of a heart attack early and thus limit it. It is impossible for a common man to often undergo expensive tests such as an ECG and thus there …
Published in Journal of Artificial Intelligence Research & Advances · Vol. 10, Issue 3, 2023 · pp. 25–30 Read article
-
Methods Based on Machine Learning for Large-scale Classification of Crop Leaf Diseases
Abstract: Worldwide productivity of crops is seriously threatened by crop leaf diseases, which can result in large crop losses and negative economic effects. Effective disease management and crop protection depend on the early and precise detection and classification of these illnesses. Machine learning approaches have gained popularity recently due to their ability to automate procedures related to illness diagnosis and classification. An overview of the several machine learning–based methods used for …
Published in International Journal of Computer Science Languages · Vol. 2, Issue 1, 2024 · pp. 11–23 Read article
-
Stock Price Prediction Using Data Science Techniques
Abstract: This study centers around persuasively presenting the potential to forecast the stability of future market stocks. Previous research has delved into predicting the trajectory of future market trends, leading to fluctuations in stock data, which opens avenues for refinement. The proposed model employs data science methodologies to predict the stock price index's value. This is achieved by contrasting supervised classification data science learning algorithms that predict either stock price increases …
Published in E-Commerce for Future & Trends · Vol. 10, Issue 2, 2023 · pp. 33–41 Read article
-
Fuzzy C-Means Clustering for Effective Segmentation and Classification of Brain Tumors in MRI Scans
Abstract: The paper discusses the importance of detecting and classifying brain tumors via MRI for effective treatment. It proposes a framework utilizing the Fuzzy C-means clustering algorithm for segmentation, demonstrating improved performance through real dataset validation. The model is trained on a large, annotated MRI dataset to identify and classify different tumor types, enabling machine learning-based classification into benign and malignant tumors. The MATLAB-based solution automates brain tumor feature extraction, aiding …
Published in Journal of Microwave Engineering and Technologies · Vol. 11, Issue 2, 2024 · pp. 23–28 Read article
-
3D Printing of Polymer-Based Functionally Graded Materials: Recent Developments and Challenges
Abstract: Additive manufacturing (AM), specifically 3D printing, has become a useful technique for fabricating functionally graded materials (FGMs) because it can facilitate the spatial distribution of materials. Polymer FGMs (P-FGMs) have gained a great deal of interest due to their lightweight, customizable, multifunctional properties. In comparison to conventional fabrication, 3D printing allows better control of composition and microstructure, which results in materials with controlled mechanical, thermal, and biological properties. This review …
Published in Journal of Polymer & Composites · Vol. 14, Issue 2, 2026 · pp. 338–347 Read article
-
Comparison and Analysis of Facial Emotion Detection Using Various Deep Learning Neural Networks
Abstract: Facial emotion recognition employs Convolutional Neural Networks (CNNs), Residual Networks (ResNet), Long Short-Term Memory (LSTM) networks, and Deep Neural Networks (DNNs) to automatically identify various emotions, including disgust, anger, fear, happiness, sadness, surprise, and neutrality. This study utilizes transfer learning along with data preprocessing techniques such as rotation, flipping, brightness adjustment, and enhancement methods. Traditional machine learning models achieve an accuracy range of 45 to 50%. In contrast, our proposed …
Published in Journal of Image Processing & Pattern Recognition Progress · Vol. 12, Issue 2, 2025 · pp. 37–42 Read article
-
Machine Learning Approaches for Phishing Detection: A Comparative Study
Abstract: These days, everyone has an internet addiction. All of us have used the internet for banking, booking, recharging, and buying. Phishing is a type of website threat that exists online. On the original website, phishing is an attempt to illegally obtain information such as login ID, password and credit card information. In this research, we proposed an efficient phishing detection system based on machine learning. Overall, the experimental findings demonstrated …
Published in International Journal of Satellite Remote Sensing · Vol. 2, Issue 1, 2024 · pp. 35–45 Read article
-
Advanced Deep Learning Techniques for Sickle Cell Anaemia Detection
