Matching Items (27)
Filtering by
- All Subjects: artificial intelligence
- Creators: Liu, Huan
- Creators: Chakrabarti, Chaitali
Description
The increasing popularity of Twitter renders improved trustworthiness and relevance assessment of tweets much more important for search. However, given the limitations on the size of tweets, it is hard to extract measures for ranking from the tweet's content alone. I propose a method of ranking tweets by generating a reputation score for each tweet that is based not just on content, but also additional information from the Twitter ecosystem that consists of users, tweets, and the web pages that tweets link to. This information is obtained by modeling the Twitter ecosystem as a three-layer graph. The reputation score is used to power two novel methods of ranking tweets by propagating the reputation over an agreement graph based on tweets' content similarity. Additionally, I show how the agreement graph helps counter tweet spam. An evaluation of my method on 16~million tweets from the TREC 2011 Microblog Dataset shows that it doubles the precision over baseline Twitter Search and achieves higher precision than current state of the art method. I present a detailed internal empirical evaluation of RAProp in comparison to several alternative approaches proposed by me, as well as external evaluation in comparison to the current state of the art method.
ContributorsRavikumar, Srijith (Author) / Kambhampati, Subbarao (Thesis advisor) / Davulcu, Hasan (Committee member) / Liu, Huan (Committee member) / Arizona State University (Publisher)
Created2013
Description
As the size and scope of valuable datasets has exploded across many industries and fields of research in recent years, an increasingly diverse audience has sought out effective tools for their large-scale data analytics needs. Over this period, machine learning researchers have also been very prolific in designing improved algorithms which are capable of finding the hidden structure within these datasets. As consumers of popular Big Data frameworks have sought to apply and benefit from these improved learning algorithms, the problems encountered with the frameworks have motivated a new generation of Big Data tools to address the shortcomings of the previous generation. One important example of this is the improved performance in the newer tools with the large class of machine learning algorithms which are highly iterative in nature. In this thesis project, I set about to implement a low-rank matrix completion algorithm (as an example of a highly iterative algorithm) within a popular Big Data framework, and to evaluate its performance processing the Netflix Prize dataset. I begin by describing several approaches which I attempted, but which did not perform adequately. These include an implementation of the Singular Value Thresholding (SVT) algorithm within the Apache Mahout framework, which runs on top of the Apache Hadoop MapReduce engine. I then describe an approach which uses the Divide-Factor-Combine (DFC) algorithmic framework to parallelize the state-of-the-art low-rank completion algorithm Orthogoal Rank-One Matrix Pursuit (OR1MP) within the Apache Spark engine. I describe the results of a series of tests running this implementation with the Netflix dataset on clusters of various sizes, with various degrees of parallelism. For these experiments, I utilized the Amazon Elastic Compute Cloud (EC2) web service. In the final analysis, I conclude that the Spark DFC + OR1MP implementation does indeed produce competitive results, in both accuracy and performance. In particular, the Spark implementation performs nearly as well as the MATLAB implementation of OR1MP without any parallelism, and improves performance to a significant degree as the parallelism increases. In addition, the experience demonstrates how Spark's flexible programming model makes it straightforward to implement this parallel and iterative machine learning algorithm.
ContributorsKrouse, Brian (Author) / Ye, Jieping (Thesis advisor) / Liu, Huan (Committee member) / Davulcu, Hasan (Committee member) / Arizona State University (Publisher)
Created2014
Description
Twitter is a micro-blogging platform where the users can be social, informational or both. In certain cases, users generate tweets that have no "hashtags" or "@mentions"; we call it an orphaned tweet. The user will be more interested to find more "context" of an orphaned tweet presumably to engage with his/her friend on that topic. Finding context for an Orphaned tweet manually is challenging because of larger social graph of a user , the enormous volume of tweets generated per second, topic diversity, and limited information from tweet length of 140 characters. To help the user to get the context of an orphaned tweet, this thesis aims at building a hashtag recommendation system called TweetSense, to suggest hashtags as a context or metadata for the orphaned tweets. This in turn would increase user's social engagement and impact Twitter to maintain its monthly active online users in its social network. In contrast to other existing systems, this hashtag recommendation system recommends personalized hashtags by exploiting the social signals of users in Twitter. The novelty with this system is that it emphasizes on selecting the suitable candidate set of hashtags from the related tweets of user's social graph (timeline).The system then rank them based on the combination of features scores computed from their tweet and user related features. It is evaluated based on its ability to predict suitable hashtags for a random sample of tweets whose existing hashtags are deliberately removed for evaluation. I present a detailed internal empirical evaluation of TweetSense, as well as an external evaluation in comparison with current state of the art method.
