Ravikumar Balakrishnan

Dr. Ravikumar Balakrishnan

Intel Corporation
orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
RBRavikumar Balakrishnan
Ravikumar Balakrishnan
MAMustafa Akdeniz
Mustafa Akdeniz
SDSagar Dhakal
Sagar Dhakal
AAArjun Anand
Arjun Anand
AZAriela Zeira
Ariela Zeira
NHNageen Himayat
Nageen Himayat
Client and Internet of Things devices are increasingly equipped with the ability to sense, process, and communicate data with high efficiency. This is resulting in a major shift in machine learning (ML) computation at the network edge. Distributed learning approaches such as federated learning that move ML training to end devices have emerged, promising lower latency and bandwidth costs and enhanced privacy of end users’ data. However, new challenges that arise from the heterogeneous nature of the devices’ communication rates, compute capabilities, and the limited observability of the training data at each device must be addressed. All these factors can significantly affect the training performance in terms of overall accuracy, model fairness, and convergence time. We present compute-communication and data importance-aware resource management schemes optimizing these metrics and evaluate the training performance on benchmark datasets. We also develop a federated meta-learning solution, based on task similarity, that serves as a sample efficient initialization for federated learning, as well as improves model personalization and generalization across non-IID (independent, identically distributed) data. We present experimental results on benchmark federated learning datasets to highlight the performance gains of the proposed methods in comparison to the well-known federated averaging algorithm and its variants.

1 November 2020
Verified
Conference Paper
Closed access
AAArjun Anand
Arjun Anand
RBRavikumar Balakrishnan
Ravikumar Balakrishnan
VSV. Srinivasa Somayazulu
V. Srinivasa Somayazulu
RVRath Vannithamby
Rath Vannithamby
Recently Deep Reinforcement Learning (DRL) based techniques have been applied to scheduling algorithms at wireless Base Stations (BS) with promising results. The Deep Neural Networks (DNN) used in the DRL approaches have a fixed size input, and this poses a problem in real usage scenarios where the system state that includes number of users can be variable. In this paper we address this important problem by proposing a model-assisted DRL approach, which combines a model based solution with DRL based scheduling and can address both the variable input size and the DNN complexity issues. We show that this approach can achieve performance comparable to a DRL algorithm where a larger DNN trained for the worst-case scenario of the maximum number of users in the system. From a practical perspective our approach of using a model-assisted DRL has the following advantages: 1) There is no strict upper bound on the maximum number of users that can be supported by the system as opposed to the case when we train for the worst case number of users, 2) Training is faster and more energy efficient, and 3) leads to low-complex DNNs achieve comparable performance.


3 September 2024 in arXiv
Verified
Preprint
Open access
PDF file
SDShenghong Dai
Shenghong Dai
JSJy-Yong Sohn
Jy-Yong Sohn
YCYicong Chen
Yicong Chen
SAS M Iftekharul Alam
S M Iftekharul Alam
RBRavikumar Balakrishnan
Ravikumar Balakrishnan
KLKangwook Lee
Kangwook Lee
Continual Federated Learning (CFL) is essential for enabling real-world applications where multiple decentralized clients adaptively learn from continuous data streams. A significant challenge in CFL is mitigating catastrophic forgetting, where models lose previously acquired knowledge when learning new information. Existing approaches often face difficulties due to the constraints of device storage capacities and the heterogeneous nature of data distributions among clients. While some CFL algorithms have addressed these challenges, they frequently rely on unrealistic assumptions about the availability of task boundaries (i.e., knowing when new tasks begin). To address these limitations, we introduce Fed-A-GEM, a federated adaptation of the A-GEM method (Chaudhry et al., 2019), which employs a buffer-based gradient projection approach. Fed-A-GEM alleviates catastrophic forgetting by leveraging local buffer samples and aggregated buffer gradients, thus preserving knowledge across multiple clients. Our method is combined with existing CFL techniques, enhancing their performance in the CFL context. Our experiments on standard benchmarks show consistent performance improvements across diverse scenarios. For example, in a task-incremental learning scenario using the CIFAR-100 dataset, our method can increase the accuracy by up to 27%. Our code is available at https://github.com/shenghongdai/Fed-A-GEM.