The theory of inventive problem solving (TRIZ) reverse is a process in which known solutions are used as a basis for seeking new problems or applications and involves steps such as back-tracing product strengths to inventive principles, selecting catchwords, conducting database searches, analyzing patent lists, and identifying opportunities for patent exploitation. This paper explores the application of TRIZ reverse methodology to the identification of new markets for patented products, with a focus on low-density, high-strength thin cement sheets. Although research in the literature has used these methods, the details are often kept confidential. This study bridges that gap by offering a detailed examination of the TRIZ reverse process and its application to cement sheets, with specific demonstrations of patent search commands and a discussion of potential exploitation avenues. The insights provided here can facilitate a broader understanding and implementation of TRIZ reverse, thus empowering researchers to identify untapped market opportunities for existing technologies. The search results revealed various potential applications of thin cement sheets beyond solar panels. These include construction materials such as outer skins for wall panels, siding, partition walls, roofing, and ceilings; food industry applications such as food containers and boxes; wet-area lining boards; and smoke containment curtains for public buildings, including hotels, restaurants, prisons, hospitals, airports, and aircraft.

There has been a universal recognition that university-industry research collaboration (UIRC) is vital to strengthen the national innovation system (NIS) and economic growth. Despite a series of research studies on the significance of UIRC, present baseline models to enhance the capabilities of NIS through URIC are still scarce, specifically in developing countries. This research has highlighted that absorptive capacity has a vital influence on the NIS as well as on the research and innovative activities of an individual. Moreover, this research highlights how education and training (E&T) can enrich the absorptive capacity of NIS and UIRC using the Theory of Inventive Problem Solving (TRIZ) approach. The methodology involves applying TRIZ tools, such as function modeling, contradiction analysis, and inventive principles to identify effective strategies for improving the absorptive capacities of universities and industries and consequently of NIS. Thus, proposed solutions include enhancing the education, training systems and programs, promoting collaboration between universities and industries, and decreasing aids, such as foreign-educated and skilled workforce, which can strengthen the absorptive capacity of NIS. Analysis of this research suggests that a strong E&T system and upgrading the standard of education are crucial factors for improving the absorptive capacity of NIS. Recommendations include developing policies that foster a culture of knowledge, promoting interdisciplinary research, and incentivizing innovation. Future research directions include exploring comparative analysis with other developed and developing countries’ strategies in customizing the education and research system to enhance the outcomes of research and innovations.

Lung cancer is a leading cause of cancer-related mortality worldwide, and accurate detection of epidermal growth factor receptor mutations is essential for personalized treatment. However, non-invasive identification of these mutations remains challenging due to the complexity of clinical and morphological patterns. This study develops an adaptive boosting (AdaBoost)-based machine learning model for detecting lung cancer mutations using clinical and morphological data. The dataset consists of clinical and morphological attributes from 80 patients, which processed through comprehensive preprocessing steps, including imputation, outlier removal, and feature selection. One-hot encoding increased the feature count beyond the original 28, and analysis of variance was employed to retain the most relevant 33 features. AdaBoost was trained with optimized hyperparameters, including learning rate and the number of estimators, which were tuned using grid search to ensure robustness. The model’s performance was evaluated using an 80/20 train-test split and k-fold cross-validation to assess generalization capability. Experimental results demonstrated that AdaBoost outperformed other models, achieving an accuracy of 83% and an area under the curve of 0.90 after feature selection. The model maintained superior cross-validation scores compared to Naive Bayes, decision tree, K-nearest neighbors, and support vector machine, reinforcing its reliability in mutation detection. The study highlights the significance of preprocessing steps in improving classification performance and suggests that AdaBoost can serve as an effective, non-invasive tool for assisting clinical decision-making in lung cancer mutation detection.

Developing and evaluating a deep learning-based method to enhance satellite image resolution has emerged as a promising approach to address challenges posed by motion, imaging blur, and noise without modifying existing optical systems. This study utilized an enhanced super-resolution generative adversarial network (SRGAN) with ResNet-50 as the generator and a modified VGG-19 in the discriminator. The model was trained on remote sensing images from the Linear Imaging Self-Scanning imagery and compared with very deep super resolution, SRGAN, and enhanced SRGAN methods using the structural similarity index measure (SSIM) and peak signal-to-noise ratio (PSNR) as evaluation metrics. Utilizing an enhanced SRGAN with ResNet-50 and modified VGG-19 significantly improved satellite image resolution. The proposed method consistently outperformed conventional convolutional neural network- and generative adversarial network-based super-resolution techniques. Across three test datasets, the method achieved SSIM scores as high as 0.862 and PSNR scores of 33.256, 32.886, and 34.885, demonstrating its superior ability to preserve image properties and enhance resolution. The incorporation of perceptual loss alongside pixel loss contributed to improved visual quality, making the approach particularly effective in maintaining fine details and naturalistic high-frequency characteristics.