ContributorsVijayakumar, Manikandan (Author) / Kambhampati, Subbarao (Thesis advisor) / Liu, Huan (Committee member) / Davulcu, Hasan (Committee member) / Arizona State University (Publisher)
Created2014
Description
As robotic technology and its various uses grow steadily more complex and ubiquitous, humans are coming into increasing contact with robotic agents. A large portion of such contact is cooperative interaction, where both humans and robots are required to work on the same application towards achieving common goals. These application scenarios are characterized by a need to leverage the strengths of each agent as part of a unified team to reach those common goals. To ensure that the robotic agent is truly a contributing team-member, it must exhibit some degree of autonomy in achieving goals that have been delegated to it. Indeed, a significant portion of the utility of such human-robot teams derives from the delegation of goals to the robot, and autonomy on the part of the robot in achieving those goals. In order to be considered truly autonomous, the robot must be able to make its own plans to achieve the goals assigned to it, with only minimal direction and assistance from the human.
Automated planning provides the solution to this problem -- indeed, one of the main motivations that underpinned the beginnings of the field of automated planning was to provide planning support for Shakey the robot with the STRIPS system. For long, however, automated planners suffered from scalability issues that precluded their application to real world, real time robotic systems. Recent decades have seen a gradual abeyance of those issues, and fast planning systems are now the norm rather than the exception. However, some of these advances in speedup and scalability have been achieved by ignoring or abstracting out challenges that real world integrated robotic systems must confront.
In this work, the problem of planning for human-hobot teaming is introduced. The central idea -- the use of automated planning systems as mediators in such human-robot teaming scenarios -- and the main challenges inspired from real world scenarios that must be addressed in order to make such planning seamless are presented: (i) Goals which can be specified or changed at execution time, after the planning process has completed; (ii) Worlds and scenarios where the state changes dynamically while a previous plan is executing; (iii) Models that are incomplete and can be changed during execution; and (iv) Information about the human agent's plan and intentions that can be used for coordination. These challenges are compounded by the fact that the human-robot team must execute in an open world, rife with dynamic events and other agents; and in a manner that encourages the exchange of information between the human and the robot. As an answer to these challenges, implemented solutions and a fielded prototype that combines all of those solutions into one planning system are discussed. Results from running this prototype in real world scenarios are presented, and extensions to some of the solutions are offered as appropriate.
Automated planning provides the solution to this problem -- indeed, one of the main motivations that underpinned the beginnings of the field of automated planning was to provide planning support for Shakey the robot with the STRIPS system. For long, however, automated planners suffered from scalability issues that precluded their application to real world, real time robotic systems. Recent decades have seen a gradual abeyance of those issues, and fast planning systems are now the norm rather than the exception. However, some of these advances in speedup and scalability have been achieved by ignoring or abstracting out challenges that real world integrated robotic systems must confront.
In this work, the problem of planning for human-hobot teaming is introduced. The central idea -- the use of automated planning systems as mediators in such human-robot teaming scenarios -- and the main challenges inspired from real world scenarios that must be addressed in order to make such planning seamless are presented: (i) Goals which can be specified or changed at execution time, after the planning process has completed; (ii) Worlds and scenarios where the state changes dynamically while a previous plan is executing; (iii) Models that are incomplete and can be changed during execution; and (iv) Information about the human agent's plan and intentions that can be used for coordination. These challenges are compounded by the fact that the human-robot team must execute in an open world, rife with dynamic events and other agents; and in a manner that encourages the exchange of information between the human and the robot. As an answer to these challenges, implemented solutions and a fielded prototype that combines all of those solutions into one planning system are discussed. Results from running this prototype in real world scenarios are presented, and extensions to some of the solutions are offered as appropriate.
