Publications by Rishi Sharma | EPFL Graph Search

Efficient Pyramidal Analysis of Gigapixel Images on a Decentralized Modest Computer Cluster

Anne-Marie Kermarrec, Rafael Pereira Pires, Rishi Sharma

Analyzing gigapixel images is recognized as computationally demanding. In this paper, we introduce PyramidAI, a technique for analyzing gigapixel images with reduced computational cost. The proposed approach adopts a gradual analysis of the image, beginnin ...

Springer Nature Switzerland2025

Boosting Resource-Constrained Federated Learning Systems with Guessed Updates

Anne-Marie Kermarrec, Akash Balasaheb Dhasade, Rafael Pereira Pires, Rishi Sharma

Federated learning (FL) enables a set of client devices to collaboratively train a model without sharing raw data. This process, though, operates under the constrained computation and communication resources of edge devices. These constraints combined with ...

Institute of Electrical and Electronics Engineers (IEEE)2025

Fair Decentralized Learning

Anne-Marie Kermarrec, Rishi Sharma, Marinus Abraham de Vos, Sayan Biswas

Decentralized learning (DL) is an emerging approach that enables nodes to collaboratively train a machine learning model without sharing raw data. In many application domains, such as healthcare, this approach faces challenges due to the high level of hete ...

IEEE2025

Boosting Asynchronous Decentralized Learning with Model Fragmentation

Anne-Marie Kermarrec, Rafael Pereira Pires, Rishi Sharma, Marinus Abraham de Vos, Sayan Biswas

Decentralized learning (DL) is an emerging technique that allows nodes on the web to collaboratively train machine learning models without sharing raw data. Dealing with stragglers, i.e., nodes with slower compute or communication than others, is a key cha ...

ACM2025

Practical Federated Learning without a Server

Anne-Marie Kermarrec, Akash Balasaheb Dhasade, Erick Lavoie, Rishi Sharma, Marinus Abraham de Vos

Federated Learning (FL) enables end-user devices to collaboratively train ML models without sharing raw data, thereby preserving data privacy. In FL, a central parameter server coordinates the learning process by iteratively aggregating the trained models ...

ACM2025

Low-Cost Privacy-Preserving Decentralized Learning

Anne-Marie Kermarrec, Rafael Pereira Pires, Rishi Sharma, Sayan Biswas

Decentralized learning (DL) is an emerging paradigm of collaborative machine learning that enables nodes in a network to train models collectively without sharing their raw data or relying on a central server. This paper introduces Zip-DL, a privacy-aware ...

Privacy Enhancing Technologies Symposium Advisory Board2025

Energy-Aware Decentralized Learning with Intermittent Model Training

Anne-Marie Kermarrec, Akash Balasaheb Dhasade, Rafael Pereira Pires, Rishi Sharma, Marinus Abraham de Vos

SKIPTRAIN is a novel Decentralized Learning (DL) algorithm, which minimizes energy consumption in decentralized learning by strategically skipping some training rounds and substituting them with synchronization rounds. These trainingsilent periods, besides ...

IEEE2024

Revisiting Ensembling in One-Shot Federated Learning

Rachid Guerraoui, Anne-Marie Kermarrec, Akash Balasaheb Dhasade, Rafael Pereira Pires, Nirupam Gupta, Rafaël Benjamin Pinot, Rishi Sharma, Youssef Allouah

Federated learning (FL) is an appealing approach to training machine learning models without sharing raw data. However, standard FL algorithms are iterative and thus induce a significant communication cost. One-shot federated learning (OFL) trades the iter ...

Curran Associates, Inc.2024

Search for high-mass exclusive diphoton production with tagged protons in proton-proton collisions at √s=13 TeV

Patrick Berger, Rakesh Chawla, Ian Smith, Jung Eung Park, Thomas Mountford, Matthias Finger, Matteo Galli, Maria Isabel Cordero Campana, Xiaoqiang Wang, Jian Zhao, James Gonzalo King, Mirko Kovac, Hua Zhang, Joo Yeon Kim, Jae-Woo Choi, Pedro Camilo de Oliveira e Silva, Young Min Kim, Myun Sik Kim, Gabriele Grosso, Qian Wang, Varun Sharma, Michele De Palma, Hyungjun Lee, Chia-Chun Chen, Siyuan Wang, Özdemir Can Kara, Shivom Sharma, Jian Wang, Yi Zhang, Bokyoung Kim, Rahul Kumar Gupta, Ling Zhou, Mithlesh Kumar, Chenyang Wang, Matthias Wolf, Chang-Hyun Park, Anna Mascellani, Yikai Yang, Zhenyu Wang, Seonwoo Lee, Wei-Hsiang Lin, , , , , , , , , , , , , , , , , , , , , , , , , , , ,

Noiseless Privacy-Preserving Decentralized Learning

Anne-Marie Kermarrec, Rafael Pereira Pires, Rishi Sharma, Marinus Abraham de Vos, Sayan Biswas

Decentralized learning (DL) enables collaborative learning without a server and without training data leaving the users' devices. However, the models shared in DL can still be used to infer training data. Conventional defenses such as differential privacy ...

Privacy Enhancing Technologies Symposium Advisory Board2024

Decentralized learning made easy with DecentralizePy

Anne-Marie Kermarrec, Akash Balasaheb Dhasade, Milos Vujasinovic, Rafael Pereira Pires, Rishi Sharma

Decentralized learning (DL) has gained prominence for its potential benefits in terms of scalability, privacy, and fault tolerance. It consists of many nodes that coordinate without a central server and exchange millions of parameters in the inherently ite ...

2023

Epidemic Learning: Boosting Decentralized Learning with Randomized Communication

Rachid Guerraoui, Anne-Marie Kermarrec, Sadegh Farhadkhani, Rafael Pereira Pires, Rishi Sharma, Marinus Abraham de Vos

We present Epidemic Learning ( EL ), a simple yet powerful decentralized learning (DL) algorithm that leverages changing communication topologies to achieve faster model convergence compared to conventional DL approaches. At each round of EL, each node sen ...

2023

Get More for Less in Decentralized Learning Systems

Anne-Marie Kermarrec, Jeffrey Wigger, Akash Balasaheb Dhasade, Milos Vujasinovic, Rafael Pereira Pires, Rishi Sharma

Decentralized learning (DL) systems have been gaining popularity because they avoid raw data sharing by communicating only model parameters, hence preserving data confidentiality. However, the large size of deep neural networks poses a significant challeng ...

IEEE2023