Volume 15, Issue 12, December 2024 Edition - IJSER Journal Publication


Publication for Volume 15, Issue 12, December 2024 Edition - IJSER Journal Publication

IJSER Research Group https://www.ijser.org/forum/index.php Register for IJSER Research Forum      IJSER Xplore IJSER Xplore Research Paper Database
Pages   [1]
 

MANURE OF ANIMALS FOR GENERATION OF ELECTRICITY AND PRODUCTION OF COMBUSTIVE (VEHICLE NATURAL GAS) AT CHITMA DISTRICT /TETE PROVINCE IN MOZAMBIQUE []


ABSTRACT: Once the energy is one of the very important in the development and the economy of Countries, all the nations are working in order to find the renewable sources for it. Knowing that the fossil fuel will varnish, associated to their impact in the environment linked to green house gases responsible of the global warming and acid rain, Mozambique is not behind this process. Researchers have been studying the away to generate energy friendly to the environment by using the cheapest sources and technologies, as well as , Municipal solid waste, water, solar, wind, ocean, thermal, now manure of animal, for instance, chickens, elephants, cows, pigs and human. The waste our bodies produce doesn't have to be a significant strain on our already limited resources. Harnessing it as a renewable energy source can also improve sanitation and reduce water pollution throughout the world and particular at Chitima district, Tete province in Mozambique. It's not a solution to the world energy problem, but the technology already exists and it is being shown to be economically feasible in a variety of situations. It is toilet to energy for domestic and cars use. It will help the rural area and the country to develop its economy, giving jobs, power and reduce the deforestation for nearly 9000 citizens in the first step, solving the objective of Mozambique as well as sustainable development goals such as; the Ensure of: availability and sustainable management of water and sanitation; access to affordable, reliable, sustainable and modern energy, for all; Ensure healthy lives and promote well-being for all at all ages, End poverty in all its forms everywhere; Make cities and human settlements inclusive, safe, resilient and sustainable; and Take urgent action to combat climate change and its impacts. Human excrement, as well as, of others animals to energy will generate power, biogas for cooking propose and be used to move cars reducing the impact of the increasing uses of fossil fuel, petrol and diesel, the principal responsible of green house gases, Keywords: Electricity, manure of animals, generation, source energy, petrol, sustainable,. Chitima, Mozambique

Detecting Weapons to inform Concerned Authority about Criminal Activity based on Computer Vision []


Computer vision has become a critical tool in security applications, enabling computers to identify and respond to potential threats. In this model, we introduce a computer vision-based approach to detect weapons (specifically, various guns and knives) and inform property authorities about potential criminal activity. When criminal threats involving firearms or knives are detected within the monitored premises, the model alerts the concerned authority through an automated system that sends SMS messages and makes calls. Additionally, audible alarms are activated at the location to prompt immediate attention from people nearby. This model is adaptable to various settings such as banks, offices, and homes, serving as a preventative measure against theft and robbery.

Mitigating Vulnerabilities in Federated Learning: Analyzing and Preventing Data and Model Poisoning Attacks []


Federated learning enables thousands of partici- pants to train a global machine learning model without sharing their private training data with each other. It distributes model training among a lot of agents guided by privacy concerns and performs training using only local data. The agents share only the model parameter updates, for iterative aggregation at the server to in turn train an overall global model. However, this lack of transparency can provide a new attack surface. Our work in this paper explores the variation of the global accuracy and loss under different hyper parameters settings like the maximum number of clients, proportion of malicious clients and number of rounds. We also experiment as to how the federated learning setting is impacted by threats like model poisoning attack. Model-poisoning attack is significantly more powerful than data poisoning attacks that target for the training data. We also attempt to prevent this attack by enabling some defensive techniques like norm thresholding in the global server side.

Pages   [1]