ContributorsTalamadupula, Kartik (Author) / Kambhampati, Subbarao (Thesis advisor) / Baral, Chitta (Committee member) / Liu, Huan (Committee member) / Scheutz, Matthias (Committee member) / Smith, David E. (Committee member) / Arizona State University (Publisher)
Created2014
Description
Deep neural networks (DNN) have shown tremendous success in various cognitive tasks, such as image classification, speech recognition, etc. However, their usage on resource-constrained edge devices has been limited due to high computation and large memory requirement.
To overcome these challenges, recent works have extensively investigated model compression techniques such as element-wise sparsity, structured sparsity and quantization. While most of these works have applied these compression techniques in isolation, there have been very few studies on application of quantization and structured sparsity together on a DNN model.
This thesis co-optimizes structured sparsity and quantization constraints on DNN models during training. Specifically, it obtains optimal setting of 2-bit weight and 2-bit activation coupled with 4X structured compression by performing combined exploration of quantization and structured compression settings. The optimal DNN model achieves 50X weight memory reduction compared to floating-point uncompressed DNN. This memory saving is significant since applying only structured sparsity constraints achieves 2X memory savings and only quantization constraints achieves 16X memory savings. The algorithm has been validated on both high and low capacity DNNs and on wide-sparse and deep-sparse DNN models. Experiments demonstrated that deep-sparse DNN outperforms shallow-dense DNN with varying level of memory savings depending on DNN precision and sparsity levels. This work further proposed a Pareto-optimal approach to systematically extract optimal DNN models from a huge set of sparse and dense DNN models. The resulting 11 optimal designs were further evaluated by considering overall DNN memory which includes activation memory and weight memory. It was found that there is only a small change in the memory footprint of the optimal designs corresponding to the low sparsity DNNs. However, activation memory cannot be ignored for high sparsity DNNs.
To overcome these challenges, recent works have extensively investigated model compression techniques such as element-wise sparsity, structured sparsity and quantization. While most of these works have applied these compression techniques in isolation, there have been very few studies on application of quantization and structured sparsity together on a DNN model.
This thesis co-optimizes structured sparsity and quantization constraints on DNN models during training. Specifically, it obtains optimal setting of 2-bit weight and 2-bit activation coupled with 4X structured compression by performing combined exploration of quantization and structured compression settings. The optimal DNN model achieves 50X weight memory reduction compared to floating-point uncompressed DNN. This memory saving is significant since applying only structured sparsity constraints achieves 2X memory savings and only quantization constraints achieves 16X memory savings. The algorithm has been validated on both high and low capacity DNNs and on wide-sparse and deep-sparse DNN models. Experiments demonstrated that deep-sparse DNN outperforms shallow-dense DNN with varying level of memory savings depending on DNN precision and sparsity levels. This work further proposed a Pareto-optimal approach to systematically extract optimal DNN models from a huge set of sparse and dense DNN models. The resulting 11 optimal designs were further evaluated by considering overall DNN memory which includes activation memory and weight memory. It was found that there is only a small change in the memory footprint of the optimal designs corresponding to the low sparsity DNNs. However, activation memory cannot be ignored for high sparsity DNNs.
ContributorsSrivastava, Gaurav (Author) / Seo, Jae-Sun (Thesis advisor) / Chakrabarti, Chaitali (Committee member) / Berisha, Visar (Committee member) / Arizona State University (Publisher)
Created2018
Description
The pervasive use of social media gives it a crucial role in helping the public perceive reliable information. Meanwhile, the openness and timeliness of social networking sites also allow for the rapid creation and dissemination of misinformation. It becomes increasingly difficult for online users to find accurate and trustworthy information. As witnessed in recent incidents of misinformation, it escalates quickly and can impact social media users with undesirable consequences and wreak havoc instantaneously. Different from some existing research in psychology and social sciences about misinformation, social media platforms pose unprecedented challenges for misinformation detection. First, intentional spreaders of misinformation will actively disguise themselves. Second, content of misinformation may be manipulated to avoid being detected, while abundant contextual information may play a vital role in detecting it. Third, not only accuracy, earliness of a detection method is also important in containing misinformation from being viral. Fourth, social media platforms have been used as a fundamental data source for various disciplines, and these research may have been conducted in the presence of misinformation. To tackle the challenges, we focus on developing machine learning algorithms that are robust to adversarial manipulation and data scarcity.
