Urban planning faces increasingly complex challenges, marked by rapid urbanization and the need for sustainable development. Recognizing these challenges, our research focuses on semantic segmentation in aerial imagery. In our research, we identify challenges such as inadequate delineation of environmental features and the difficulty of obtaining critical insights from aerial imagery. To tackle these issues, we propose a state-of-the-art solution employing a modified Inception-ResNet-V2 U-Net architecture. Our research encompasses a comprehensive methodology that includes meticulous dataset preparation, bespoke model architecture design, refined loss function optimization, and robust training protocols augmented by data enrichment techniques. Our results showcase an impressive model accuracy of 84.6%, underlining the method's superior performance in accurately delineating environmental features from aerial imagery. This land use and land cover classification empowers urban planners and developers with critical insights, facilitating informed decision-making and sustainable urban development. Our paper also focuses on contour detection performed on natural vegetation which is one of the crucial aspect for urban planning. This approach offers a tool that can be used in real time for urban planning and improved accuracy of environmental feature delineation.

In the domain of medicinal plant classification, where the diversity of plant species is vast and increasing, precise categorization is essential. Traditional approaches for the classification of species often struggle to accommodate unidentified or unclassified species. Our research introduces a framework to address the issue. Our intuition is to pose the medicinal plants classification as an open-world problem and predict the class for unknown samples in the hierarchy with confidence to the best-known label. To achieve this we propose a unique custom model that combines the concept of VGG-16 and cascaded classifier facilitating the identification of unknown medicinal plant species. This innovative approach significantly enhances the precision and adaptability of our classification system, addressing the challenges posed by unknown or unrecorded plant species. Given an unknown species, our custom model can predict the taxonomic categories of the species. We employed a dataset comprising ten diverse medicinal plant species, serving as the basis for training. The results demonstrated promising accuracy for unknown medicinal plants species. We used three unknown species named Wood Sorel, Noni and Curry Leaves to test our model. For the unknown species the model obtained average accuracies of 81.66% and 76.11 % for predicting phylum and class respectively.

In the realm of deep learning, the escalating complexity of neural networks has posed substantial computational challenges. This paper introduces a novel solution, Adaptive Low-Rank Approximations (ALRA), which marks a paradigm shift in model optimization. ALRA's distinctiveness lies in its adaptive nature; unlike conventional fixed-rank approximations, it dynamically adjusts the rank of factorized weight matrices during model training through Singular Value Decomposition (SVD). This adaptive capability enables deep learning models to autonomously tailor their complexity to the inherent data distribution, leading to a reduction in the number of model parameters and an improvement in predictive accuracy. We propose a method where the rank adjustment is integrated into the training process, allowing for continuous optimization of model complexity. Our experiments across diverse datasets and network architectures, including CNNs and ANNs, demonstrate ALRA's remarkable potential, consistently outperforming traditional fixed-rank approximations. ALRA's dynamic rank adaptation strategy not only reduces model size but also enhances convergence speed and robustness, making it particularly suited for real-time applications. By introducing ALRA, we offer a novel approach that combines adaptability and low-rank factorization to reshape the landscape of deep learning. We tested ALRA on various datasets like CIFAR-10, MNIST, SVHM, MNIST-Fashion, Iris, Wine and Diabetes. Results obtained after the integration of ALRA in CNNs and ANNs are promising and it outperforms the baseline CNNs and ANNs model.

The utilization of binary representation of the embeddings over real valued features represents a promising avenue, in terms of memory savings and faster operations for various machine learning models. In this research paper, we delve into the exploration of barcode representation for text embeddings derived from BERT, which is optimized using Co-ordinate Search algorithm. These binary embeddings present a compact representation of text, thereby mitigating memory and computational demands, which is especially advantageous in the context of resource-intensive large-scale text processing tasks. In our study, we introduce a novel optimal threshold technique, coupled with the Coordinate Search algorithm to transform continuous BERT embeddings into binary barcodes thereby enabling effective Natural Language Processing while sustaining computational efficiency. The optimal barcode representations have been applied in Natural Language Processing applications, showcasing its innovative potential in revolutionizing text representation. Through an extensive series of experiments on various NLP task encompassing diverse datasets, we comprehensively evaluate our approach, comparing it against a spectrum of thresholding techniques. The binary embeddings achieved by optimal thresholds outperform traditional binarization methods in terms of accuracy. The proposed method for generating a binary representations is versatile, being independent of the model, data and task, making it applicable across various machine learning applications.