Search

Your search keyword '"Kumar, Ashish"' showing total 10,041 results
Start Over Author "Kumar, Ashish"

Refine your results

more »

more »

more »

more »

more »

more »
10,041 results on '"Kumar, Ashish"'

18. NeuFair: Neural Network Fairness Repair with Dropout

Author

Dasu, Vishnu Asutosh, Kumar, Ashish, Tizpaz-Niari, Saeid, and Tan, Gang

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Software Engineering
Abstract

This paper investigates neuron dropout as a post-processing bias mitigation for deep neural networks (DNNs). Neural-driven software solutions are increasingly applied in socially critical domains with significant fairness implications. While neural networks are exceptionally good at finding statistical patterns from data, they may encode and amplify existing biases from the historical data. Existing bias mitigation algorithms often require modifying the input dataset or the learning algorithms. We posit that the prevalent dropout methods that prevent over-fitting during training by randomly dropping neurons may be an effective and less intrusive approach to improve the fairness of pre-trained DNNs. However, finding the ideal set of neurons to drop is a combinatorial problem. We propose NeuFair, a family of post-processing randomized algorithms that mitigate unfairness in pre-trained DNNs via dropouts during inference after training. Our randomized search is guided by an objective to minimize discrimination while maintaining the model's utility. We show that our design of randomized algorithms is effective and efficient in improving fairness (up to 69%) with minimal or no model performance degradation. We provide intuitive explanations of these phenomena and carefully examine the influence of various hyperparameters of search algorithms on the results. Finally, we empirically and conceptually compare NeuFair to different state-of-the-art bias mitigators., Comment: Paper accepted at ACM ISSTA 2024

Published
2024

25. Pick-or-Mix: Dynamic Channel Sampling for ConvNets

Author

Kumar, Ashish, Kim, Daneul, Park, Jaesik, and Behera, Laxmidhar

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Channel pruning approaches for convolutional neural networks (ConvNets) deactivate the channels, statically or dynamically, and require special implementation. In addition, channel squeezing in representative ConvNets is carried out via 1x1 convolutions which dominates a large portion of computations and network parameters. Given these challenges, we propose an effective multi-purpose module for dynamic channel sampling, namely Pick-or-Mix (PiX), which does not require special implementation. PiX divides a set of channels into subsets and then picks from them, where the picking decision is dynamically made per each pixel based on the input activations. We plug PiX into prominent ConvNet architectures and verify its multi-purpose utilities. After replacing 1x1 channel squeezing layers in ResNet with PiX, the network becomes 25% faster without losing accuracy. We show that PiX allows ConvNets to learn better data representation than widely adopted approaches to enhance networks' representation power (e.g., SE, CBAM, AFF, SKNet, and DWP). We also show that PiX achieves state-of-the-art performance on network downscaling and dynamic channel pruning applications., Comment: Published in Computer Vision and Pattern Recognition (CVPR 2024)

Published
2024

28. COVID-19 Vaccination in Patients with Inborn Errors of Immunity Reduces Hospitalization and Critical Care Needs Related to COVID-19: a USIDNET Report.