The main objective of this dissertation is to provide a systematic study of misinformation detection in social media. To tackle the challenges of adversarial attacks, I propose adaptive detection algorithms to deal with the active manipulations of misinformation spreaders via content and networks. To facilitate content-based approaches, I analyze the contextual data of misinformation and propose to incorporate the specific contextual patterns of misinformation into a principled detection framework. Considering its rapidly growing nature, I study how misinformation can be detected at an early stage. In particular, I focus on the challenge of data scarcity and propose a novel framework to enable historical data to be utilized for emerging incidents that are seemingly irrelevant. With misinformation being viral, applications that rely on social media data face the challenge of corrupted data. To this end, I present robust statistical relational learning and personalization algorithms to minimize the negative effect of misinformation.
The main objective of this dissertation is to provide a systematic study of misinformation detection in social media. To tackle the challenges of adversarial attacks, I propose adaptive detection algorithms to deal with the active manipulations of misinformation spreaders via content and networks. To facilitate content-based approaches, I analyze the contextual data of misinformation and propose to incorporate the specific contextual patterns of misinformation into a principled detection framework. Considering its rapidly growing nature, I study how misinformation can be detected at an early stage. In particular, I focus on the challenge of data scarcity and propose a novel framework to enable historical data to be utilized for emerging incidents that are seemingly irrelevant. With misinformation being viral, applications that rely on social media data face the challenge of corrupted data. To this end, I present robust statistical relational learning and personalization algorithms to minimize the negative effect of misinformation.
ContributorsWu, Liang (Author) / Liu, Huan (Thesis advisor) / Tong, Hanghang (Committee member) / Doupe, Adam (Committee member) / Davison, Brian D. (Committee member) / Arizona State University (Publisher)
Created2019
Description
Deep learning (DL) has proved itself be one of the most important developements till date with far reaching impacts in numerous fields like robotics, computer vision, surveillance, speech processing, machine translation, finance, etc. They are now widely used for countless applications because of their ability to generalize real world data, robustness to noise in previously unseen data and high inference accuracy. With the ability to learn useful features from raw sensor data, deep learning algorithms have out-performed tradinal AI algorithms and pushed the boundaries of what can be achieved with AI. In this work, we demonstrate the power of deep learning by developing a neural network to automatically detect cough instances from audio recorded in un-constrained environments. For this, 24 hours long recordings from 9 dierent patients is collected and carefully labeled by medical personel. A pre-processing algorithm is proposed to convert event based cough dataset to a more informative dataset with start and end of coughs and also introduce data augmentation for regularizing the training procedure. The proposed neural network achieves 92.3% leave-one-out accuracy on data captured in real world.
Deep neural networks are composed of multiple layers that are compute/memory intensive. This makes it difficult to execute these algorithms real-time with low power consumption using existing general purpose computers. In this work, we propose hardware accelerators for a traditional AI algorithm based on random forest trees and two representative deep convolutional neural networks (AlexNet and VGG). With the proposed acceleration techniques, ~ 30x performance improvement was achieved compared to CPU for random forest trees. For deep CNNS, we demonstrate that much higher performance can be achieved with architecture space exploration using any optimization algorithms with system level performance and area models for hardware primitives as inputs and goal of minimizing latency with given resource constraints. With this method, ~30GOPs performance was achieved for Stratix V FPGA boards.
Hardware acceleration of DL algorithms alone is not always the most ecient way and sucient to achieve desired performance. There is a huge headroom available for performance improvement provided the algorithms are designed keeping in mind the hardware limitations and bottlenecks. This work achieves hardware-software co-optimization for Non-Maximal Suppression (NMS) algorithm. Using the proposed algorithmic changes and hardware architecture
With CMOS scaling coming to an end and increasing memory bandwidth bottlenecks, CMOS based system might not scale enough to accommodate requirements of more complicated and deeper neural networks in future. In this work, we explore RRAM crossbars and arrays as compact, high performing and energy efficient alternative to CMOS accelerators for deep learning training and inference. We propose and implement RRAM periphery read and write circuits and achieved ~3000x performance improvement in online dictionary learning compared to CPU.
