Soumyajit Gupta
YOU?
Author Swipe
View article: Finding Pareto trade-offs in fair and accurate detection of toxic speech
Finding Pareto trade-offs in fair and accurate detection of toxic speech Open
Introduction. Optimizing NLP models for fairness poses many challenges. Lack of differentiable fairness measures prevents gradient-based loss training or requires surrogate losses that diverge from the true metric of interest. In addition,…
View article: DeepEyeNet: Adaptive Genetic Bayesian Algorithm Based Hybrid ConvNeXtTiny Framework For Multi-Feature Glaucoma Eye Diagnosis
DeepEyeNet: Adaptive Genetic Bayesian Algorithm Based Hybrid ConvNeXtTiny Framework For Multi-Feature Glaucoma Eye Diagnosis Open
Glaucoma is a leading cause of irreversible blindness worldwide, emphasizing the critical need for early detection and intervention. In this paper, we present DeepEyeNet, a novel and comprehensive framework for automated glaucoma detection…
View article: Fairly Accurate: Fairness-aware Multi-group Target Detection in Online Discussion
Fairly Accurate: Fairness-aware Multi-group Target Detection in Online Discussion Open
Target-group detection is the task of detecting which group(s) a social media post is ``directed at or about'', with various applications, such as targeted-marketing. In this work, we focus on the fairness implications of target-group dete…
View article: Same Same, But Different: Conditional Multi-Task Learning for Demographic-Specific Toxicity Detection
Same Same, But Different: Conditional Multi-Task Learning for Demographic-Specific Toxicity Detection Open
Algorithmic bias often arises as a result of differential subgroup validity, in which predictive relationships vary across groups. For example, in toxic language detection, comments targeting different demographic groups can vary markedly …
View article: Same Same, But Different: Conditional Multi-Task Learning for Demographic-Specific Toxicity Detection
Same Same, But Different: Conditional Multi-Task Learning for Demographic-Specific Toxicity Detection Open
Algorithmic bias often arises as a result of differential subgroup validity, in which predictive relationships vary across groups. For example, in toxic language detection, comments targeting different demographic groups can vary markedly …
View article: HOFS: Higher order mutual information approximation for feature selection in R
HOFS: Higher order mutual information approximation for feature selection in R Open
View article: Finding Pareto Trade-offs in Fair and Accurate Detection of Toxic Speech
Finding Pareto Trade-offs in Fair and Accurate Detection of Toxic Speech Open
Optimizing NLP models for fairness poses many challenges. Lack of differentiable fairness measures prevents gradient-based loss training or requires surrogate losses that diverge from the true metric of interest. In addition, competing obj…
View article: Hofs: Higher Order Mutual Information Approximation for Feature Selection in R
Hofs: Higher Order Mutual Information Approximation for Feature Selection in R Open
View article: Scalable Unidirectional Pareto Optimality for Multi-Task Learning with Constraints
Scalable Unidirectional Pareto Optimality for Multi-Task Learning with Constraints Open
Multi-objective optimization (MOO) problems require balancing competing objectives, often under constraints. The Pareto optimal solution set defines all possible optimal trade-offs over such objectives. In this work, we present a novel met…
View article: Scalable Unidirectional Pareto Optimality for Multi-Task Learning with\n Constraints
Scalable Unidirectional Pareto Optimality for Multi-Task Learning with\n Constraints Open
Multi-objective optimization (MOO) problems require balancing competing\nobjectives, often under constraints. The Pareto optimal solution set defines\nall possible optimal trade-offs over such objectives. In this work, we present\na novel …
View article: Towards a Biomanufactory on Mars
Towards a Biomanufactory on Mars Open
A crewed mission to and from Mars may include an exciting array of enabling biotechnologies that leverage inherent mass, power, and volume advantages over traditional abiotic approaches. In this perspective, we articulate the scientific an…
View article: Tail-Net: Extracting Lowest Singular Triplets for Big Data Applications
Tail-Net: Extracting Lowest Singular Triplets for Big Data Applications Open
SVD serves as an exploratory tool in identifying the dominant features in the form of top rank-r singular factors corresponding to the largest singular values. For Big Data applications it is well known that Singular Value Decomposition (S…
View article: SCA-Net: A Self-Correcting Two-Layer Autoencoder for Hyper-spectral\n Unmixing
SCA-Net: A Self-Correcting Two-Layer Autoencoder for Hyper-spectral\n Unmixing Open
Hyperspectral unmixing involves separating a pixel as a weighted combination\nof its constituent endmembers and corresponding fractional abundances, with the\ncurrent state of the art results achieved by neural models on benchmark\ndataset…
View article: SCA-Net: A Self-Correcting Two-Layer Autoencoder for Hyper-spectral Unmixing
SCA-Net: A Self-Correcting Two-Layer Autoencoder for Hyper-spectral Unmixing Open
Hyperspectral unmixing involves separating a pixel as a weighted combination of its constituent endmembers and corresponding fractional abundances, with the current state of the art results achieved by neural models on benchmark datasets. …
View article: A Hybrid 2-stage Neural Optimization for Pareto Front Extraction
A Hybrid 2-stage Neural Optimization for Pareto Front Extraction Open
Classification, recommendation, and ranking problems often involve competing goals with additional constraints (e.g., to satisfy fairness or diversity criteria). Such optimization problems are quite challenging, often involving non-convex …
View article: Towards a Biomanufactory on Mars
Towards a Biomanufactory on Mars Open
A crewed mission to and from Mars may include an exciting array of enabling biotechnologies that leverage inherent mass, power, and volume advantages over traditional abiotic approaches. In this perspective, we articulate the scientific an…
View article: Streaming Singular Value Decomposition for Big Data Applications.
