Kaushik Roy
Purdue University
H-index: 123
North America-United States
Description
Kaushik Roy, With an exceptional h-index of 123 and a recent h-index of 72 (since 2020), a distinguished researcher at Purdue University, specializes in the field of Neuromorphic computing, Machine Learning, ML hardware, Neuro-mimetic devices, Energy-efficient computing.
His recent articles reflect a diverse array of research interests and contributions to the field:
EV-Planner: Energy-Efficient Robot Navigation via Event-Based Physics-Guided Neuromorphic Planner
Prompt-Based Bias Calibration for Better Zero/Few-Shot Learning of Language Models
AdaGossip: Adaptive Consensus Step-size for Decentralized Deep Learning with Communication Compression
Verifix: Post-Training Correction to Improve Label Noise Robustness with Verified Samples
HALSIE: Hybrid Approach to Learning Segmentation by Simultaneously Exploiting Image and Event Modalities
Global update tracking: A decentralized learning algorithm for heterogeneous data
Vehicle battery current sensing system
Sparsity-aware reconfigurable compute-in-memory (CIM) static random access memory (SRAM)
Professor Information
University | Purdue University |
---|---|
Position | Professor of Electrical and Computer Engineering |
Citations(all) | 65501 |
Citations(since 2020) | 23695 |
Cited By | 51661 |
hIndex(all) | 123 |
hIndex(since 2020) | 72 |
i10Index(all) | 813 |
i10Index(since 2020) | 386 |
University Profile Page | Purdue University |
Research & Interests List
Neuromorphic computing
Machine Learning
ML hardware
Neuro-mimetic devices
Energy-efficient computing
Top articles of Kaushik Roy
EV-Planner: Energy-Efficient Robot Navigation via Event-Based Physics-Guided Neuromorphic Planner
Vision-based object tracking is an essential precursor to performing autonomous aerial navigation in order to avoid obstacles. Biologically inspired neuromorphic event cameras are emerging as a powerful alternative to frame-based cameras, due to their ability to asynchronously detect varying intensities (even in poor lighting conditions), high dynamic range, and robustness to motion blur. Spiking neural networks (SNNs) have gained traction for processing events asynchronously in an energy-efficient manner. On the other hand, physics-based artificial intelligence (AI) has gained prominence recently, as they enable embedding system knowledge via physical modeling inside traditional analog neural networks (ANNs). In this letter, we present an event-based physics-guided neuromorphic planner (EV-Planner) to perform obstacle avoidance using neuromorphic event cameras and physics-based AI. We consider …
Authors
Sourav Sanyal,Rohan Kumar Manna,Kaushik Roy
Journal
IEEE Robotics and Automation Letters
Published Date
2024/1/8
Prompt-Based Bias Calibration for Better Zero/Few-Shot Learning of Language Models
Prompt learning is susceptible to intrinsic bias present in pre-trained language models (LMs), resulting in sub-optimal performance of prompt-based zero/few-shot learning. In this work, we propose a null-input prompting method to calibrate intrinsic bias encoded in pre-trained LMs. Different from prior efforts that address intrinsic bias primarily for social fairness and often involve excessive computational cost, our objective is to explore enhancing LMs' performance in downstream zero/few-shot learning while emphasizing the efficiency of intrinsic bias calibration. Specifically, we leverage a diverse set of auto-selected null-meaning inputs generated from GPT-4 to prompt pre-trained LMs for intrinsic bias probing. Utilizing the bias-reflected probability distribution, we formulate a distribution disparity loss for bias calibration, where we exclusively update bias parameters ( of total parameters) of LMs towards equal probability distribution. Experimental results show that the calibration promotes an equitable starting point for LMs while preserving language modeling abilities. Across a wide range of datasets, including sentiment analysis and topic classification, our method significantly improves zero/few-shot learning performance of LMs for both in-context learning and prompt-based fine-tuning (on average and , respectively).