This work also examines the realistic RRAM devices and their non-idealities. We do an in-depth study of the effects of RRAM non-idealities on inference accuracy when a pretrained model is mapped to RRAM based accelerators. To mitigate this issue, we propose Random Sparse Adaptation (RSA), a novel scheme aimed at tuning the model to take care of the faults of the RRAM array on which it is mapped. Our proposed method can achieve inference accuracy much higher than what traditional Read-Verify-Write (R-V-W) method could achieve. RSA can also recover lost inference accuracy 100x ~ 1000x faster compared to R-V-W. Using 32-bit high precision RSA cells, we achieved ~10% higher accuracy using fautly RRAM arrays compared to what can be achieved by mapping a deep network to an 32 level RRAM array with no variations.
Deep neural networks are composed of multiple layers that are compute/memory intensive. This makes it difficult to execute these algorithms real-time with low power consumption using existing general purpose computers. In this work, we propose hardware accelerators for a traditional AI algorithm based on random forest trees and two representative deep convolutional neural networks (AlexNet and VGG). With the proposed acceleration techniques, ~ 30x performance improvement was achieved compared to CPU for random forest trees. For deep CNNS, we demonstrate that much higher performance can be achieved with architecture space exploration using any optimization algorithms with system level performance and area models for hardware primitives as inputs and goal of minimizing latency with given resource constraints. With this method, ~30GOPs performance was achieved for Stratix V FPGA boards.
Hardware acceleration of DL algorithms alone is not always the most ecient way and sucient to achieve desired performance. There is a huge headroom available for performance improvement provided the algorithms are designed keeping in mind the hardware limitations and bottlenecks. This work achieves hardware-software co-optimization for Non-Maximal Suppression (NMS) algorithm. Using the proposed algorithmic changes and hardware architecture
With CMOS scaling coming to an end and increasing memory bandwidth bottlenecks, CMOS based system might not scale enough to accommodate requirements of more complicated and deeper neural networks in future. In this work, we explore RRAM crossbars and arrays as compact, high performing and energy efficient alternative to CMOS accelerators for deep learning training and inference. We propose and implement RRAM periphery read and write circuits and achieved ~3000x performance improvement in online dictionary learning compared to CPU.
This work also examines the realistic RRAM devices and their non-idealities. We do an in-depth study of the effects of RRAM non-idealities on inference accuracy when a pretrained model is mapped to RRAM based accelerators. To mitigate this issue, we propose Random Sparse Adaptation (RSA), a novel scheme aimed at tuning the model to take care of the faults of the RRAM array on which it is mapped. Our proposed method can achieve inference accuracy much higher than what traditional Read-Verify-Write (R-V-W) method could achieve. RSA can also recover lost inference accuracy 100x ~ 1000x faster compared to R-V-W. Using 32-bit high precision RSA cells, we achieved ~10% higher accuracy using fautly RRAM arrays compared to what can be achieved by mapping a deep network to an 32 level RRAM array with no variations.
ContributorsMohanty, Abinash (Author) / Cao, Yu (Thesis advisor) / Seo, Jae-Sun (Committee member) / Vrudhula, Sarma (Committee member) / Chakrabarti, Chaitali (Committee member) / Arizona State University (Publisher)
Created2018
Description
Speech recognition and keyword detection are becoming increasingly popular applications for mobile systems. While deep neural network (DNN) implementation of these systems have very good performance,
they have large memory and compute resource requirements, making their implementation on a mobile device quite challenging. In this thesis, techniques to reduce the memory and computation cost
of keyword detection and speech recognition networks (or DNNs) are presented.
The first technique is based on representing all weights and biases by a small number of bits and mapping all nodal computations into fixed-point ones with minimal degradation in the
accuracy. Experiments conducted on the Resource Management (RM) database show that for the keyword detection neural network, representing the weights by 5 bits results in a 6 fold reduction in memory compared to a floating point implementation with very little loss in performance. Similarly, for the speech recognition neural network, representing the weights by 6 bits results in a 5 fold reduction in memory while maintaining an error rate similar to a floating point implementation. Additional reduction in memory is achieved by a technique called weight pruning,
where the weights are classified as sensitive and insensitive and the sensitive weights are represented with higher precision. A combination of these two techniques helps reduce the memory
footprint by 81 - 84% for speech recognition and keyword detection networks respectively.
