Matching Items (462)
Filtering by
- Creators: School of Sustainability
- Member of: Theses and Dissertations
Description
Geology and its tangential studies, collectively known and referred to in this thesis as geosciences, have been paramount to the transformation and advancement of society, fundamentally changing the way we view, interact and live with the surrounding natural and built environment. It is important to recognize the value and importance of this interdisciplinary scientific field while reconciling its ties to imperial and colonizing extractive systems which have led to harmful and invasive endeavors. This intersection among geosciences, (environmental) justice studies, and decolonization is intended to promote inclusive pedagogical models through just and equitable methodologies and frameworks as to prevent further injustices and promote recognition and healing of old wounds. By utilizing decolonial frameworks and highlighting the voices of peoples from colonized and exploited landscapes, this annotated syllabus tackles the issues previously described while proposing solutions involving place-based education and the recentering of land within geoscience pedagogical models. (abstract)
ContributorsReed, Cameron E (Author) / Richter, Jennifer (Thesis director) / Semken, Steven (Committee member) / School of Earth and Space Exploration (Contributor, Contributor) / School of Sustainability (Contributor) / Barrett, The Honors College (Contributor)
Created2021-05
Description
Generating real-world content for VR is challenging in terms of capturing and processing at high resolution and high frame-rates. The content needs to represent a truly immersive experience, where the user can look around in 360-degree view and perceive the depth of the scene. The existing solutions only capture and offload the compute load to the server. But offloading large amounts of raw camera feeds takes longer latencies and poses difficulties for real-time applications. By capturing and computing on the edge, we can closely integrate the systems and optimize for low latency. However, moving the traditional stitching algorithms to battery constrained device needs at least three orders of magnitude reduction in power. We believe that close integration of capture and compute stages will lead to reduced overall system power.
We approach the problem by building a hardware prototype and characterize the end-to-end system bottlenecks of power and performance. The prototype has 6 IMX274 cameras and uses Nvidia Jetson TX2 development board for capture and computation. We found that capturing is bottlenecked by sensor power and data-rates across interfaces, whereas compute is limited by the total number of computations per frame. Our characterization shows that redundant capture and redundant computations lead to high power, huge memory footprint, and high latency. The existing systems lack hardware-software co-design aspects, leading to excessive data transfers across the interfaces and expensive computations within the individual subsystems. Finally, we propose mechanisms to optimize the system for low power and low latency. We emphasize the importance of co-design of different subsystems to reduce and reuse the data. For example, reusing the motion vectors of the ISP stage reduces the memory footprint of the stereo correspondence stage. Our estimates show that pipelining and parallelization on custom FPGA can achieve real time stitching.
We approach the problem by building a hardware prototype and characterize the end-to-end system bottlenecks of power and performance. The prototype has 6 IMX274 cameras and uses Nvidia Jetson TX2 development board for capture and computation. We found that capturing is bottlenecked by sensor power and data-rates across interfaces, whereas compute is limited by the total number of computations per frame. Our characterization shows that redundant capture and redundant computations lead to high power, huge memory footprint, and high latency. The existing systems lack hardware-software co-design aspects, leading to excessive data transfers across the interfaces and expensive computations within the individual subsystems. Finally, we propose mechanisms to optimize the system for low power and low latency. We emphasize the importance of co-design of different subsystems to reduce and reuse the data. For example, reusing the motion vectors of the ISP stage reduces the memory footprint of the stereo correspondence stage. Our estimates show that pipelining and parallelization on custom FPGA can achieve real time stitching.
ContributorsGunnam, Sridhar (Author) / LiKamWa, Robert (Thesis advisor) / Turaga, Pavan (Committee member) / Jayasuriya, Suren (Committee member) / Arizona State University (Publisher)
Created2018
Description
Mixture of experts is a machine learning ensemble approach that consists of individual models that are trained to be ``experts'' on subsets of the data, and a gating network that provides weights to output a combination of the expert predictions. Mixture of experts models do not currently see wide use due to difficulty in training diverse experts and high computational requirements. This work presents modifications of the mixture of experts formulation that use domain knowledge to improve training, and incorporate parameter sharing among experts to reduce computational requirements.
