Arnab Neelim Mazumder

Arnab Neelim Mazumder

Auto-generated profile, based on publicly available information

Arnab Neelim Mazumder

University of Maryland, Baltimore County, Baltimore, MD, USA

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
25 November 2021 in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
Verified
Journal Article
Open access
AMArnab Neelim Mazumder
Arnab Neelim Mazumder
JMJian Meng
Jian Meng
HRHasib-Al Rashid
Hasib-Al Rashid
UKUtteja Kallakuri
Utteja Kallakuri
XZXin Zhang
Xin Zhang
TMTinoosh Mohsenin
Tinoosh Mohsenin
Deep neural networks (DNNs) are being prototyped for a variety of artificial intelligence (AI) tasks including computer vision, data analytics, robotics, etc. The efficacy of DNNs coincides with the fact that they can provide state-of-the-art inference accuracy for these applications. However, this advantage comes from the high computational complexity of the DNNs in use. Hence, it is becoming increasingly important to scale these DNNs so that they can fit on resource-constrained hardware and edge devices. The main goal is to allow efficient processing of the DNNs on low-power micro-AI platforms without compromising hardware resources and accuracy. In this work, we aim to provide a comprehensive survey about the recent developments in the domain of energy-efficient deployment of DNNs on micro-AI platforms. To this extent, we look at different neural architecture search strategies as part of micro-AI model design, provide extensive details about model compression and quantization strategies in practice, and finally elaborate on the current hardware approaches towards efficient deployment of the micro-AI models on hardware. The main takeaways for a reader from this article will be understanding of different search spaces to pinpoint the best micro-AI model configuration, ability to interpret different quantization and sparsification techniques, and the realization of the micro-AI models on resource-constrained hardware and different design considerations associated with it.

11 May 2021 in IEEE Design & Test
Verified
Journal Article
Closed access
AMArnab Neelim Mazumder
Arnab Neelim Mazumder
HRHaoran Ren
Haoran Ren
HRHasib-Al Rashid
Hasib-Al Rashid
MHMorteza Hosseini
Morteza Hosseini
VCVandana Chandrareddy
Vandana Chandrareddy
TMTinoosh Mohsenin
Tinoosh Mohsenin
Respiratory diseases are a leading cause of death in the world. This article proposes a deep convolutional neural network (DCNN) that can detect respiratory symptoms in a user using audio input, and with the assistance of additional physiological or demographic information.

1 October 2021 in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Verified
Journal Article
Closed access
MHMorteza Hosseini
Morteza Hosseini
NMNitheesh Kumar Manjunath
Nitheesh Kumar Manjunath
BPBharat Prakash
Bharat Prakash
AMArnab Mazumder
Arnab Mazumder
VCVandana Chandrareddy
Vandana Chandrareddy
TMTinoosh Mohsenin
Tinoosh Mohsenin
In deep convolutional neural networks (DCNNs), model size and computation complexity are two important factors governing throughput and energy efficiency when deployed to hardware for inference. Recent works on compact DCNNs as well as pruning methods are effective, yet with drawbacks. For instance, more than half the size of all MobileNet models lies in their last two layers, mainly because compact separable convolution (CONV) layers are not applicable to their last fully connected (FC) layers. Also, in pruning methods, the compression is gained at the expense of irregularity in the DCNN architecture, which necessitates additional indexing memory to address nonzero weights, thereby increasing memory footprint, decompression delays, and energy consumption. In this article, we propose cyclic sparsely connected (CSC) architectures, with memory/computation complexity of O(NlogN), where N is the number of nodes/channels given a DCNN layer that, contrary to compact depthwise separable layers, can be used as an overlay for both FC and CONV layers of O(N²). Also, contrary to pruning methods, CSC architectures are structurally sparse and require no indexing due to their cyclic nature. We show that both standard convolution and depthwise convolution layers are special cases of the CSC layers, whose mathematical function, along with FC layers, can be unified into one single formulation and whose hardware implementation can be carried out under one arithmetic logic component. We examine the efficacy of the CSC architectures for compression of LeNet, AlexNet, and MobileNet DCNNs with precision ranging from 2 to 32 bits. More specifically, we surge upon the compact 8-bit quantized 0.5 MobileNet V1 and show that by compressing its last two layers with CSC architectures, the model is compressed by ~1.5x with a size of only 873 kB and little accuracy loss. Finally, we design a configurable hardware that implements all types of DCNN layers including FC, CONV, depthwise, CSC-FC, and CSC-CONV indistinguishably within a unified pipeline. We implement the hardware on a tiny Xilinx field-programmable gate array (FPGA) for total on-chip processing of the compressed MobileNet that, compared to the related work, has the highest Inference/J while utilizing the smallest FPGA.