Further reduction in memory size is achieved by judiciously dropping connections for large blocks of weights. The corresponding technique, termed coarse-grain sparsification, introduces
hardware-aware sparsity during DNN training, which leads to efficient weight memory compression and significant reduction in the number of computations during classification without
loss of accuracy. Keyword detection and speech recognition DNNs trained with 75% of the weights dropped and classified with 5-6 bit weight precision effectively reduced the weight memory
requirement by ~95% compared to a fully-connected network with double precision, while showing similar performance in keyword detection accuracy and word error rate.
they have large memory and compute resource requirements, making their implementation on a mobile device quite challenging. In this thesis, techniques to reduce the memory and computation cost
of keyword detection and speech recognition networks (or DNNs) are presented.
The first technique is based on representing all weights and biases by a small number of bits and mapping all nodal computations into fixed-point ones with minimal degradation in the
accuracy. Experiments conducted on the Resource Management (RM) database show that for the keyword detection neural network, representing the weights by 5 bits results in a 6 fold reduction in memory compared to a floating point implementation with very little loss in performance. Similarly, for the speech recognition neural network, representing the weights by 6 bits results in a 5 fold reduction in memory while maintaining an error rate similar to a floating point implementation. Additional reduction in memory is achieved by a technique called weight pruning,
where the weights are classified as sensitive and insensitive and the sensitive weights are represented with higher precision. A combination of these two techniques helps reduce the memory
footprint by 81 - 84% for speech recognition and keyword detection networks respectively.
Further reduction in memory size is achieved by judiciously dropping connections for large blocks of weights. The corresponding technique, termed coarse-grain sparsification, introduces
hardware-aware sparsity during DNN training, which leads to efficient weight memory compression and significant reduction in the number of computations during classification without
loss of accuracy. Keyword detection and speech recognition DNNs trained with 75% of the weights dropped and classified with 5-6 bit weight precision effectively reduced the weight memory
requirement by ~95% compared to a fully-connected network with double precision, while showing similar performance in keyword detection accuracy and word error rate.
ContributorsArunachalam, Sairam (Author) / Chakrabarti, Chaitali (Thesis advisor) / Seo, Jae-Sun (Thesis advisor) / Cao, Yu (Committee member) / Arizona State University (Publisher)
Created2016
Description
Exabytes of data are created online every day. This deluge of data is no more apparent than it is on social media. Naturally, finding ways to leverage this unprecedented source of human information is an active area of research. Social media platforms have become laboratories for conducting experiments about people at scales thought unimaginable only a few years ago.
Researchers and practitioners use social media to extract actionable patterns such as where aid should be distributed in a crisis. However, the validity of these patterns relies on having a representative dataset. As this dissertation shows, the data collected from social media is seldom representative of the activity of the site itself, and less so of human activity. This means that the results of many studies are limited by the quality of data they collect.
The finding that social media data is biased inspires the main challenge addressed by this thesis. I introduce three sets of methodologies to correct for bias. First, I design methods to deal with data collection bias. I offer a methodology which can find bias within a social media dataset. This methodology works by comparing the collected data with other sources to find bias in a stream. The dissertation also outlines a data collection strategy which minimizes the amount of bias that will appear in a given dataset. It introduces a crawling strategy which mitigates the amount of bias in the resulting dataset. Second, I introduce a methodology to identify bots and shills within a social media dataset. This directly addresses the concern that the users of a social media site are not representative. Applying these methodologies allows the population under study on a social media site to better match that of the real world. Finally, the dissertation discusses perceptual biases, explains how they affect analysis, and introduces computational approaches to mitigate them.
The results of the dissertation allow for the discovery and removal of different levels of bias within a social media dataset. This has important implications for social media mining, namely that the behavioral patterns and insights extracted from social media will be more representative of the populations under study.
Researchers and practitioners use social media to extract actionable patterns such as where aid should be distributed in a crisis. However, the validity of these patterns relies on having a representative dataset. As this dissertation shows, the data collected from social media is seldom representative of the activity of the site itself, and less so of human activity. This means that the results of many studies are limited by the quality of data they collect.
The finding that social media data is biased inspires the main challenge addressed by this thesis. I introduce three sets of methodologies to correct for bias. First, I design methods to deal with data collection bias. I offer a methodology which can find bias within a social media dataset. This methodology works by comparing the collected data with other sources to find bias in a stream. The dissertation also outlines a data collection strategy which minimizes the amount of bias that will appear in a given dataset. It introduces a crawling strategy which mitigates the amount of bias in the resulting dataset. Second, I introduce a methodology to identify bots and shills within a social media dataset. This directly addresses the concern that the users of a social media site are not representative. Applying these methodologies allows the population under study on a social media site to better match that of the real world. Finally, the dissertation discusses perceptual biases, explains how they affect analysis, and introduces computational approaches to mitigate them.
