Kartikey Sharma - Academia.edu (original) (raw)

Uploads

Papers by Kartikey Sharma

Journal of Resources, Energy and Development

There are various facets that can play a telling role in mitigating the aspect of climate change ... more There are various facets that can play a telling role in mitigating the aspect of climate change and the linkages associated with it. While largely the global focus has been on energy transition and moving away from fossil fuel-based consumption, a lot needs to be explored on how we as a society need to move towards responsible production and consumption patterns and the elements that can enable it. The phenomenon of climate change and sustainability is rooted in the products/services and their eventual mass consumption that warrant their degradation. Given the wide range of interdependencies that exist within the ecosystem of sustainability, this paper brings to light the need for analysing our consumption and production patterns through the lens of complex system mechanism, and its subsequent impact on India and the world.

Frontiers in artificial intelligence, May 2, 2022

Sharing labeled data is crucial to acquire large datasets for various Deep Learning applications.... more Sharing labeled data is crucial to acquire large datasets for various Deep Learning applications. In medical imaging, this is often not feasible due to privacy regulations. Whereas anonymization would be a solution, standard techniques have been shown to be partially reversible. Here, synthetic data using a Generative Adversarial Network (GAN) with differential privacy guarantees could be a solution to ensure the patient's privacy while maintaining the predictive properties of the data. In this study, we implemented a Wasserstein GAN (WGAN) with and without differential privacy guarantees to generate privacy-preserving labeled Time-of-Flight Magnetic Resonance Angiography (TOF-MRA) image patches for brain vessel segmentation. The synthesized image-label pairs were used to train a U-net which was evaluated in terms of the segmentation performance on real patient images from two different datasets. Additionally, the Fréchet Inception Distance (FID) was calculated between the generated images and the real images to assess their similarity. During the evaluation using the U-Net and the FID, we explored the effect of different levels of privacy which was represented by the parameter ǫ. With stricter privacy guarantees, the segmentation performance and the similarity to the real patient images in terms of FID decreased. Our best segmentation model, trained on synthetic and private data, achieved a Dice Similarity Coefficient (DSC) of 0.75 for ǫ = 7.4 compared to 0.84 for ǫ = ∞ in a brain vessel segmentation paradigm (DSC of 0.69 and 0.88 on the second test set, respectively). We identified a threshold of ǫ < 5 for which the Kossen et al. Labeled TOF-MRA With Differential Privacy performance (DSC < 0.61) became unstable and not usable. Our synthesized labeled TOF-MRA images with strict privacy guarantees retained predictive properties necessary for segmenting the brain vessels. Although further research is warranted regarding generalizability to other imaging modalities and performance improvement, our results mark an encouraging first step for privacy-preserving data sharing in medical imaging.

INFORMS Journal on Optimization, 2022

Robust optimization methods have shown practical advantages in a wide range of decision-making ap... more Robust optimization methods have shown practical advantages in a wide range of decision-making applications under uncertainty. Recently, their efficacy has been extended to multiperiod settings. Current approaches model uncertainty either independent of the past or in an implicit fashion by budgeting the aggregate uncertainty. In many applications, however, past realizations directly influence future uncertainties. For this class of problems, we develop a modeling framework that explicitly incorporates this dependence via connected uncertainty sets, whose parameters at each period depend on previous uncertainty realizations. To find optimal here-and-now solutions, we reformulate robust and distributionally robust constraints for popular set structures and demonstrate this modeling framework numerically on broadly applicable knapsack and portfolio-optimization problems.

Now a days, environmental issues are increasing at alarming rate. We have reviewed manyresearch p... more Now a days, environmental issues are increasing at alarming rate. We have reviewed manyresearch papers on climate change and environment issues..In review paper an overview of manyresearches by many researchers on environment and climate change. This paper reports acomprehensive literature review for 1991–2019 (inclusive), the years in which this topicappeared in scientific journals. In our review paper includes all ecological factors, criticalthinking on environment, space climate, global environment change and many more.