First, this work presents an application of mixture of experts models for quality robust visual recognition. First it is shown that human subjects outperform deep neural networks on classification of distorted images, and then propose a model, MixQualNet, that is more robust to distortions. The proposed model consists of ``experts'' that are trained on a particular type of image distortion. The final output of the model is a weighted sum of the expert models, where the weights are determined by a separate gating network. The proposed model also incorporates weight sharing to reduce the number of parameters, as well as increase performance.
Second, an application of mixture of experts to predict visual saliency is presented. A computational saliency model attempts to predict where humans will look in an image. In the proposed model, each expert network is trained to predict saliency for a set of closely related images. The final saliency map is computed as a weighted mixture of the expert networks' outputs, with weights determined by a separate gating network. The proposed model achieves better performance than several other visual saliency models and a baseline non-mixture model.
Finally, this work introduces a saliency model that is a weighted mixture of models trained for different levels of saliency. Levels of saliency include high saliency, which corresponds to regions where almost all subjects look, and low saliency, which corresponds to regions where some, but not all subjects look. The weighted mixture shows improved performance compared with baseline models because of the diversity of the individual model predictions.
First, this work presents an application of mixture of experts models for quality robust visual recognition. First it is shown that human subjects outperform deep neural networks on classification of distorted images, and then propose a model, MixQualNet, that is more robust to distortions. The proposed model consists of ``experts'' that are trained on a particular type of image distortion. The final output of the model is a weighted sum of the expert models, where the weights are determined by a separate gating network. The proposed model also incorporates weight sharing to reduce the number of parameters, as well as increase performance.
Second, an application of mixture of experts to predict visual saliency is presented. A computational saliency model attempts to predict where humans will look in an image. In the proposed model, each expert network is trained to predict saliency for a set of closely related images. The final saliency map is computed as a weighted mixture of the expert networks' outputs, with weights determined by a separate gating network. The proposed model achieves better performance than several other visual saliency models and a baseline non-mixture model.
Finally, this work introduces a saliency model that is a weighted mixture of models trained for different levels of saliency. Levels of saliency include high saliency, which corresponds to regions where almost all subjects look, and low saliency, which corresponds to regions where some, but not all subjects look. The weighted mixture shows improved performance compared with baseline models because of the diversity of the individual model predictions.
ContributorsDodge, Samuel Fuller (Author) / Karam, Lina (Thesis advisor) / Jayasuriya, Suren (Committee member) / Li, Baoxin (Committee member) / Turaga, Pavan (Committee member) / Arizona State University (Publisher)
Created2018
Description
Non-line-of-sight (NLOS) imaging of objects not visible to either the camera or illumina-
tion source is a challenging task with vital applications including surveillance and robotics.
Recent NLOS reconstruction advances have been achieved using time-resolved measure-
ments. Acquiring these time-resolved measurements requires expensive and specialized
detectors and laser sources. In work proposes a data-driven approach for NLOS 3D local-
ization requiring only a conventional camera and projector. The localisation is performed
using a voxelisation and a regression problem. Accuracy of greater than 90% is achieved
in localizing a NLOS object to a 5cm × 5cm × 5cm volume in real data. By adopting
the regression approach an object of width 10cm to localised to approximately 1.5cm. To
generalize to line-of-sight (LOS) scenes with non-planar surfaces, an adaptive lighting al-
gorithm is adopted. This algorithm, based on radiosity, identifies and illuminates scene
patches in the LOS which most contribute to the NLOS light paths, and can factor in sys-
tem power constraints. Improvements ranging from 6%-15% in accuracy with a non-planar
LOS wall using adaptive lighting is reported, demonstrating the advantage of combining
the physics of light transport with active illumination for data-driven NLOS imaging.
tion source is a challenging task with vital applications including surveillance and robotics.