Abstract: Sickle Cell Anemia (SCA) is a prevalent genetic blood disorder characterized by the presence of abnormal hemoglobin, resulting in the distinctive sickle shape of red blood cells. Timely and accurate identification of Sickle Cell Anemia (SCA) is essential for effective management and treatment. This study presents a new method that utilizes Convolutional Neural Networks (CNNs), a deep learning model particularly effective for image analysis. The process involves using microscopic images …
Published in Research and Reviews: A Journal of Medicine · Vol. 14, Issue 3, 2024 · pp. 9–15 Read article
-
Electric Vehicle Range Prediction
Abstract: The introduction of new energy vehicles has emerged as a new trend in the automotive industry inresponse to growing energy and environmental issues. The electric vehicle (EV) is the driving force behind newenergy vehicles. The one major problem electric vehicles have always been the distance(range) of the travel and mapto the nearby charging stations. For range prediction in the present study, four machine learningalgorithms—multiple linear regression, random forest regression, polynomial …
Published in Trends in Electrical Engineering · Vol. 13, Issue 3, 2023 · pp. 24–32 Read article
-
A Comparative Machine Learning Framework for Early Prediction of Liver Cancer Using Clinical Attributes
Abstract: One of the main causes of cancer-related death globally is liver cancer, and improving patient outcomes depends heavily on early detection. However, low contrast, noise, organ similarity, and tumor shape and size variability make it difficult to accurately identify and segment liver tumors from medical imaging. Automated liver cancer diagnosis, segmentation, and prognosis have been greatly improved by recent developments in artificial intelligence (AI), especially deep learning. This work presents …
Published in Research and Reviews: Journal of Oncology and Hematology · Vol. 15, Issue 2, 2026 · pp. 39–47 Read article
-
Matching Minutiae Fingerprint Q-Learning Approach for Detail Coordination: Identifiable Mark Point
Abstract: The use of fingerprints for high-precision recognition and identification of people is one of the most reliable biometric symbols because it is non-invasive. In this paper, we propose an innovative approach to detect details on low contrast resolution image quality of fingerprint images. Existing algorithms are not very susceptible to sound and image excellence due to the lack of level of intensity. We recommend a reliable route to find fingerprints …
Published in International Journal of Information Security Engineering · Vol. 1, Issue 1, 2023 · pp. 1–15 Read article
-
Lightweight Models for Per-PC Energy Consumption Forecasting: Comparative Study with ML and DL Approaches
Abstract: We have collected primary data from automated logging of parameters like CPU utilization, estimated power, active or idle state, user logging activity, and the type of day. Additionally, survey data showed user awareness, energy-saving behaviour, and PC usage patterns. The data is pre-processed and merged by applying processes such as data cleaning, normalization, and feature extraction, i.e., determining the peak active timings and downtime. Developed lightweight prediction models based on …
Published in Journal of Alternate Energy Sources & Technologies · Vol. 17, Issue 1, 2026 Read article
-
Review on CBIR Image Based on Colour, Texture and Shape Features of Biomedical Image Applications
Abstract: This study seeks to understand how different image enhancing methods affect the sensitivity of contrast-based textural measures and morphological traits derived from high-resolution satellite data (three-band SPOT-5). The built-up/non-built-up detection framework is the backbone of every biomedical application. Using supervised learning while working with a low-resolution reference layer reduces uncertainty and boosts the reference layer's quality in a roundabout way. The image's histogram is recalculated based on contrast in order …
Published in Journal of Image Processing & Pattern Recognition Progress · Vol. 11, Issue 1, 2024 · pp. 8–13 Read article
-
Optimization algorithms for multi objective problems and its application to sustainable sources for supplementing energy crisis
Abstract: Optimization, also called mathematical programming, is a collection of mathematical ideas and techniques used to solve quantitative issues in a variety of areas, including physics, biology, architecture, economics, and business. The subject evolved out of the recognition that quantitative issues in apparently disparate disciplines share significant mathematical features. As a result of this similarity, a wide variety of issues can be articulated and solved utilising the unified set of concepts …
Published in Journal of Microelectronics and Solid State Devices · Vol. 9, Issue 1, 2022 · pp. 19–25 Read article