Optimization under Variable Uncertainty Kartikey Sharma In this dissertation, we study models and... more Optimization under Variable Uncertainty Kartikey Sharma In this dissertation, we study models and methods to address uncertainties that can vary in optimization problems. Robust optimization is a popular approach for optimization under uncertainty, especially if limited information is available about the distribution of the uncertainty. It models the uncertainty through sets and finds a robust optimal solution that is feasible for all realizations of the uncertainty within the set and is optimal for the worst-case realization. The structure of these sets determines the complexity of the resulting optimization problem. In most models, the uncertainty set is assumed to be exogenous i.e., predetermined and is unaffected by decisions or other uncertainty realizations in the problem. This thesis introduces endogenous uncertainty models, which may be affected by decisions that are made in the problem or by other uncertainty realizations within the problem. In the first chapter, we take a step towards generalizing robust linear optimization to problems with decision dependent uncertainties. We show these problems to be NPcomplete in general settings. To alleviate these computational inefficiencies, we introduce Dedication Table of Contents List of Tables List of Figures Chapter 1. Introduction Chapter 2. Decision Dependent Uncertainty List of Tables 2.1 Size of Big-M formulation of (RO-DDU) for U i (x) with respect to (i) x ∈ {0, 1} n and (ii) x ∈ R n with x i taking s possible values: dim(y) = p, K constraints in U i (x), and m constraints in the complete problem. 2.2 Comparison of (LC) reformulations for the set U Π (x)

SIAM Journal on Optimization, 2018

The efficacy of robust optimization spans a variety of settings with uncertainties bounded in pre... more The efficacy of robust optimization spans a variety of settings with uncertainties bounded in predetermined sets. In many applications, uncertainties are affected by decisions and cannot be modeled with current frameworks. This paper takes a step towards generalizing robust linear optimization to problems with decisiondependent uncertainties. In general settings, we show these problems to be NP-complete. To alleviate the computational inefficiencies, we introduce a class of uncertainty sets whose size depends on binary decisions. We propose reformulations that improve upon alternative standard linearization techniques. To illustrate the advantages of this framework, a shortest path problem is discussed, where the uncertain arc lengths are affected by decisions. Beyond the modeling and performance advantages, the proposed notion of proactive uncertainty control also mitigates over conservatism of current robust optimization approaches.

Journal of Resources, Energy and Development

There are various facets that can play a telling role in mitigating the aspect of climate change ... more There are various facets that can play a telling role in mitigating the aspect of climate change and the linkages associated with it. While largely the global focus has been on energy transition and moving away from fossil fuel-based consumption, a lot needs to be explored on how we as a society need to move towards responsible production and consumption patterns and the elements that can enable it. The phenomenon of climate change and sustainability is rooted in the products/services and their eventual mass consumption that warrant their degradation. Given the wide range of interdependencies that exist within the ecosystem of sustainability, this paper brings to light the need for analysing our consumption and production patterns through the lens of complex system mechanism, and its subsequent impact on India and the world.

Frontiers in artificial intelligence, May 2, 2022

Sharing labeled data is crucial to acquire large datasets for various Deep Learning applications.... more Sharing labeled data is crucial to acquire large datasets for various Deep Learning applications. In medical imaging, this is often not feasible due to privacy regulations. Whereas anonymization would be a solution, standard techniques have been shown to be partially reversible. Here, synthetic data using a Generative Adversarial Network (GAN) with differential privacy guarantees could be a solution to ensure the patient's privacy while maintaining the predictive properties of the data. In this study, we implemented a Wasserstein GAN (WGAN) with and without differential privacy guarantees to generate privacy-preserving labeled Time-of-Flight Magnetic Resonance Angiography (TOF-MRA) image patches for brain vessel segmentation. The synthesized image-label pairs were used to train a U-net which was evaluated in terms of the segmentation performance on real patient images from two different datasets. Additionally, the Fréchet Inception Distance (FID) was calculated between the generated images and the real images to assess their similarity. During the evaluation using the U-Net and the FID, we explored the effect of different levels of privacy which was represented by the parameter ǫ. With stricter privacy guarantees, the segmentation performance and the similarity to the real patient images in terms of FID decreased. Our best segmentation model, trained on synthetic and private data, achieved a Dice Similarity Coefficient (DSC) of 0.75 for ǫ = 7.4 compared to 0.84 for ǫ = ∞ in a brain vessel segmentation paradigm (DSC of 0.69 and 0.88 on the second test set, respectively). We identified a threshold of ǫ < 5 for which the Kossen et al. Labeled TOF-MRA With Differential Privacy performance (DSC < 0.61) became unstable and not usable. Our synthesized labeled TOF-MRA images with strict privacy guarantees retained predictive properties necessary for segmenting the brain vessels. Although further research is warranted regarding generalizability to other imaging modalities and performance improvement, our results mark an encouraging first step for privacy-preserving data sharing in medical imaging.