Recent NLOS reconstruction advances have been achieved using time-resolved measure-
ments. Acquiring these time-resolved measurements requires expensive and specialized
detectors and laser sources. In work proposes a data-driven approach for NLOS 3D local-
ization requiring only a conventional camera and projector. The localisation is performed
using a voxelisation and a regression problem. Accuracy of greater than 90% is achieved
in localizing a NLOS object to a 5cm × 5cm × 5cm volume in real data. By adopting
the regression approach an object of width 10cm to localised to approximately 1.5cm. To
generalize to line-of-sight (LOS) scenes with non-planar surfaces, an adaptive lighting al-
gorithm is adopted. This algorithm, based on radiosity, identifies and illuminates scene
patches in the LOS which most contribute to the NLOS light paths, and can factor in sys-
tem power constraints. Improvements ranging from 6%-15% in accuracy with a non-planar
LOS wall using adaptive lighting is reported, demonstrating the advantage of combining
the physics of light transport with active illumination for data-driven NLOS imaging.
ContributorsChandran, Sreenithy (Author) / Jayasuriya, Suren (Thesis advisor) / Turaga, Pavan (Committee member) / Dasarathy, Gautam (Committee member) / Arizona State University (Publisher)
Created2019
Description
The detection and segmentation of objects appearing in a natural scene, often referred to as Object Detection, has gained a lot of interest in the computer vision field. Although most existing object detectors aim to detect all the objects in a given scene, it is important to evaluate whether these methods are capable of detecting the salient objects in the scene when constraining the number of proposals that can be generated due to constraints on timing or computations during execution. Salient objects are objects that tend to be more fixated by human subjects. The detection of salient objects is important in applications such as image collection browsing, image display on small devices, and perceptual compression.
This thesis proposes a novel evaluation framework that analyses the performance of popular existing object proposal generators in detecting the most salient objects. This work also shows that, by incorporating saliency constraints, the number of generated object proposals and thus the computational cost can be decreased significantly for a target true positive detection rate (TPR).
As part of the proposed framework, salient ground-truth masks are generated from the given original ground-truth masks for a given dataset. Given an object detection dataset, this work constructs salient object location ground-truth data, referred to here as salient ground-truth data for short, that only denotes the locations of salient objects. This is obtained by first computing a saliency map for the input image and then using it to assign a saliency score to each object in the image. Objects whose saliency scores are sufficiently high are referred to as salient objects. The detection rates are analyzed for existing object proposal generators with respect to the original ground-truth masks and the generated salient ground-truth masks.
As part of this work, a salient object detection database with salient ground-truth masks was constructed from the PASCAL VOC 2007 dataset. Not only does this dataset aid in analyzing the performance of existing object detectors for salient object detection, but it also helps in the development of new object detection methods and evaluating their performance in terms of successful detection of salient objects.
This thesis proposes a novel evaluation framework that analyses the performance of popular existing object proposal generators in detecting the most salient objects. This work also shows that, by incorporating saliency constraints, the number of generated object proposals and thus the computational cost can be decreased significantly for a target true positive detection rate (TPR).
As part of the proposed framework, salient ground-truth masks are generated from the given original ground-truth masks for a given dataset. Given an object detection dataset, this work constructs salient object location ground-truth data, referred to here as salient ground-truth data for short, that only denotes the locations of salient objects. This is obtained by first computing a saliency map for the input image and then using it to assign a saliency score to each object in the image. Objects whose saliency scores are sufficiently high are referred to as salient objects. The detection rates are analyzed for existing object proposal generators with respect to the original ground-truth masks and the generated salient ground-truth masks.
As part of this work, a salient object detection database with salient ground-truth masks was constructed from the PASCAL VOC 2007 dataset. Not only does this dataset aid in analyzing the performance of existing object detectors for salient object detection, but it also helps in the development of new object detection methods and evaluating their performance in terms of successful detection of salient objects.
