Computer Sciences Commons^™

Bayesian Networks To Assess The Newborn Stool Microbiome, William E. Bennett Jr.

McKelvey School of Engineering Graduate Student Theses & Dissertations

In human stool, a large population of bacterial genes and transcripts from hundreds of genera coexist with host genes and transcripts. Assessments of the metagenome and transcriptome are particularly challenging, since there is a great deal of sequence overlap among related species and related genes. We sequenced the total RNA content from stool samples in a neonate using previously-described methods. We then performed stepwise alignment of different populations of RNA sequence reads to different indices, including ribosomal databases, the human genome, and all sequenced bacterial genomes. Each pool of RNA at each alignment step was subjected to compression to assess …

Mercator (Mapping Enumerator For Cuda) User's Manual, Stephen V. Cole, Jeremy Buhler

All Computer Science and Engineering Research

Welcome to the MERCATOR user's manual! MERCATOR is a CUDA/C++ system designed to assist you in writing efficient CUDA applications by automatically generating significant portions of the GPU-side application code. We hope you find it helpful; please feel free to contact the authors with any questions or feedback.

In-Network Retransmissions In Named Data Networking, Hila Ben Abraham, Patrick Crowley

All Computer Science and Engineering Research

The strategy layer is an important architectural component in both Content-Centric Networking (CCN) and Named Data Networking (NDN). This component introduces a new forwarding model that allows an application to configure its namespace with a forwarding strategy. A core mechanism in every forwarding strategy is the decision of whether to retransmit an unsatisfied Interest or to wait for an application retransmission. While some applications request control of all retransmissions, others rely on the assumption that the strategy will retransmit an Interest when it is not satisfied. Although an application can select the forwarding strategy used in the local host, it …

Deep Semantic Image Interpolation, Joshua D. Little

McKelvey School of Engineering Graduate Student Theses & Dissertations

Image datasets often live on a continuum: Images from an outdoor scene vary from day to night, across different weather conditions, and over the course of seasons. Faces age and exhibit different expressions. We consider the problem of taking individual images from these datasets and explicitly manipulating those images to change where they lie on the continuum. We focus on a version of this problem that requires as little input as possible, and we build off of previous work using CNN features to construct an intermediate image manifold on which to manipulate the images. We also investigate a novel way …

Revelation Of Yin-Yang Balance In Microbial Cell Factories By Data Mining, Flux Modeling, And Metabolic Engineering, Gang Wu

McKelvey School of Engineering Graduate Student Theses & Dissertations

The long-held assumption of never-ending rapid growth in biotechnology and especially in synthetic biology has been recently questioned, due to lack of substantial return of investment. One of the main reasons for failures in synthetic biology and metabolic engineering is the metabolic burdens that result in resource losses. Metabolic burden is defined as the portion of a host cells resources either energy molecules (e.g., NADH, NADPH and ATP) or carbon building blocks (e.g., amino acids) that is used to maintain the engineered components (e.g., pathways). As a result, the effectiveness of synthetic biology tools heavily dependents on cell capability to …

Automatically Characterizing Product And Process Incentives In Collective Intelligence, Allen Brockhurst Lavoie

McKelvey School of Engineering Graduate Student Theses & Dissertations

Social media facilitate interaction and information dissemination among an unprecedented number of participants. Why do users contribute, and why do they contribute to a specific venue? Does the information they receive cover all relevant points of view, or is it biased? The substantial and increasing importance of online communication makes these questions more pressing, but also puts answers within reach of automated methods. I investigate scalable algorithms for understanding two classes of incentives which arise in collective intelligence processes. Product incentives exist when contributors have a stake in the information delivered to other users. I investigate product-relevant user behavior changes, …

Exploiting The Weak Generational Hypothesis For Write Reduction And Object Recycling, Jonathan Andrew Shidal

McKelvey School of Engineering Graduate Student Theses & Dissertations

Programming languages with automatic memory management are continuing to grow in popularity due to ease of programming. However, these languages tend to allocate objects excessively, leading to inefficient use of memory and large garbage collection and allocation overheads.