Streaming Singular Value Decomposition for Big Data Applications. Open
View article: Range-Net: A High Precision Streaming SVD for Big Data Applications
Range-Net: A High Precision Streaming SVD for Big Data Applications Open
In a Big Data setting computing the dominant SVD factors is restrictive due to the main memory requirements. Recently introduced streaming Randomized SVD schemes work under the restrictive assumption that the singular value spectrum of the…
View article: Extracting Optimal Solution Manifolds using Constrained Neural Optimization
Extracting Optimal Solution Manifolds using Constrained Neural Optimization Open
Constrained Optimization solution algorithms are restricted to point based solutions. In practice, single or multiple objectives must be satisfied, wherein both the objective function and constraints can be non-convex resulting in multiple…
View article: Prevention is Better than Cure: Handling Basis Collapse and Transparency in Dense Networks
Prevention is Better than Cure: Handling Basis Collapse and Transparency in Dense Networks Open
Dense nets are an integral part of any classification and regression problem. Recently, these networks have found a new application as solvers for known representations in various domains. However, one crucial issue with dense nets is it's…
View article: TIME: A Transparent, Interpretable, Model-Adaptive and Explainable Neural Network for Dynamic Physical Processes
TIME: A Transparent, Interpretable, Model-Adaptive and Explainable Neural Network for Dynamic Physical Processes Open
Partial Differential Equations are infinite dimensional encoded representations of physical processes. However, imbibing multiple observation data towards a coupled representation presents significant challenges. We present a fully convolu…
View article: Fast Probabilistic Uncertainty Quantification and Sensitivity Analysis of a Mars Life Support System Model
Fast Probabilistic Uncertainty Quantification and Sensitivity Analysis of a Mars Life Support System Model Open
Mars life support system models consist of numerous mission-critical, interrelated, and scenario-specific parameters. The large size and involved nature of these models make them computationally expensive, with parameters that are subject …
View article: A Streaming model for Generalized Rayleigh with extension to Minimum Noise Fraction
A Streaming model for Generalized Rayleigh with extension to Minimum Noise Fraction Open
The Rayleigh quotient optimization is the maximization of a rational function, or a max-min problem, with simultaneous maximization of the numerator function and minimization of the denominator function. Here, we describe a low-rank, strea…
View article: A fully automated, faster noise rejection approach to increasing the analytical capability of chemical imaging for digital histopathology
A fully automated, faster noise rejection approach to increasing the analytical capability of chemical imaging for digital histopathology Open
Chemical hyperspectral imaging (HSI) data is naturally high dimensional and large. There are thus inherent manual trade-offs in acquisition time, and the quality of data. Minimum Noise Fraction (MNF) developed by Green et al. [1] has been …
View article: A Fully Automated, Faster Noise Reduction Approach to Increasing the Analytical Capability of Chemical Imaging for Digital Histopathology
A Fully Automated, Faster Noise Reduction Approach to Increasing the Analytical Capability of Chemical Imaging for Digital Histopathology Open
High dimensional data, for example from infrared spectral imaging, involves an inherent trade-off in the acquisition time and quality of spatial-spectral data. Minimum Noise Fraction (MNF) developed by Green et al . [1] has been extensivel…
View article: Efficient Clustering-Based Noise Covariance Estimation for Maximum Noise Fraction
Efficient Clustering-Based Noise Covariance Estimation for Maximum Noise Fraction Open
View article: Optimization of integrated microalgal biorefinery producing fuel and value‐added products
Optimization of integrated microalgal biorefinery producing fuel and value‐added products Open
An integrated microalgal biorefinery is desirable from an economic standpoint but challenging to synthesize, due to diversity of options. This work uses a model‐based optimization approach to address this challenge in a systematic manner. …
View article: Higher Order Mutual Information Approximation for Feature Selection
Higher Order Mutual Information Approximation for Feature Selection Open
Feature selection is a process of choosing a subset of relevant features so that the quality of prediction models can be improved. An extensive body of work exists on information-theoretic feature selection, based on maximizing Mutual Info…