Authors
Kang He,Yinghan Long,Kaushik Roy
Journal
arXiv preprint arXiv:2402.10353
Published Date
2024/2/15
AdaGossip: Adaptive Consensus Step-size for Decentralized Deep Learning with Communication Compression
Decentralized learning is crucial in supporting on-device learning over large distributed datasets, eliminating the need for a central server. However, the communication overhead remains a major bottleneck for the practical realization of such decentralized setups. To tackle this issue, several algorithms for decentralized training with compressed communication have been proposed in the literature. Most of these algorithms introduce an additional hyper-parameter referred to as consensus step-size which is tuned based on the compression ratio at the beginning of the training. In this work, we propose AdaGossip, a novel technique that adaptively adjusts the consensus step-size based on the compressed model differences between neighboring agents. We demonstrate the effectiveness of the proposed method through an exhaustive set of experiments on various Computer Vision datasets (CIFAR-10, CIFAR-100 …
Authors
Sai Aparna Aketi,Abolfazl Hashemi,Kaushik Roy
Journal
arXiv e-prints
Published Date
2024/4
Verifix: Post-Training Correction to Improve Label Noise Robustness with Verified Samples
Label corruption, where training samples have incorrect labels, can significantly degrade the performance of machine learning models. This corruption often arises from non-expert labeling or adversarial attacks. Acquiring large, perfectly labeled datasets is costly, and retraining large models from scratch when a clean dataset becomes available is computationally expensive. To address this challenge, we propose Post-Training Correction, a new paradigm that adjusts model parameters after initial training to mitigate label noise, eliminating the need for retraining. We introduce Verifix, a novel Singular Value Decomposition (SVD) based algorithm that leverages a small, verified dataset to correct the model weights using a single update. Verifix uses SVD to estimate a Clean Activation Space and then projects the model's weights onto this space to suppress activations corresponding to corrupted data. We demonstrate Verifix's effectiveness on both synthetic and real-world label noise. Experiments on the CIFAR dataset with 25% synthetic corruption show 7.36% generalization improvements on average. Additionally, we observe generalization improvements of up to 2.63% on naturally corrupted datasets like WebVision1.0 and Clothing1M.
Authors
Sangamesh Kodge,Deepak Ravikumar,Gobinda Saha,Kaushik Roy
Journal
arXiv preprint arXiv:2403.08618
Published Date
2024/3/13
HALSIE: Hybrid Approach to Learning Segmentation by Simultaneously Exploiting Image and Event Modalities
Event cameras detect changes in per-pixel intensity to generate asynchronous `event streams'. They offer great potential for accurate semantic map retrieval in real-time autonomous systems owing to their much higher temporal resolution and high dynamic range (HDR) compared to conventional cameras. However, existing implementations for event-based segmentation suffer from sub-optimal performance since these temporally dense events only measure the varying component of a visual signal, limiting their ability to encode dense spatial context compared to frames. To address this issue, we propose a hybrid end-to-end learning framework HALSIE, utilizing three key concepts to reduce inference cost by up to versus prior art while retaining similar performance: First, a simple and efficient cross-domain learning scheme to extract complementary spatio-temporal embeddings from both frames and events. Second, a specially designed dual-encoder scheme with Spiking Neural Network (SNN) and Artificial Neural Network (ANN) branches to minimize latency while retaining cross-domain feature aggregation. Third, a multi-scale cue mixer to model rich representations of the fused embeddings. These qualities of HALSIE allow for a very lightweight architecture achieving state-of-the-art segmentation performance on DDD-17, MVSEC, and DSEC-Semantic datasets with up to higher parameter efficiency and favorable inference cost (17.9mJ per cycle). Our ablation study also brings new insights into effective design choices that can prove beneficial for research across other vision tasks.