INFORMS Journal on Optimization, 2022

Robust optimization methods have shown practical advantages in a wide range of decision-making ap... more Robust optimization methods have shown practical advantages in a wide range of decision-making applications under uncertainty. Recently, their efficacy has been extended to multiperiod settings. Current approaches model uncertainty either independent of the past or in an implicit fashion by budgeting the aggregate uncertainty. In many applications, however, past realizations directly influence future uncertainties. For this class of problems, we develop a modeling framework that explicitly incorporates this dependence via connected uncertainty sets, whose parameters at each period depend on previous uncertainty realizations. To find optimal here-and-now solutions, we reformulate robust and distributionally robust constraints for popular set structures and demonstrate this modeling framework numerically on broadly applicable knapsack and portfolio-optimization problems.

Now a days, environmental issues are increasing at alarming rate. We have reviewed manyresearch p... more Now a days, environmental issues are increasing at alarming rate. We have reviewed manyresearch papers on climate change and environment issues..In review paper an overview of manyresearches by many researchers on environment and climate change. This paper reports acomprehensive literature review for 1991–2019 (inclusive), the years in which this topicappeared in scientific journals. In our review paper includes all ecological factors, criticalthinking on environment, space climate, global environment change and many more.

Optimization under Variable Uncertainty Kartikey Sharma In this dissertation, we study models and... more Optimization under Variable Uncertainty Kartikey Sharma In this dissertation, we study models and methods to address uncertainties that can vary in optimization problems. Robust optimization is a popular approach for optimization under uncertainty, especially if limited information is available about the distribution of the uncertainty. It models the uncertainty through sets and finds a robust optimal solution that is feasible for all realizations of the uncertainty within the set and is optimal for the worst-case realization. The structure of these sets determines the complexity of the resulting optimization problem. In most models, the uncertainty set is assumed to be exogenous i.e., predetermined and is unaffected by decisions or other uncertainty realizations in the problem. This thesis introduces endogenous uncertainty models, which may be affected by decisions that are made in the problem or by other uncertainty realizations within the problem. In the first chapter, we take a step towards generalizing robust linear optimization to problems with decision dependent uncertainties. We show these problems to be NPcomplete in general settings. To alleviate these computational inefficiencies, we introduce Dedication Table of Contents List of Tables List of Figures Chapter 1. Introduction Chapter 2. Decision Dependent Uncertainty List of Tables 2.1 Size of Big-M formulation of (RO-DDU) for U i (x) with respect to (i) x ∈ {0, 1} n and (ii) x ∈ R n with x i taking s possible values: dim(y) = p, K constraints in U i (x), and m constraints in the complete problem. 2.2 Comparison of (LC) reformulations for the set U Π (x)

SIAM Journal on Optimization, 2018

The efficacy of robust optimization spans a variety of settings with uncertainties bounded in pre... more The efficacy of robust optimization spans a variety of settings with uncertainties bounded in predetermined sets. In many applications, uncertainties are affected by decisions and cannot be modeled with current frameworks. This paper takes a step towards generalizing robust linear optimization to problems with decisiondependent uncertainties. In general settings, we show these problems to be NP-complete. To alleviate the computational inefficiencies, we introduce a class of uncertainty sets whose size depends on binary decisions. We propose reformulations that improve upon alternative standard linearization techniques. To illustrate the advantages of this framework, a shortest path problem is discussed, where the uncertain arc lengths are affected by decisions. Beyond the modeling and performance advantages, the proposed notion of proactive uncertainty control also mitigates over conservatism of current robust optimization approaches.