Author

McDonnell, John, Cousins, Kimberley, Younger, M, Lane, Adam, Abolhassani, Hassan, Abraham, Roshini, Al-Tamemi, Salem, Aldave-Becerra, Juan, Al-Faris, Eman, Alfaro-Murillo, Alberto, AlKhater, Suzan, Alsaati, Nouf, Doss, Alexa, Anderson, Melissa, Angarola, Ernestina, Ariue, Barbara, Arnold, Danielle, Assaad, Amal, Aytekin, Caner, Bank, Meaghan, Bergerson, Jenna, Bleesing, Jack, Boesing, John, Bouso, Carolina, Brodszki, Nicholas, Cabanillas, Diana, Cady, Carol, Callahan, Meghan, Caorsi, Roberta, Carbone, Javier, Carrabba, Maria, Castagnoli, Riccardo, Catanzaro, Jason, Chan, Samantha, Chandra, Sharat, Chapdelaine, Hugo, Chavoshzadeh, Zahra, Chong, Hey, Connors, Lori, Consonni, Filippo, Correa-Jimenez, Oscar, Cunningham-Rundles, Charlotte, DAstous-Gauthier, Katherine, Delmonte, Ottavia, Demirdag, Yesim, Deshpande, Deepti, Diaz-Cabrera, Natalie, Dimitriades, Victoria, El-Owaidy, Rasha, ElGhazali, Gehad, Al-Hammadi, Suleiman, Fabio, Giovanna, Faure, Astrid, Feng, Jin, Fernandez, James, Fill, Lauren, Franco, Guacira, Frenck, Robert, Fuleihan, Ramsay, Giardino, Giuliana, Galant-Swafford, Jessica, Gambineri, Eleonora, Garabedian, Elizabeth, Geerlinks, Ashley, Goudouris, Ekaterini, Grecco, Octavio, Pan-Hammarström, Qiang, Khani, Hedieh, Hammarström, Lennart, Hartog, Nicholas, Heimall, Jennifer, Hernandez-Molina, Gabriela, Horner, Caroline, Hostoffer, Robert, Hristova, Nataliya, Hsiao, Kuang-Chih, Ivankovich-Escoto, Gabriela, Jaber, Faris, Jalil, Maaz, Jamee, Mahnaz, Jean, Tiffany, Jeong, Stephanie, Jhaveri, Devi, Jordan, Michael, Joshi, Avni, Kalkat, Amanpreet, Kanarek, Henry, Kellner, Erinn, Khojah, Amer, Khoury, Ruby, Kokron, Cristina, Kumar, Ashish, Lecerf, Kelsey, Lehman, Heather, Leiding, Jennifer, Lesmana, Harry, Lim, Xin, Lopes, Joao, López, Ana, and Tarquini, Lucia

Subjects
Immunization, Immunodeficiency, Outcomes, Viruses: respiratory diseases
Abstract

BACKGROUND: The CDC and ACIP recommend COVID-19 vaccination for patients with inborn errors of immunity (IEI). Not much is known about vaccine safety in IEI, and whether vaccination attenuates infection severity in IEI. OBJECTIVE: To estimate COVID-19 vaccination safety and examine effect on outcomes in patients with IEI. METHODS: We built a secure registry database in conjunction with the US Immunodeficiency Network to examine vaccination frequency and indicators of safety and effectiveness in IEI patients. The registry opened on January 1, 2022, and closed on August 19, 2022. RESULTS: Physicians entered data on 1245 patients from 24 countries. The most common diagnoses were antibody deficiencies (63.7%). At least one COVID-19 vaccine was administered to 806 patients (64.7%), and 216 patients received vaccination prior to the development of COVID-19. The most common vaccines administered were mRNA-based (84.0%). Seventeen patients were reported to seek outpatient clinic or emergency room care for a vaccine-related complication, and one patient was hospitalized for symptomatic anemia. Eight hundred twenty-three patients (66.1%) experienced COVID-19 infection. Of these, 156 patients required hospitalization (19.0%), 47 required ICU care (5.7%), and 28 died (3.4%). Rates of hospitalization (9.3% versus 24.4%, p

Published
2024

29. High-Speed Detector For Low-Powered Devices In Aerial Grasping

Author

Kumar, Ashish and Behera, Laxmidhar

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Robotics
Abstract

Autonomous aerial harvesting is a highly complex problem because it requires numerous interdisciplinary algorithms to be executed on mini low-powered computing devices. Object detection is one such algorithm that is compute-hungry. In this context, we make the following contributions: (i) Fast Fruit Detector (FFD), a resource-efficient, single-stage, and postprocessing-free object detector based on our novel latent object representation (LOR) module, query assignment, and prediction strategy. FFD achieves 100FPS@FP32 precision on the latest 10W NVIDIA Jetson-NX embedded device while co-existing with other time-critical sub-systems such as control, grasping, SLAM, a major achievement of this work. (ii) a method to generate vast amounts of training data without exhaustive manual labelling of fruit images since they consist of a large number of instances, which increases the labelling cost and time. (iii) an open-source fruit detection dataset having plenty of very small-sized instances that are difficult to detect. Our exhaustive evaluations on our and MinneApple dataset show that FFD, being only a single-scale detector, is more accurate than many representative detectors, e.g. FFD is better than single-scale Faster-RCNN by 10.7AP, multi-scale Faster-RCNN by 2.3AP, and better than latest single-scale YOLO-v8 by 8AP and multi-scale YOLO-v8 by 0.3 while being considerably faster., Comment: 8 Pages, 9 Figures, 8 Tables, IEEE Robotics and Automation Letters (IEEE RA-L)

Published
2024

Catalog

Books, media, physical & digital resources
See catalog results