Authors
Shristi Das Biswas,Adarsh Kosta,Chamika Liyanagedera,Marco Apolinario,Kaushik Roy
Journal
arXiv preprint arXiv:2211.10754
Published Date
2022/11/19
Global update tracking: A decentralized learning algorithm for heterogeneous data
Decentralized learning enables the training of deep learning models over large distributed datasets generated at different locations, without the need for a central server. However, in practical scenarios, the data distribution across these devices can be significantly different, leading to a degradation in model performance. In this paper, we focus on designing a decentralized learning algorithm that is less susceptible to variations in data distribution across devices. We propose Global Update Tracking (GUT), a novel tracking-based method that aims to mitigate the impact of heterogeneous data in decentralized learning without introducing any communication overhead. We demonstrate the effectiveness of the proposed technique through an exhaustive set of experiments on various Computer Vision datasets (CIFAR-10, CIFAR-100, Fashion MNIST, and ImageNette), model architectures, and network topologies. Our experiments show that the proposed method achieves state-of-the-art performance for decentralized learning on heterogeneous data via a 1-6% improvement in test accuracy compared to other existing techniques.
Authors
Sai Aparna Aketi,Abolfazl Hashemi,Kaushik Roy
Journal
Advances in Neural Information Processing Systems
Published Date
2024/2/13
Vehicle battery current sensing system
A method of sensing a current in a conductor includes controlling a digital to analog converter output to cancel residual offset voltage in a magnetic tunnel junction device prior to sensing the current with the magnetic tunnel junction device. The method includes switching input to the magnetic tunnel junction device between a fixed voltage and an output of a digital to analog converter while switching input to a band pass filter between a lower and an upper voltage output of the magnetic tunnel junction device. The output of the digital to analog converter is modified to provide a low-amplitude unsaturated sine-wave at an output of the band pass filter, at which point changes in the output of the band pass filter are associated with the amount of current in a sensed conductor.
Published Date
2024/4/2
Sparsity-aware reconfigurable compute-in-memory (CIM) static random access memory (SRAM)
Sparsity-aware reconfiguration compute-in-memory (CIM) static random access memory (SRAM) systems are disclosed. In one aspect, a reconfigurable precision succession approximation register (SAR) analog-to-digital converter (ADC) that has the ability to form (n+ m) bit precision using n-bit and m-bit sub-ADCs is provided. By controlling which sub-ADCs are used based on data sparsity, precision may be maintained as needed while providing a more energy efficient design.
Published Date
2024/3/12
Professor FAQs
What is Kaushik Roy's h-index at Purdue University?
The h-index of Kaushik Roy has been 72 since 2020 and 123 in total.
What are Kaushik Roy's top articles?
The articles with the titles of
EV-Planner: Energy-Efficient Robot Navigation via Event-Based Physics-Guided Neuromorphic Planner
Prompt-Based Bias Calibration for Better Zero/Few-Shot Learning of Language Models
AdaGossip: Adaptive Consensus Step-size for Decentralized Deep Learning with Communication Compression
Verifix: Post-Training Correction to Improve Label Noise Robustness with Verified Samples
HALSIE: Hybrid Approach to Learning Segmentation by Simultaneously Exploiting Image and Event Modalities
Global update tracking: A decentralized learning algorithm for heterogeneous data
Vehicle battery current sensing system
Sparsity-aware reconfigurable compute-in-memory (CIM) static random access memory (SRAM)
...
are the top articles of Kaushik Roy at Purdue University.
What are Kaushik Roy's research interests?
The research interests of Kaushik Roy are: Neuromorphic computing, Machine Learning, ML hardware, Neuro-mimetic devices, Energy-efficient computing
What is Kaushik Roy's total number of citations?
Kaushik Roy has 65,501 citations in total.
What are the co-authors of Kaushik Roy?
The co-authors of Kaushik Roy are Anand Raghunathan, Muhammad Ashraful Alam, Swarup Bhunia, Saibal Mukhopadhyay, Sumeet Gupta, Priyadarshini (Priya) Panda.