IIT Roorkee Home

Events

Events Search and Views Navigation

Event Views Navigation

Today

Latest Past Events

Invited Talk by Dr. Vinay Kumar B. R. , Postdoctoral Researcher at Eindhoven University of Technology ; Talk Title: Community detection on geometric graphs

ECE Conference Room ECE Department, Roorkee

All of you are cordially invited to attend an invited talk by Dr. Vinay Kumar B. R., Postdoctoral Researcher at Eindhoven University of Technology on 13th January 2025 (Monday) at 11 am in the ECE Conference Room (ECE Department). Talk Title: Community detection on geometric graphs Abstract In many real-world networks, such as co-authorship and social networks, the graph structure is correlated with the locations of the nodes. The geometric dependence is typically evidenced by the absence of long-distance edges and the abundance of triangles. Detecting latent communities on such geometric graphs has been an important direction of research. We consider the community recovery problem on a random geometric graph where every node has two independent labels: a location label and a community label. A geometric kernel maps the locations of pairs of nodes to probabilities. Edges are drawn between pairs of nodes based on their communities and the value of the kernel corresponding to the respective node locations. Given the graph so generated along with the location labels, the latent communities of the nodes are to be inferred. In this talk, we will look into the fundamental limits for recovering the communities in such models. Additionally, we propose a linear time algorithm (in the number of edges) and show that it recovers the communities of nodes exactly up to the information-theoretic threshold. About the speaker: Dr. Vinay Kumar B.R. did his PhD in the Dept. of Electrical Communication Engineering at the Indian Institute of Science, Bangalore, under the guidance of Prof. Navin Kashyap. His thesis was titled “Probabilistic Forwarding of Coded Packets for Broadcasting over Networks”. He was a post-doctoral researcher at INRIA Sophia Antipolis - Méditerranée working with Konstantin Avrachenkov in the NEO team prior to joining the NETWORKS-COFUND program in 2024, where he currently works with Nelly Litvak and Remco van der Hofstad at the Eindhoven University of Technology, Netherland. Broadly, his research is in the areas of random graphs and network science. He is interested in problems that involve a graph structure and complex interactions between the network elements. His research goal is to propose and analyse robust mathematical models that capture different physical phenomena observed on practical networks.

Invited talk by Dr. Harshan Jagadeesh, Talk Title: On Spatial-Provenance Recovery in Wireless Networks with Relaxed-Privacy Constraints

ECE Conference Room ECE Department, Roorkee

Talk Abstract: A number of applications in next-generation networks impose low-latency requirements on learning the provenance of the packets, which in turn could be used for learning the topology of the network. While the state-of-the-art provenance embedding methods focus on the footprint of information flow, there exist interesting use-cases in vehicle-to-everything networks, wherein the road side units intend to learn the spatial-provenance of the packets to offer various location-based services and for detecting security threats on the network. Although vehicles use the global positioning system for navigation, they may refrain from sharing their exact GPS coordinates to the RSUs due to privacy considerations. Thus, to address the localization expectations of the RSUs and the privacy concerns of the vehicles, in this talk, we will present a relaxed-privacy model wherein vehicles share their partial location information through spatial-provenance in order to avail the location-based services. To implement this notion of relaxed privacy, we discuss a low-latency protocol for spatial-provenance recovery, wherein vehicles use correlated linear Bloom filters to embed their position information. Our proposed spatial-provenance recovery process takes into account the resolution of localization, the underlying ad hoc protocol, and the coverage range of the wireless technology used by the vehicles. Through a rigorous theoretical analysis, we present extensive analysis on the underlying trade-off between relaxed-privacy and the communication-overhead of the protocol. Finally, using a wireless testbed, we show that our proposed method requires a few bits in the packet header to provide security features such as localizing a low-power jammer executing a denial-of-service attack. About the speaker: Dr. Harshan Jagadeesh is an Associate Professor in the Department of Electrical Engineering, Indian Institute of Technology Delhi. He is also the co-coordinator of the center of excellence on cybersecurity and information assurance at IIT Delhi. Prior to joining IIT Delhi, he worked as a Researcher in the CyberSecurity group at Advanced Digital Sciences Center, Singapore. Before that he worked as a Research Fellow in the Division of Mathematical Sciences, Nanyang Technological University, Singapore and in the Department of Electrical and Computer Systems Engineering at Monash University, Australia. He obtained the Ph.D. degree from the Department of Electrical Communication Engineering, Indian Institute of Science, India. His research interests are in the broad areas of security and privacy applied to wireless and storage networks. All are cordially invited to attend the lecture. 

Invited talk on “Federated Learning over Wireless Networks

ECE Conference Room ECE Department, Roorkee

Abstract: Responsible AI refers to the development and deployment of artificial intelligence (AI) technologies in an ethical, transparent and fair manner, aligned with societal values. The key pillars of responsible AI include fairness, explainability, transparency, privacy and security. We, at Intelicom Lab, IIITD, address research problems pertaining to many of these pillars. In this talk, we will discuss our recent contributions towards continual federated learning (FL) wherein ML models collaboratively adapt to new tasks without forgetting previous knowledge. In particular, we will discuss our recently proposed convergent continual adaptive optimization framework, wherein continual FL is realised  by adapting the learning rate at each edge user. Subsequently, we will discuss the robust FL methods applied to wireless networks, alongside their convergence analysis, followed by a testbed based demonstration of FL over a network of USRPs. About the speaker:  Dr. Ranjitha Prasad is an Assistant Professor in the Department of Electronics and Communications Engineering at the Indraprastha Institute of Information Technology Delhi. She obtained her Ph.D. from Indian Institute of Science in 2015. Her experience is in the general areas of signal processing, Bayesian statistics, and more recently, machine learning and deep neural networks. She has been a postdoctoral researcher at Nanyang Technological University and National University of Singapore, Singapore, and a scientist at TCS Innovation Labs, Delhi. She is the recipient of the Best Ph.D. thesis award (The Seshagiri Kaikini Medal) for 2014- 2015 from the Council of Indian Institute of Science, and the recipient of the Best Paper in the Communications Track at NCC 2014, held at IIT Kanpur. Her current research interests are Federated Learning, Explainable and Responsible AI.   All are cordially invited to attend the invited talk.