Souvik Kundu

Souvik Kundu

Auto-generated profile, based on publicly available information

Souvik Kundu

Intel Labs

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

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

Representative Publications

View All Publications
1 January 2020 in IEEE Transactions on Computers
Verified
Journal Article
Closed access
SKSouvik Kundu
Souvik Kundu
MNMahdi Nazemi
Mahdi Nazemi
MPMassoud Pedram
Massoud Pedram
KCKeith M. Chugg
Keith M. Chugg
PBPeter Beeral
Peter Beeral
The high energy cost of processing deep convolutional neural networks impedes their ubiquitous deployment in energy-constrained platforms such as embedded systems and IoT devices. This work introduces convolutional layers with pre-defined sparse 2D kernels that have support sets that repeat periodically within and across filters. Due to the efficient storage of our periodic sparse kernels, the parameter savings can translate into considerable improvements in energy efficiency due to reduced DRAM accesses, thus promising significant improvements in the trade-off between energy consumption and accuracy for both training and inference. To evaluate this approach, we performed experiments with two widely accepted datasets, CIFAR-10 and Tiny ImageNet in sparse variants of the ResNet18 and VGG16 architectures. Compared to baseline models, our proposed sparse variants require up to ∼82% fewer model parameters with 5.6× fewer FLOPs with negligible loss in accuracy for ResNet18 on CIFAR-10. For VGG16 trained on Tiny ImageNet, our approach requires 5.8× fewer FLOPs and up to ∼83.3% fewer model parameters with a drop in top-5 (top-1) accuracy of only 1.2%(∼2.1%). We also compared the performance of our proposed architectures with that of ShuffleNet and MobileNetV2. Using similar hyperparameters and FLOPs, our ResNet18 variants yield an average accuracy improvement of ∼2.8%.

1 August 2022
Verified
Conference Paper
Closed access
YHYang Hu
Yang Hu
CIConnor Imes
Connor Imes
XZXuanang Zhao
Xuanang Zhao
SKSouvik Kundu
Souvik Kundu
PBPeter Anthony Beerel
Peter Anthony Beerel
JWJohn Paul Walters
John Paul Walters
Deep neural networks with large model sizes achieve state-of-the-art results for tasks in computer vision and natural language processing. However, such models are too compute- or memory-intensive for resource-constrained edge devices. Prior works on parallel and distributed execution primarily focus on training-rather than inference-using homogeneous accelerators in data centers. We propose PipeEdge, a distributed framework for edge systems that uses pipeline parallelism to both speed up inference and enable running larger, more accurate models that otherwise cannot fit on single edge devices. PipeEdge uses an optimal partition strategy that considers heterogeneity in compute, memory, and network bandwidth. Our empirical evaluation demonstrates that PipeEdge achieves 11.88× and 12.78× speedup using 16 edge devices for the ViT-Huge and BERT-Large models, respectively, with no accuracy loss. Similarly, PipeEdge improves throughput for ViT-Huge (which cannot fit in a single device) by 3.93× over a 4-device baseline using 16 edge devices. Finally, we show up to 4.16× throughput improvement over the state-of-the-art PipeDream when using a heterogeneous set of devices.

1 January 2024
Verified
Conference Paper
Open access
PDF file
ZLZeyu Liu
Zeyu Liu
SKSouvik Kundu
Souvik Kundu
ALAnni Li
Anni Li
JWJunrui Wan
Junrui Wan
LJLianghao Jiang
Lianghao Jiang
PBPeter Beerel
Peter Beerel
Zeyu Liu, Souvik Kundu, Anni Li, Junrui Wan, Lianghao Jiang, Peter Beerel. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 2024.