ePrints@IIScePrints@IISc Home | About | Browse | Latest Additions | Advanced Search | Contact | Help

A Novel Privacy-Centric Training Routine for Maintaining Accuracy in Traditional Machine Learning Systems

Haritas, HK and Haritas, CK and Kallimani, JS (2023) A Novel Privacy-Centric Training Routine for Maintaining Accuracy in Traditional Machine Learning Systems. In: 7th International Conference on Information and Communication Technology for Intelligent Systems, ICTIS 2023, 27-28 April 2023, Ahmedabad, pp. 257-263.

Full text not available from this repository. (Request a copy)
Official URL: https://link.springer.com/chapter/10.1007/978-981-...


The increasingly prevalent concerns on user privacy and user’s abstinence on data sharing assuredly spells a decline in novel model development for machine learning. On-Device AI is an exciting paradigm with local training and no requirements of externalizing personal data. Hybridizing these techniques of traditional data upload and server-side training to maximize accuracy and an On-Device approach, protecting privacy standards yields us a novel training routine, which can find itself extensively invoked with increasing privacy concerns. We enumerate multiple use cases in addition to examining feasibility and an abstract implementation of this concept. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Item Type: Conference Paper
Publication: Smart Innovation, Systems and Technologies
Publisher: Springer Science and Business Media Deutschland GmbH
Additional Information: The copyright of this conference proceeding belongs to the Springer Science and Business Media Deutschland GmbH.
Keywords: Data privacy; Machine learning; User profile, Attack modeling; Customer profiling; Invertibility; Malicious attack; Malicious attack/model; Naive model; Non-invertibility of training data; On-device AI; Personalized medicines; Smart watch; Training data; Transmission efficiency, Data acquisition
Department/Centre: UG Programme
Date Deposited: 27 Nov 2023 08:32
Last Modified: 27 Nov 2023 08:32
URI: https://eprints.iisc.ac.in/id/eprint/83254

Actions (login required)

View Item View Item