The weak generational hypothesis notes that objects tend to die young in languages with automatic dynamic memory management. Much work has been done to optimize allocation and garbage collection algorithms based on this observation. Previous work has largely focused on developing efficient software algorithms for allocation and collection. However, much less work has studied architectural solutions. In this work, we propose and …

Visualization Of Deep Convolutional Neural Networks, Dingwen Li

McKelvey School of Engineering Graduate Student Theses & Dissertations

Deep learning has achieved great accuracy in large scale image classification and scene recognition tasks, especially after the Convolutional Neural Network (CNN) model was introduced. Although a CNN often demonstrates very good classification results, it is usually unclear how or why a classification result is achieved. The objective of this thesis is to explore several existing visualization approaches which offer intuitive visual results. The thesis focuses on three visualization approaches: (1) image masking which highlights the region of image with high influence on the classification, (2) Taylor decomposition back-propagation which generates a per pixel heat map that describes each pixel's …

Learning With Scalability And Compactness, Wenlin Chen

McKelvey School of Engineering Graduate Student Theses & Dissertations

Artificial Intelligence has been thriving for decades since its birth. Traditional AI features heuristic search and planning, providing good strategy for tasks that are inherently search-based problems, such as games and GPS searching. In the meantime, machine learning, arguably the hottest subfield of AI, embraces data-driven methodology with great success in a wide range of applications such as computer vision and speech recognition. As a new trend, the applications of both learning and search have shifted toward mobile and embedded devices which entails not only scalability but also compactness of the models. Under this general paradigm, we propose a series …

Grafalgo - A Library Of Graph Algorithms And Supporting Data Structures (Revised), Jonathan Turner

All Computer Science and Engineering Research

This report provides an (updated) overview of Grafalgo, an open-source library of graph algorithms and the data structures used to implement them. The programs in this library were originally written to support a graduate class in advanced data structures and algorithms at Washington University. Because the code's primary purpose was pedagogical, it was written to be as straightforward as possible, while still being highly efficient. Grafalgo is implemented in C++ and incorporates some features of C++11. The library is available on an open-source basis and may be downloaded from https://code.google.com/p/grafalgo/. Source code documentation is at www.arl.wustl.edu/~jst/doc/grafalgo.

Locality-Aware Dynamic Task Graph Scheduling, Jordyn Maglalang, Sriram Krishnamoorthy, Kunal Agrawal

All Computer Science and Engineering Research

Dynamic task graph schedulers automatically balance work across processor cores by scheduling tasks among available threads while preserving dependences. In this paper, we design NabbitC, a provably efficient dynamic task graph scheduler that accounts for data locality on NUMA systems. NabbitC allows users to assign a color to each task representing the location (e.g., a processor core) that has the most efficient access to data needed during that node’s execution. NabbitC then automatically adjusts the scheduling so as to preferentially execute each node at the location that matches its color—leading to better locality because the node is likely to make …

Geometric Algorithms For Modeling Protein Structures, Dan Zeng

Undergraduate Research Symposium Posters

Understanding the structures of macromolecular assemblies is necessary to describe the mechanics of a wide variety of cellular processes. Such assemblies often consist of hundreds of proteins and nucleic acid, each with a unique shape. Modern imaging techniques, including electron cryo-microscopy, create 3D density maps to portray the shapes of these assemblies. However, computational methods for deriving structural models from these images are not fully developed because structural geometry varies depending on the image resolution. To address this issue, we developed geometric algorithms in Gorgon, a molecular visualization software. First, we created an interface for Pathwalking. This algorithm determines protein …

Applying Bayesian Machine Learning Methods To Theoretical Surface Science, Shane Carr

McKelvey School of Engineering Graduate Student Theses & Dissertations

Machine learning is a rapidly evolving field in computer science with increasingly many applications to other domains. In this thesis, I present a Bayesian machine learning approach to solving a problem in theoretical surface science: calculating the preferred active site on a catalyst surface for a given adsorbate molecule. I formulate the problem as a low-dimensional objective function. I show how the objective function can be approximated into a certain confidence interval using just one iteration of the self-consistent field (SCF) loop in density functional theory (DFT). I then use Bayesian optimization to perform a global search for the solution. …

Faster Maximium Priority Matchings In Bipartite Graphs, Jonathan Turner

All Computer Science and Engineering Research

A maximum priority matching is a matching in an undirected graph that maximizes a priority score defined with respect to given vertex priorities. An earlier paper showed how to find maximum priority matchings in unweighted graphs. This paper describes an algorithm for bipartite graphs that is faster when the number of distinct priority classes is limited. For graphs with k distinct priority classes it runs in O(kmn1/2) time, where n is the number of vertices in the graph and m is the number of edges.

