In the evolving landscape of computer science, a dedicated master’s student, Mr Shridhar Singh, is making waves in the field of Generative Large-Language Models (LLMs) and revolutionising the way humans interact with artificial intelligence.
Singh is among the 29 students representing the School of Mathematics, Statistics, and Computer Sciences at the 2023 Postgraduate Research and Innovation Symposium (PRIS) hosted by UKZN’s College of Agriculture, Engineering, and Science at Coastlands Hotel, Musgrave, on 2 and 3 November. The symposium’s central theme is Water for Sustainability into the 21st Century.
Singh’s research explores the use of Zero Knowledge Proofs (ZKPs) to authenticate the computations of Large-Language Models. His research focuses on solving the critical issue of ensuring the authenticity of computations carried out by these models. LLMs have become instrumental in various applications, from natural language understanding to generating human-like text. However, the rise in the usage of these models has raised concerns about the credibility of information they produce and the privacy of data processed.
Enter Zero Knowledge Proofs (ZKPs), a cryptographic technique that allows LLMs to verify the authenticity of the data source before accepting it for computation, all while preserving the privacy and anonymity of user inputs. This approach not only enhances the reliability of LLMs, but also offers a strong defence against data tampering and infection. His work is not only about analysing sensitive and mission-critical data but also securing it in a way that is free from tampering and data breaches. It’s a journey into the cutting-edge of technology, one that aligns with his passion for Artificial Intelligence and its limitless potential.
Generative AI, a burgeoning field with applications across diverse domains, provides a conversational experience that closely resembles interaction with an expert in any given field. Singh’s exploration of Zero Knowledge Cryptography aligns with his interest in authenticating Large-Language Models.
The significance of his research reaches far beyond academia. Its potential applications are vast, and the scalability of the proof system enables it to handle the projected scope. He employs suggestions that it could be employed in mission-critical or sensitive data contexts. Organisations and individuals could create their own “Expert-GPT” systems and employ Zero Knowledge Proofs to validate the trustworthiness of the interactions with the LLMs, all while safeguarding sensitive and unnecessary information.
Shridhar Singh is a name to watch in the world of computer science and artificial intelligence. His research not only represents the cutting edge of technology, but also embodies a vision for a future where innovation and privacy can coexist.