The results of the dissertation allow for the discovery and removal of different levels of bias within a social media dataset. This has important implications for social media mining, namely that the behavioral patterns and insights extracted from social media will be more representative of the populations under study.
ContributorsMorstatter, Fred (Author) / Liu, Huan (Thesis advisor) / Kambhampati, Subbarao (Committee member) / Maciejewski, Ross (Committee member) / Carley, Kathleen M. (Committee member) / Arizona State University (Publisher)
Created2017
Description
Using stereo vision for 3D reconstruction and depth estimation has become a popular and promising research area as it has a simple setup with passive cameras and relatively efficient processing procedure. The work in this dissertation focuses on locally adaptive stereo vision methods and applications to different imaging setups and image scenes.
Solder ball height and substrate coplanarity inspection is essential to the detection of potential connectivity issues in semi-conductor units. Current ball height and substrate coplanarity inspection tools are expensive and slow, which makes them difficult to use in a real-time manufacturing setting. In this dissertation, an automatic, stereo vision based, in-line ball height and coplanarity inspection method is presented. The proposed method includes an imaging setup together with a computer vision algorithm for reliable, in-line ball height measurement. The imaging setup and calibration, ball height estimation and substrate coplanarity calculation are presented with novel stereo vision methods. The results of the proposed method are evaluated in a measurement capability analysis (MCA) procedure and compared with the ground-truth obtained by an existing laser scanning tool and an existing confocal inspection tool. The proposed system outperforms existing inspection tools in terms of accuracy and stability.
In a rectified stereo vision system, stereo matching methods can be categorized into global methods and local methods. Local stereo methods are more suitable for real-time processing purposes with competitive accuracy as compared with global methods. This work proposes a stereo matching method based on sparse locally adaptive cost aggregation. In order to reduce outlier disparity values that correspond to mis-matches, a novel sparse disparity subset selection method is proposed by assigning a significance status to candidate disparity values, and selecting the significant disparity values adaptively. An adaptive guided filtering method using the disparity subset for refined cost aggregation and disparity calculation is demonstrated. The proposed stereo matching algorithm is tested on the Middlebury and the KITTI stereo evaluation benchmark images. A performance analysis of the proposed method in terms of the I0 norm of the disparity subset is presented to demonstrate the achieved efficiency and accuracy.
Solder ball height and substrate coplanarity inspection is essential to the detection of potential connectivity issues in semi-conductor units. Current ball height and substrate coplanarity inspection tools are expensive and slow, which makes them difficult to use in a real-time manufacturing setting. In this dissertation, an automatic, stereo vision based, in-line ball height and coplanarity inspection method is presented. The proposed method includes an imaging setup together with a computer vision algorithm for reliable, in-line ball height measurement. The imaging setup and calibration, ball height estimation and substrate coplanarity calculation are presented with novel stereo vision methods. The results of the proposed method are evaluated in a measurement capability analysis (MCA) procedure and compared with the ground-truth obtained by an existing laser scanning tool and an existing confocal inspection tool. The proposed system outperforms existing inspection tools in terms of accuracy and stability.
In a rectified stereo vision system, stereo matching methods can be categorized into global methods and local methods. Local stereo methods are more suitable for real-time processing purposes with competitive accuracy as compared with global methods. This work proposes a stereo matching method based on sparse locally adaptive cost aggregation. In order to reduce outlier disparity values that correspond to mis-matches, a novel sparse disparity subset selection method is proposed by assigning a significance status to candidate disparity values, and selecting the significant disparity values adaptively. An adaptive guided filtering method using the disparity subset for refined cost aggregation and disparity calculation is demonstrated. The proposed stereo matching algorithm is tested on the Middlebury and the KITTI stereo evaluation benchmark images. A performance analysis of the proposed method in terms of the I0 norm of the disparity subset is presented to demonstrate the achieved efficiency and accuracy.
ContributorsLi, Jinjin (Author) / Karam, Lina (Thesis advisor) / Chakrabarti, Chaitali (Committee member) / Patel, Nital (Committee member) / Spanias, Andreas (Committee member) / Arizona State University (Publisher)
Created2017