The Bounded Edge Coloring Problem And Offline Crossbar Scheduling, Jonathan Turner

All Computer Science and Engineering Research

This paper introduces a variant of the classical edge coloring problem in graphs that can be applied to an offline scheduling problem for crossbar switches. We show that the problem is NP-complete, develop three lower bounds bounds on the optimal solution value and evaluate the performance of several approximation algorithms, both analytically and experimentally. We show how to approximate an optimal solution with a worst-case performance ratio of 3/2 and our experimental results demonstrate that the best algorithms produce results that very closely track a lower bound.

Maximum Priority Matchings, Jonathan Turner

All Computer Science and Engineering Research

Let G=(V,E) be an undirected graph with n vertices and m edges, in which each vertex u is assigned an integer priority in [1,n], with 1 being the ``highest'' priority. Let M be a matching of G. We define the priority score of M to be an n-ary integer in which the i-th most-significant digit is the number of vertices with priority i that are incident to an edge in M. We describe a variation of the augmenting path method (Edmonds' algorithm) that finds a matching with maximum priority score in O(mn) time.

Conflict-Aware Real-Time Routing For Industrial Wireless Sensor-Actuator Networks, Chengjie Wu, Dolvara Gunatilaka, Mo Sha, Chenyang Lu

All Computer Science and Engineering Research

Process industries are adopting wireless sensor-actuator networks (WSANs) as the communication infrastructure. WirelessHART is an open industrial standard for WSANs that have seen world-wide deployments. Real-time scheduling and delay analysis have been studied for WSAN extensively. End-to-end delay in WSANs highly depends on routing, which is still open problem. This paper presents the first real-time routing design for WSAN. We first discuss end-to-end delays of WSANs, then present our real-time routing design. We have implemented and experimented our routing designs on a wireless testbed of 69 nodes. Both experimental results and simulations show that our routing design can improve the …

Maximizing Network Lifetime Of Wireless Sensor-Actuator Networks Under Graph Routing, Chengjie Wu, Dolvara Gunatilaka, Abusayeed Saifullah, Mo Sha, Paras Tiwari, Chenyang Lu, Yixin Chen

All Computer Science and Engineering Research

Process industries are adopting wireless sensor-actuator networks (WSANs) as the communication infrastructure. The dynamics of industrial environments and stringent reliability requirements necessitate high degrees of fault tolerance in routing. WirelessHART is an open industrial standard for WSANs that have seen world-wide deployments. WirelessHART employs graph routing schemes to achieve network reliability through multiple paths. Since many industrial devices operate on batteries in harsh environments where changing batteries are prohibitively labor-intensive, WSANs need to achieve long network lifetime. To meet industrial demand for long-term reliable communication, this paper studies the problem of maximizing network lifetime for WSANs under graph routing. We …

Woodstocc: Extracting Latent Parallelism From A Dna Sequence Aligner On A Gpu, Stephen V. Cole, Jacob R. Gardner, Jeremy D. Buhler

All Computer Science and Engineering Research

An exponential increase in the speed of DNA sequencing over the past decade has driven demand for fast, space-efficient algorithms to process the resultant data. The first step in processing is alignment of many short DNA sequences, or reads, against a large reference sequence. This work presents WOODSTOCC, an implementation of short-read alignment designed for Graphics Processing Unit (GPU) architectures. WOODSTOCC translates a novel CPU implementation of gapped short-read alignment, which has guaranteed optimal and complete results, to the GPU. Our implementation combines an irregular trie search with dynamic programming to expose regularly structured parallelism. We first describe this implementation, …

The Edge Group Coloring Problem With Applications To Multicast Switching, Jonathan Turner

All Computer Science and Engineering Research

This paper introduces a natural generalization of the classical edge coloring problem in graphs that provides a useful abstraction for two well-known problems in multicast switching. We show that the problem is NP-hard and evaluate the performance of several approximation algorithms, both analytically and experimentally. We find that for random χ-colorable graphs, the number of colors used by the best algorithms falls within a small constant factor of χ, where the constant factor is mainly a function of the ratio of the number of outputs to inputs. When this ratio is less than 10, the best algorithms produces solutions that …