This research investigates the role of mobile multimedia platforms and artificial intelligence (AI) in driving innovation and ensuring the sustainability of entrepreneurial businesses, focusing particularly on technology acquisition, integration, and infrastructure. For data collection, the study employed a quantitative research design and surveyed 150 Indian technology firms that had adopted mobile multimedia applications. Structural equation modeling was used to analyze the data, supported by descriptive statistics, correlation, regression analysis, and mixed methods to understand the adoption and use of digital technologies for innovation activities. The results show that AI-driven applications, when combined with multimedia content and real-time analytics, significantly enhance entrepreneurial innovation by improving operational efficiency, increasing customer engagement, and facilitating expansion into new international markets. Companies utilizing mobile multimedia platforms gain a competitive advantage, translating into long-term business growth and sustainability. This research contributes to the literature on AI and entrepreneurship in the context of digital transformation, highlighting the need for startups to invest in AI-enabled mobile technologies. It equally serves policymakers by informing the regulation of an environment that promotes innovation and business sustainability through digital initiatives. This research addresses a significant gap in the literature by providing evidence on how AI acts as a driver of change and provides insight into the adoption of new technologies in the context of entrepreneurship, an area that remains largely underexplored.

Speech emotion recognition in Marathi presents considerable hurdles due to the language’s distinct grammatical and emotional characteristics. This paper presents a robust methodology for classifying emotions in Marathi speech utilizing advanced signal processing, feature extraction, and machine learning techniques. The method entails collecting diverse Marathi speech samples and using pre-processing steps such as pre-emphasis and voice activity detection to improve signal quality. Speech signals are segmented using the Hamming window to reduce discontinuities, and features such as Mel-frequency cepstral coefficients, pitch, intensity, and spectral properties are retrieved. For classification, an attentive deep belief network is paired with a support vector machine, which uses attention techniques and batch normalization to improve performance and reduce overfitting. The suggested approach surpasses existing models, with 98% accuracy, 98% F1-score, 99% specificity, 99% sensitivity, 98% precision, and 98% recall.

Fingerprint-based authentication is a critical biometric approach for ensuring security and accuracy. Traditional methods often face challenges such as noise and suboptimal feature extraction. To address the challenges, Fusion Net-3, an extensive model, is proposed to improve the speed, precision, and security level of fingerprint-based authentication systems. Fusion Net-3 operates through two separate stages: enrollment and authentication. During the enrollment phase, advanced pre-processing of fingerprint images was performed, incorporating an enhanced bilateral filter optimized with the seagull optimization algorithm. After pre-processing, features were obtained using a two-phase method: Zernike moments for shape-based features and local binary patterns for texture-based features. This helped ensure that fingerprint features were considered comprehensive for representation. For feature selection optimization, the falcon-inspired jackal optimization algorithm was proposed, a hybrid method combining the strengths of the golden jackal optimization and falcon optimization algorithm. Then, the selected features were combined using a combination of the geometric mean and the Fisher score to facilitate classification for a balanced and novel representation. During authentication, fingerprints were processed using similar techniques for consistency. Each fingerprint was labeled as genuine or fraudulent with the aid of the Fusion Net-3 model, which leverages the combined strengths of convolutional neural networks, ResNet-50, and U-Net. The model achieved an accuracy of 98.956% and a mean squared error of 0.0234 when implemented on a Python platform. Overall, the Fusion Net-3 model demonstrated superior performance compared to existing methods, effectively enhancing authentication accuracy and security.

In recent years, location-based social networks (LBSNs) have gained significant popularity, enabling users to interact with points of interest (POIs) using modern technologies. As more people rely on LBSNs for finding interesting venues, contextually aware and relevant recommendation systems have become very beneficial with practical applications. In this research, we propose an enhanced hybrid recommendation system, designed for LBSNs to improve the accuracy of suggestions by integrating collaborative filtering methods with singular value decomposition to handle sparse data, along with context-aware modeling to tailor recommendations based on user interests, and group recommendation to accommodate multi-user scenarios. In addition, we incorporate contextual aspects, such as spatial proximity and temporal behavior, into the model to ensure recommendations align closely with the user’s present surroundings and preferences. The proposed method extends further to group recommendations by considering individual inclinations into cohesive suggestions for groups interested in visiting POIs together. The proposed method is assessed using precision, recall, and F1 score, ensuring a thorough evaluation of its performance. To further highlight context-aware recommendations, we use clustering based on user preference, temporal behavior, and category-wise interaction to identify patterns across various venue types. The proposed method shows improved recommendations, specifically based on data from LBSNs, and develops an efficient solution for balanced user preferences with contextual influences.