IIT-Madras’s CeRAI and Ericsson Join Forces to Advance Responsible AI in 5G and 6G Era – News18

The IIT-Madras’s Centre for Responsible AI (CeRAI) and Ericsson have entered right into a partnership aimed toward advancing the sphere of Responsible AI. The collaboration was formally introduced throughout a symposium on ‘Responsible AI for Networks of the Future’ held on the premier institutes campus on Monday.

The spotlight of the occasion was the signing of an settlement by Ericsson, designating CeRAI as a ‘Platinum Consortium Member’ for a five-year time period. Under this Memorandum of Understanding (MoU), Ericsson Research will actively assist and have interaction in all analysis endeavours undertaken by CeRAI.

The Centre for Responsible AI at IIT-Madras is recognised as an interdisciplinary analysis hub with the imaginative and prescient to change into a premier establishment for each elementary and utilized analysis in Responsible AI. Its rapid aim is to deploy AI techniques throughout the Indian ecosystem whereas guaranteeing moral and accountable AI practices.

This partnership between CeRAI and Ericsson is predicted to make vital contributions to the event of moral and accountable AI practices within the evolving technological panorama within the nation.

What is Responsible AI

Responsible AI is an method to growing and deploying AI techniques in a protected, reliable, and moral means. It isn’t just about following a algorithm or tips, relatively about having a considerate and intentional method to AI growth and deployment. It is about contemplating the potential dangers and advantages of AI and ensuring that AI techniques are utilized in a means that’s honest, equitable, and helpful to all.

AI analysis has been gaining paramount significance in recent times, particularly within the context of the forthcoming 6G networks that can be pushed by AI algorithms. Dr Magnus Frodigh, Global Head of Ericsson Research, highlighted the importance of accountable AI within the growth of 6G networks. He emphasised that whereas AI-controlled sensors will join people and machines, accountable AI practices are important to make sure belief, equity, and privateness compliance.

Addressing the symposium, Prof Manu Santhanam, Dean of Industrial Consultancy and Sponsored Research at IIT-Madras, expressed optimism in regards to the collaboration, stating that analysis in AI will form the instruments for working companies sooner or later. He emphasised on IIT-Madras’s dedication to impactful translational work in collaboration with the trade.

Prof B Ravindran, Faculty Head at CeRAI, IIT-Madras, and Robert Bosch Centre for Data Science and AI (RBCDSAI), IIT-Madras, elaborated on the partnership, stating that the networks of the longer term will facilitate simpler entry to high-performing AI techniques.

Prof Ravindran pressured on the significance of embedding accountable AI rules from the outset in such techniques. He additionally highlighted that with the appearance of 5G and 6G networks, new analysis is required to make sure that AI fashions are explainable and may present efficiency ensures appropriate for numerous purposes.

Some of the tasks by the establishment showcased in the course of the occasion included:

• Large-Language Models (LLMs) in Healthcare: This challenge focuses on detecting biases exhibited by giant language fashions, growing scoring strategies to evaluate their real-world applicability, and lowering biases in LLMs. Custom scoring strategies are being designed primarily based on the Risk Management Framework (RMF) proposed by the National Institute of Standards and Technology (NIST).
• Participatory AI: This challenge addresses the black-box nature of AI at numerous levels of its lifecycle, from pre-development to post-deployment and audit. Drawing inspiration from domains like city planning and forest rights, the challenge explores governance mechanisms that allow stakeholders to supply constructive inputs, thereby enhancing AI customization, accuracy, and reliability whereas addressing potential adverse impacts.
• Generative AI Models Based on Attention Mechanisms: Generative AI fashions primarily based on consideration mechanisms have gained consideration for his or her distinctive efficiency in numerous duties. However, these fashions are sometimes complicated and difficult to interpret. This challenge focuses on enhancing the interpretability of attention-based fashions, understanding their limitations, and figuring out patterns they have an inclination to study from information.
• Multi-Agent Reinforcement Learning for Trade-off and Conflict Resolution in Intent-Based Networks: With the rising significance of intent-based administration in telecom networks, this challenge explores a Multi-Agent Reinforcement Learning (MARL) method to deal with complicated coordination and conflicts amongst community intents. It goals to leverage explainability and causality for joint actions of community brokers.