The Effect of Adversarial Machine Learning Attack on Agriculture Field and Food Security

Machine Learning and Deep Learning models had enabled several smart applications such as classification and recognition of objects. Smart machine and deep learning models can provide effective solutions to real life problems in various sectors such as agricultural and food security. Agriculture plays a significant role in food security where it supplies nations with the needed nutritious. Food security matters for nations as it ensures continuity and sufficiency of food and nutritious for humans. Machine and Deep learning models are exposed to a cybersecurity attack that is called the adversarial machine learning attack. This attack employes different methods and different algorithms to hamper the performance of the models. The attack initiates a noise attack over the images, videos, audios, text data that is being processed by the model. This results in the model misfunctioning and misclassifying the data which reduces the number of correctly classified samples and increase the number of wrongly classified samples by the models. The interesting aspect of the attack is that the human eye cannot detect the change of the data such as images. The data/images look identical before and after the adversarial attack. Only the models can detect the change in pixels and therefore the performance of the model degrades. This research paper provides a comprehensive literature review over the most common and widely used algorithms of the adversarial attack. The paper explores into the critical issue of food security, shedding light on its significance in global contexts. It highlights the potential consequences of adversarial attacks, emphasizing their unfavorable impact on food security and production systems. By drawing attention to these challenges, the paper calls for urgent measures to safeguard food security infrastructure against emerging cyber threats.