Security and Privacy Challenges of Deep Learning

Abstract

Deep learning is the buzz word in recent times in the research field due to its various advantages in the fields of healthcare, medicine, automobiles, etc. A huge amount of data is required for deep learning to achieve better accuracy; thus, it is important to protect the data from security and privacy breaches. In this chapter, a comprehensive survey of security and privacy challenges in deep learning is presented. The security attacks such as poisoning attacks, evasion attacks, and black-box attacks are explored with its prevention and defence techniques. A comparative analysis is done on various techniques to prevent the data from such security attacks. Privacy is another major challenge in deep learning. In this chapter, the authors presented an in-depth survey on various privacy-preserving techniques for deep learning such as differential privacy, homomorphic encryption, secret sharing, and secure multi-party computation. A detailed comparison table to compare the various privacy-preserving techniques and approaches is also presented.

Privacy‐preserving data mining and machine learning in healthcare: Applications, challenges, and solutions

“…Furthermore, it presented various adversarial attacks and defense methods for many ML models. Onesimu et al (2021) presented a survey of security and privacy challenges in DL and also analyzed various techniques for preventing the DL networks from poisoning, evasion, and black box attacks and also discussed the PP computation techniques for DL such as differential privacy, homomorphic encryption (HE), secret sharing (SS), and secure multiparty computation. Boulemtafes et al (2022) reviewed the existing PP solutions for DL in a pervasive health monitoring environment.…”

“…Onesimu et al (2021) presented a survey of security and privacy challenges in DL and also analyzed various techniques for preventing the DL networks from poisoning, evasion, and black box attacks and also discussed the PP computation techniques for DL such as differential privacy, homomorphic encryption (HE), secret sharing (SS), and secure multiparty computation.…”

An Efficient Way of Hybridizing Edge Detectors Depending on Embedding Demand

“…Many of such applications use edge detection algorithm for localizing and visualizing target area in image and data. Before applying machine learning, if one wishes to implant privacy-preserving and securityrelated data [40]- [42] in the detected edge information that could be a promising technique to be used in the tele-medicine applications. Therefore, it is interesting to associate an edge detection method to divide the image contents into the edge and non-edge areas and to hide the data there [5,6,10,31,43].…”