Umang Gupta
YOU?
Author Swipe
View article: Comparison of Rule-based Chat Bots with Different Machine Learning Models
Comparison of Rule-based Chat Bots with Different Machine Learning Models Open
View article: A federated learning architecture for secure and private neuroimaging analysis
A federated learning architecture for secure and private neuroimaging analysis Open
The amount of biomedical data continues to grow rapidly. However, collecting data from multiple sites for joint analysis remains challenging due to security, privacy, and regulatory concerns. To overcome this challenge, we use federated le…
View article: "Define Your Terms" : Enhancing Efficient Offensive Speech Classification with Definition
"Define Your Terms" : Enhancing Efficient Offensive Speech Classification with Definition Open
The propagation of offensive content through social media channels has garnered attention of the research community. Multiple works have proposed various semantically related yet subtle distinct categories of offensive speech. In this work…
View article: “Define Your Terms” : Enhancing Efficient Offensive Speech Classification with Definition
“Define Your Terms” : Enhancing Efficient Offensive Speech Classification with Definition Open
View article: Jointly Reparametrized Multi-Layer Adaptation for Efficient and Private Tuning
Jointly Reparametrized Multi-Layer Adaptation for Efficient and Private Tuning Open
Efficient finetuning of pretrained language transformers is becoming increasingly prevalent for solving natural language processing tasks. While effective, it can still require a large number of tunable parameters. This can be a drawback f…
View article: Transferring Models Trained on Natural Images to 3D MRI via Position Encoded Slice Models
Transferring Models Trained on Natural Images to 3D MRI via Position Encoded Slice Models Open
Transfer learning has remarkably improved computer vision. These advances also promise improvements in neuroimaging, where training set sizes are often small. However, various difficulties arise in directly applying models pretrained on na…
View article: Jointly Reparametrized Multi-Layer Adaptation for Efficient and Private Tuning
Jointly Reparametrized Multi-Layer Adaptation for Efficient and Private Tuning Open
Efficient finetuning of pretrained language transformers is becoming increasingly prevalent for solving natural language processing tasks. While effective, it can still require a large number of tunable parameters. This can be a drawback f…
View article: UNSUPERVISED HARMONIZATION OF BRAIN MRI USING 3D CYCLE GANS AND ITS EFFECT ON BRAIN AGE PREDICTION
UNSUPERVISED HARMONIZATION OF BRAIN MRI USING 3D CYCLE GANS AND ITS EFFECT ON BRAIN AGE PREDICTION Open
Deep learning methods trained on brain MRI data from one scanner or imaging protocol can fail catastrophically when tested on data from other sites or protocols - a problem known as domain shift . To address this, here we propose a domain …
View article: Towards Sparsified Federated Neuroimaging Models via Weight Pruning
Towards Sparsified Federated Neuroimaging Models via Weight Pruning Open
Federated training of large deep neural networks can often be restrictive due to the increasing costs of communicating the updates with increasing model sizes. Various model pruning techniques have been designed in centralized settings to …
View article: Secure & Private Federated Neuroimaging
Secure & Private Federated Neuroimaging Open
The amount of biomedical data continues to grow rapidly. However, collecting data from multiple sites for joint analysis remains challenging due to security, privacy, and regulatory concerns. To overcome this challenge, we use Federated Le…
View article: Federated Progressive Sparsification (Purge, Merge, Tune)+
Federated Progressive Sparsification (Purge, Merge, Tune)+ Open
To improve federated training of neural networks, we develop FedSparsify, a sparsification strategy based on progressive weight magnitude pruning. Our method has several benefits. First, since the size of the network becomes increasingly s…
View article: Mitigating Gender Bias in Distilled Language Models via Counterfactual Role Reversal
Mitigating Gender Bias in Distilled Language Models via Counterfactual Role Reversal Open
Language models excel at generating coherent text, and model compression techniques such as knowledge distillation have enabled their use in resource-constrained settings. However, these models can be biased in multiple ways, including the…
View article: Mitigating Gender Bias in Distilled Language Models via Counterfactual Role Reversal
Mitigating Gender Bias in Distilled Language Models via Counterfactual Role Reversal Open
Umang Gupta, Jwala Dhamala, Varun Kumar, Apurv Verma, Yada Pruksachatkun, Satyapriya Krishna, Rahul Gupta, Kai-Wei Chang, Greg Ver Steeg, Aram Galstyan. Findings of the Association for Computational Linguistics: ACL 2022. 2022.