11 May 2024 in ACM Transactions on Embedded Computing Systems
Verified
Journal Article
Open access
AMArnab Neelim Mazumder
Arnab Neelim Mazumder
FSFarshad Safavi
Farshad Safavi
MRMaryam Rahnemoonfar
Maryam Rahnemoonfar
TMTinoosh Mohsenin
Tinoosh Mohsenin
Fine-tuning deep neural networks is pivotal for creating inference modules that can be suitably imported to edge or field-programmable gate array (FPGA) platforms. Traditionally, exploration of different parameters throughout the layers of deep neural networks has been done using grid search and other brute force techniques. Although these methods lead to the optimal choice of network parameters, the search process can be very time consuming and may not consider deployment constraints across different target platforms. This work addresses this problem by proposing Reg-Tune, a regression-based profiling approach to quickly determine the trend of different metrics in relation to hardware deployment of neural networks on tinyML platforms like FPGAs and edge devices. We start by training a handful of configurations belonging to different combinations of \(\mathcal {NN}\scriptstyle \langle q (quantization),\,s (scaling)\rangle \displaystyle\) or \(\mathcal {NN}\scriptstyle \langle r (resolution),\,s\rangle \displaystyle\) workloads to generate the accuracy values respectively for their corresponding application. Next, we deploy these configurations on the target device to generate energy/latency values. According to our hypothesis, the most energy-efficient configuration suitable for deployment on the target device is a function of the variables q , r , and s . Finally, these trained and deployed configurations and their related results are used as data points for polynomial regression with the variables q , r , and s to realize the trend for accuracy/energy/latency on the target device. Our setup allows us to choose the near-optimal energy-consuming or latency-driven configuration for the desired accuracy from the contour profiles of energy/latency across different tinyML device platforms. To this extent, we demonstrate the profiling process for three different case studies and across two platforms for energy and latency fine-tuning. Our approach results in at least 5.7 \(\times\) better energy efficiency when compared to recent implementations for human activity recognition on FPGA and 74.6% reduction in latency for semantic segmentation of aerial imagery on edge devices compared to baseline deployments.

2 March 2021 in IEEE Design & Test
Verified
Journal Article
Closed access
ASAidin Shiri
Aidin Shiri
AMArnab Neelim Mazumder
Arnab Neelim Mazumder
BPBharat Prakash
Bharat Prakash
HHHouman Homayoun
Houman Homayoun
NWNicholas R. Waytowich
Nicholas R. Waytowich
TMTinoosh Mohsenin
Tinoosh Mohsenin
Reinforcement learning guided with language instructions can potentially improve efficiency and better align with overall system design goals. This article introduces a hardware-friendly architecture for training reinforcement learning equipped with natural language instructions.

18 September 2023 in IEEE Micro
Verified
Journal Article
Closed access
AMArnab Neelim Mazumder
Arnab Neelim Mazumder
TMTinoosh Mohsenin
Tinoosh Mohsenin
Fine-tuning Deep Neural Networks (DNNs) for deployment has traditionally relied on computationally intensive methods such as grid search and neural architecture search (NAS), which may not consider hardware-aware metrics. Moreover, it is essential to consider multiple objectives to develop a range of solutions for tinyML hardware deployment with real-time latency and low power constraints. To address these problems, we propose Reg-TuneV2, a systematic approach to fine-tune DNNs for hardware deployment by considering multiple objectives, including accuracy, power, and latency contours. In addition, this approach uses metric learning to achieve smaller and better-suited configurations for deployment, achieving 90.5% accuracy with only 340 KB of memory for keyword spotting on FPGA. When compared to baselines for keyword spotting and image classification on the Nvidia Jetson Nano 4 GB SDK, the proposed method achieves a 14.5× and 101.8× reduction in model size coupled with 2.5× and 5.9× better inference efficiency, respectively.