no code implementations • 22 Feb 2024 • Norah Alballa, Marco Canini
This research investigates the enhancement of knowledge distillation (KD) processes in pre-trained models, an emerging field in knowledge transfer with significant implications for distributed training and federated learning environments.
no code implementations • 8 Feb 2024 • Mohammed Aljahdali, Ahmed M. Abdelmoniem, Marco Canini, Samuel Horváth
In Federated Learning (FL), forgetting, or the loss of knowledge across rounds, hampers algorithm convergence, particularly in the presence of severe data heterogeneity among clients.
no code implementations • 13 Dec 2023 • Jihao Xin, Ivan Ilin, Shunkang Zhang, Marco Canini, Peter Richtárik
In distributed training, communication often emerges as a bottleneck.
no code implementations • 29 May 2023 • Jihao Xin, Marco Canini, Peter Richtárik, Samuel Horváth
To obtain theoretical guarantees, we generalize the notion of standard unbiased compression operators to incorporate Global-QSGD.
no code implementations • 13 Feb 2023 • Fares Fourati, Salma Kharrat, Vaneet Aggarwal, Mohamed-Slim Alouini, Marco Canini
Federated learning is an emerging machine learning paradigm that enables clients to train collaboratively without exchanging local data.
1 code implementation • 1 Nov 2021 • Ahmed M. Abdelmoniem, Atal Narayan Sahu, Marco Canini, Suhaib A. Fahmy
Federated Learning (FL) enables distributed training by learners using local data, thereby enhancing privacy and reducing communication.
no code implementations • NeurIPS 2021 • Atal Narayan Sahu, Aritra Dutta, Ahmed M. Abdelmoniem, Trambak Banerjee, Marco Canini, Panos Kalnis
Unlike with Top-$k$ sparsifier, we show that hard-threshold has the same asymptotic convergence and linear speedup property as SGD in the convex case and has no impact on the data-heterogeneity in the non-convex case.
1 code implementation • ICLR 2021 • Yuchen Jin, Tianyi Zhou, Liangyu Zhao, Yibo Zhu, Chuanxiong Guo, Marco Canini, Arvind Krishnamurthy
This mutual-training process between BO and the loss-prediction model allows us to limit the training steps invested in the BO search.
no code implementations • 15 Feb 2021 • Ahmed M. Abdelmoniem, Chen-Yu Ho, Pantelis Papageorgiou, Muhammad Bilal, Marco Canini
Federated learning (FL) is becoming a popular paradigm for collaborative learning over distributed, private datasets owned by non-trusting entities.
1 code implementation • 26 Jan 2021 • Ahmed M. Abdelmoniem, Ahmed Elzanaty, Mohamed-Slim Alouini, Marco Canini
Many proposals exploit the compressibility of the gradients and propose lossy compression techniques to speed up the communication stage of distributed training.
1 code implementation • 19 Nov 2019 • Aritra Dutta, El Houcine Bergou, Ahmed M. Abdelmoniem, Chen-Yu Ho, Atal Narayan Sahu, Marco Canini, Panos Kalnis
Compressed communication, in the form of sparsification or quantization of stochastic gradients, is employed to reduce communication costs in distributed data-parallel training of deep neural networks.
1 code implementation • 28 May 2019 • Aritra Dutta, El Houcine Bergou, Yunming Xiao, Marco Canini, Peter Richtárik
In contrast to RNA which computes extrapolation coefficients by (approximately) setting the gradient of the objective function to zero at the extrapolated point, we propose a more direct approach, which we call direct nonlinear acceleration (DNA).
no code implementations • 27 May 2019 • Samuel Horvath, Chen-Yu Ho, Ludovit Horvath, Atal Narayan Sahu, Marco Canini, Peter Richtarik
Our technique is applied individually to all entries of the to-be-compressed update vector and works by randomized rounding to the nearest (negative or positive) power of two, which can be computed in a "natural" way by ignoring the mantissa.
2 code implementations • 22 Feb 2019 • Amedeo Sapio, Marco Canini, Chen-Yu Ho, Jacob Nelson, Panos Kalnis, Changhoon Kim, Arvind Krishnamurthy, Masoud Moshref, Dan R. K. Ports, Peter Richtárik
Training machine learning models in parallel is an increasingly important workload.
1 code implementation • 20 Dec 2017 • Marco Canini, Iosif Salem, Liron Schiff, Elad Michael Schiller, Stefan Schmid
By introducing programmability, automated verification, and innovative debugging tools, Software-Defined Networks (SDNs) are poised to meet the increasingly stringent dependability requirements of today's communication networks.
Networking and Internet Architecture Distributed, Parallel, and Cluster Computing Data Structures and Algorithms