View article: Attributing Fair Decisions with Attention Interventions
Attributing Fair Decisions with Attention Interventions Open
The widespread use of Artificial Intelligence (AI) in consequential domains, such as health-care and parole decision-making systems, has drawn intense scrutiny on the fairness of these methods. However, ensuring fairness is often insuffici…
View article: Attributing Fair Decisions with Attention Interventions
Attributing Fair Decisions with Attention Interventions Open
The widespread use of Artificial Intelligence (AI) in consequential domains, such as healthcare and parole decision-making systems, has drawn intense scrutiny on the fairness of these methods. However, ensuring fairness is often insufficie…
View article: Secure Neuroimaging Analysis using Federated Learning with Homomorphic Encryption
Secure Neuroimaging Analysis using Federated Learning with Homomorphic Encryption Open
Federated learning (FL) enables distributed computation of machine learning models over various disparate, remote data sources, without requiring to transfer any individual data to a centralized location. This results in an improved genera…
View article: Controllable Guarantees for Fair Outcomes via Contrastive Information Estimation
Controllable Guarantees for Fair Outcomes via Contrastive Information Estimation Open
Controlling bias in training datasets is vital for ensuring equal treatment, or parity, between different groups in downstream applications. A naive solution is to transform the data so that it is statistically independent of group members…
View article: Membership Inference Attacks on Deep Regression Models for Neuroimaging
Membership Inference Attacks on Deep Regression Models for Neuroimaging Open
Ensuring the privacy of research participants is vital, even more so in healthcare environments. Deep learning approaches to neuroimaging require large datasets, and this often necessitates sharing data between multiple sites, which is ant…
View article: Improved Brain Age Estimation With Slice-Based Set Networks
Improved Brain Age Estimation With Slice-Based Set Networks Open
Deep Learning for neuroimaging data is a promising but challenging direction. The high dimensionality of 3D MRI scans makes this endeavor compute and data-intensive. Most conventional 3D neuroimaging methods use 3D-CNN-based architectures …
View article: Comparison of Multiple Battery Chemistries in the Cost Minimization of a Residential Vehicle-to-Grid System
Comparison of Multiple Battery Chemistries in the Cost Minimization of a Residential Vehicle-to-Grid System Open
The economic feasibility of Vehicle-to-Grid (V2G) technology is subject to a lot of debate due to the battery degradation costs associated with it.In this paper, the daily operational cost of a V2G system is calculated for three different …
View article: Controllable Guarantees for Fair Outcomes via Contrastive Information\n Estimation
Controllable Guarantees for Fair Outcomes via Contrastive Information\n Estimation Open
Controlling bias in training datasets is vital for ensuring equal treatment,\nor parity, between different groups in downstream applications. A naive\nsolution is to transform the data so that it is statistically independent of\ngroup memb…
View article: Clustering on Ranked Data for Campaign Selection
Clustering on Ranked Data for Campaign Selection Open
Recently the ranked data are commonly seen in the era of Internet and e-commerce where the consumers give their opinion in the form of ranks of a set of items online. The consumers are asked to put the ranks on items according to their ord…
View article: Intrusion Detection System using Outlier Analysis
Intrusion Detection System using Outlier Analysis Open
The tremendous advancement and increase in usage of internet have increased the number of intruders and hackers over the years. These cyber-attacks increase in complexity and sophistication day by day, and it is thus proposed that data…
View article: Ruuh: A Deep Learning Based Conversational Social Agent
Ruuh: A Deep Learning Based Conversational Social Agent Open
Dialogue systems and conversational agents are becoming increasingly popular in the modern society but building an agent capable of holding intelligent conversation with its users is a challenging problem for artificial intelligence. In th…
View article: Policy Learning for Continuous Space Security Games Using Neural Networks
Policy Learning for Continuous Space Security Games Using Neural Networks Open
A wealth of algorithms centered around (integer) linear programming have been proposed to compute equilibrium strategies in security games with discrete states and actions. However, in practice many domains possess continuous state and act…
View article: A Case Study on using Crowdsourcing for Ambiguous Tasks
A Case Study on using Crowdsourcing for Ambiguous Tasks Open
View article: Deep Generative Dual Memory Network for Continual Learning
Deep Generative Dual Memory Network for Continual Learning Open
Despite advances in deep learning, neural networks can only learn multiple tasks when trained on them jointly. When tasks arrive sequentially, they lose performance on previously learnt tasks. This phenomenon called catastrophic forgetting…
View article: A Sentiment-and-Semantics-Based Approach for Emotion Detection in Textual Conversations
A Sentiment-and-Semantics-Based Approach for Emotion Detection in Textual Conversations Open
Emotions are physiological states generated in humans in reaction to internal or external events. They are complex and studied across numerous fields including computer science. As humans, on reading "Why don't you ever text me!" we can ei…
View article: Emotion Detection from Text
Emotion Detection from Text Open
Emotions are perceptions of changes in the human body such as heart rate, breathing rate, perspiration, and hormone levels. These conscious experiences are complex and studied extensively in different fields including computer science. Lac…
View article: Automated determination of g-mode period spacing of red giant stars
Automated determination of g-mode period spacing of red giant stars Open
The Kepler satellite has provided photometric timeseries data of\nunprecedented length, duty cycle and precision. To fully analyse these data for\nthe tens of thousands of stars observed by Kepler, automated methods are a\nprerequisite. He…