ContributorsKotamraju, Sai Prajwal (Author) / Karam, Lina J (Thesis advisor) / Yu, Hongbin (Committee member) / Jayasuriya, Suren (Committee member) / Arizona State University (Publisher)
Created2019
Description
The purpose of this paper is to understand how companies are finding high potential employees and if they are leaving top talent behind in their approach. Eugene Burke stated in 2014 that 55% of employees that are labeled as a High Potential Employee will turn over and move companies. Burke (2014) also states that the average high potential employee tenure is five years. The Corporate Leadership Council says that on average, 27% of a company's development budget is spent on its high potential program (CEB 2017). For a midsize company, the high potential development budget is almost a million dollars for only a handful of employees, only to see half of the investment walking out the door to another company . Furthermore, the Corporate Leadership Council said that a study done in 2005 revealed that 50% of high potential employees had significant problems within their job (Kotlyar and Karkowsky 2014). Are time and resources are being given to the wrong employees and the right employees are being overlooked? This paper exams how companies traditionally select high potential employees and where companies are potentially omitting employees who would be better suited for the program. This paper proposes that how a company discovers their top talent will correlate to the number of turnovers or struggles that a high potential employee has on their job. Future research direction and practical considerations are also presented in this paper.
ContributorsHarrison, Carrie (Author) / Mizzi, Philip (Thesis director) / Ruediger, Stefan (Committee member) / Department of Management and Entrepreneurship (Contributor) / School of Sustainability (Contributor) / Department of Supply Chain Management (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
Description
This study estimates the capitalization effect of golf courses in Maricopa County using the hedonic pricing method. It draws upon a dataset of 574,989 residential transactions from 2000 to 2006 to examine how the aesthetic, non-golf benefits of golf courses capitalize across a gradient of proximity measures. The measures for amenity value extend beyond home adjacency and include considerations for homes within a range of discrete walkability buffers of golf courses. The models also distinguish between public and private golf courses as a proxy for the level of golf course access perceived by non-golfers. Unobserved spatial characteristics of the neighborhoods around golf courses are controlled for by increasing the extent of spatial fixed effects from city, to census tract, and finally to 2000 meter golf course ‘neighborhoods.’ The estimation results support two primary conclusions. First, golf course proximity is found to be highly valued for adjacent homes and homes up to 50 meters way from a course, still evident but minimal between 50 and 150 meters, and insignificant at all other distance ranges. Second, private golf courses do not command a higher proximity premia compared to public courses with the exception of homes within 25 to 50 meters of a course, indicating that the non-golf benefits of courses capitalize similarly, regardless of course type. The results of this study motivate further investigation into golf course features that signal access or add value to homes in the range of capitalization, particularly for near-adjacent homes between 50 and 150 meters thought previously not to capitalize.
ContributorsJoiner, Emily (Author) / Abbott, Joshua (Thesis director) / Smith, Kerry (Committee member) / Economics Program in CLAS (Contributor) / School of Sustainability (Contributor) / School of Mathematical and Statistical Sciences (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
Description
As climate change and air pollution continue to plague the world today, committed citizens are doing their part to minimize their environmental impact. However, financial limitations have hindered a majority of individuals from adopting clean, renewable energy such as rooftop photovoltaic solar systems. England Sustainability Consulting plans to reverse this limitation and increase affordability for residents across Northern California to install solar panel systems for their energy needs. The purpose of this proposal is to showcase a new approach to procuring solar panel system components while offering the same products needed by each customer. We will examine market data to further prove the feasibility of this business approach while remaining profitable and spread our company's vision across all of Northern California.
ContributorsEngland, Kaysey (Author) / Dooley, Kevin (Thesis director) / Keahey, Jennifer (Committee member) / Department of Supply Chain Management (Contributor) / School of Social and Behavioral Sciences (Contributor) / W.P. Carey School of Business (Contributor) / School of Sustainability (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
Hiking and Hegemony: Destabilizing the nature/culture and gender binaries through outdoor recreation
Description
This paper explores the contested relationships between nature, culture, and gender. In order to analyze these relationships, we look specifically at outdoor recreation. Furthermore, we employ poststructuralist feminist theory in order to produce three frameworks; the first of which is titled Mother Nature’s Promiscuous Past. Rooted in Old World and colonial values, this framework illustrates the flawed feminization of nature by masculinity, and its subsequent extortion of anything related to femininity — including women and nature itself. This belief barred women from nature, resulting in a lack of access for women to outdoor recreation.
Our second framework, titled The Pleasurable Potential of Outdoor Recreation, cites second-wave feminism as a catalyst for women’s participation in wilderness exploration and outdoor recreation. The work of radical feminists and the women’s liberation movement in 1960s and 1970s empowered women at home, in the workplace, and eventually, in the outdoors; women reclaimed their wilderness, yet they continued to employ Framework One’s feminization of nature. Ecofeminsim brought together nature and women, seeking to bring justice to two groups wronged by the same entity: masculinity. In this context, outdoor recreation is empowering for women.
Despite the potential of Framework Two to reinscribe and better the experiences of women in outdoor recreation, we argue that both Frameworks One and Two perpetuate the gender binary and the nature/culture binary, because they are based upon the notion that women are in fact fundamentally different and separate from men, the notion that nature is an entity separate from culture, or human society, as well as the notion that nature is in fact a feminine entity.
Our third framework, Deer Pay No Mind to Your Genitals, engages poststructuralism, asserting that outdoor recreation and activities that occur in nature can serve to destabilize and deconstruct notions of the gender binary. However, we argue that care must be exercised during this process as not to perpetuate the problematic nature/culture binary, a phenomenon that is unproductive in terms of both sustainability and gender liberation. Outdoor recreation has been used by many as a tool to deconstruct numerous societal constraints, including the gender binary; this, however, continues to attribute escapist and isolationist qualities toward nature, and therefore perpetuating the nature/culture divide. Ultimately, we argue outdoor recreation can and should be used as a tool deconstruct the gender binary, however needs to account for the fact that if nature is helping to construct elements of culture, then the two cannot be separate.
Our second framework, titled The Pleasurable Potential of Outdoor Recreation, cites second-wave feminism as a catalyst for women’s participation in wilderness exploration and outdoor recreation. The work of radical feminists and the women’s liberation movement in 1960s and 1970s empowered women at home, in the workplace, and eventually, in the outdoors; women reclaimed their wilderness, yet they continued to employ Framework One’s feminization of nature. Ecofeminsim brought together nature and women, seeking to bring justice to two groups wronged by the same entity: masculinity. In this context, outdoor recreation is empowering for women.
Despite the potential of Framework Two to reinscribe and better the experiences of women in outdoor recreation, we argue that both Frameworks One and Two perpetuate the gender binary and the nature/culture binary, because they are based upon the notion that women are in fact fundamentally different and separate from men, the notion that nature is an entity separate from culture, or human society, as well as the notion that nature is in fact a feminine entity.
Our third framework, Deer Pay No Mind to Your Genitals, engages poststructuralism, asserting that outdoor recreation and activities that occur in nature can serve to destabilize and deconstruct notions of the gender binary. However, we argue that care must be exercised during this process as not to perpetuate the problematic nature/culture binary, a phenomenon that is unproductive in terms of both sustainability and gender liberation. Outdoor recreation has been used by many as a tool to deconstruct numerous societal constraints, including the gender binary; this, however, continues to attribute escapist and isolationist qualities toward nature, and therefore perpetuating the nature/culture divide. Ultimately, we argue outdoor recreation can and should be used as a tool deconstruct the gender binary, however needs to account for the fact that if nature is helping to construct elements of culture, then the two cannot be separate.
ContributorsPolick-Kirkpatrick, Kaelyn (Co-author) / Downing, Haley Marie (Co-author) / Dove-Viebahn, Aviva (Thesis director) / Schoon, Michael (Committee member) / School of Sustainability (Contributor) / School of Social Transformation (Contributor) / Economics Program in CLAS (Contributor) / Barrett, The Honors College (Contributor)
Created2016-05
Description
In this study, the packaging and labeling of milk and coffee was compared between Walmart and Sprouts. The pricing, the sourcing, the certifications and the overall shelf presence of the items was taken under consideration. After studying the packaging of both, a new design incorporating the applicable labels, customer appeal and appropriate green marketing was created for both the commodities.
ContributorsBhatt, Rashi Hitesh (Author) / Collins, Shari (Thesis director) / Keahey, Jennifer (Committee member) / School of International Letters and Cultures (Contributor) / School of Earth and Space Exploration (Contributor) / School of Sustainability (Contributor) / Barrett, The Honors College (Contributor)
Created2016-05