Grafalgo - A Library Of Graph Algorithms And Supporting Data Structures, Jonathan Turner

All Computer Science and Engineering Research

This report provides an overview of Grafalgo, an open-source library of graph algorithms and the data structures used to implement them. The programs in this library were originally written to support a graduate class in advanced data structures and algorithms at Washington University. Because the code's primary purpose was pedagogical, it was written to be as straightforward as possible, while still being highly efficient. Grafalgo is implemented in C++ and incorporates some features of C++11. The library is available on an open-source basis and may be downloaded from https://code.google.com/p/grafalgo/. Source code documentation is at www.arl.wustl.edu/~jst/doc/grafalgo. While not designed as production …

The Effects Of Gamifying Optional Lessons On Motivation, Aaron Zemach

McKelvey School of Engineering Graduate Student Theses & Dissertations

Adding video-game elements to non-video-game interfaces (“gamification”) has become a common engagement strategy over the past several years in the domain of education. While prior studies have found that adding game elements to mandatory educational materials can increase students’ motivation to complete the materials, there has yet to be a study to investigate if game elements can make users more likely to engage with optional educational materials. In this study, we investigate whether users of a gamified educational interface are more motivated than users of a non-gamified interface to voluntarily complete educational materials. We found users of a gamified interface …

Data Transport System, Rahav Dor

All Computer Science and Engineering Research

To facilitate the WU Smart Home research [21] we built a system that collects data from sensors and uploads the data to the cloud. The system supports data collection from multiple locations (typically apartments) that are independent from each other, endowing the system with two benefit: distributed data collection and alleviating privacy concerns. Each location is managed by a local micro-server (μServer) that is responsible for receiving data packets from sensors and managing their transient storage. Periodically the μServer triggers a data transport process that moves the data to a cloud server where it is stored in a centralized database. …

Exploring User-Provided Connectivity, Mohammad H. Afrasiabi, Roch Guerin

All Computer Science and Engineering Research

Network services often exhibit positive and negative externalities that affect users' adoption decisions. One such service is "user-provided connectivity" or UPC. The service offers an alternative to traditional infrastructure-based communication services by allowing users to share their "home base" connectivity with other users, thereby increasing their access to connectivity. More users mean more connectivity alternatives, i.e., a positive externality, but also greater odds of having to share one's own connectivity, i.e., a negative externality. The tug of war between positive and negative externalities together with the fact that they often depend not just on how many but also which users …

Migrating To Ipv6 - The Role Of Basic Coordination, Mehdi Nikkhah, Roch Guerin

All Computer Science and Engineering Research

The need for a larger Internet address space was acknowledged early on, and a solution (IPv6) standardized years ago. Its adoption has, however, been anything but easy and still faces significant challenges. The situation begs the questions of "why has it been so difficult?" and "what could have been (or still be) done to facilitate this migration?" There has been significant recent interest in those questions, and the paper builds on a line of work based on technology adoption models to explore them. The results confirm the impact of several known factors, but also provide new insight. In particular, they …

Global Edf Scheduling For Parallel Real-Time Tasks, Jing Li

McKelvey School of Engineering Graduate Student Theses & Dissertations

As multicore processors become ever more prevalent, it is important for real-time programs to take advantage of intra-task parallelism in order to support computation-intensive applications with tight deadlines. In this thesis, we consider the Global Earliest Deadline First (GEDF) scheduling policy for task sets consisting of parallel tasks. Each task can be represented by a directed acyclic graph (DAG) where nodes represent computational work and edges represent dependences between nodes. In this model, we prove that GEDF provides a capacity augmentation bound of 4-2/m and a resource augmentation bound of 2-1/m. The capacity augmentation bound acts as a linear-time schedulability …

Federated Scheduling For Stochastic Parallel Real-Time Tasks, Jing Li, Kunal Agrawal, Christopher Gill, Chenyang Lu

All Computer Science and Engineering Research

Federated scheduling is a strategy to schedule parallel real-time tasks: It allocates a dedicated cluster of cores to high-utilization task (utilization >1); It uses a multiprocessor scheduling algorithm to schedule and execute all low-utilization tasks sequentially, on a shared cluster of the remaining cores. Prior work has shown that federated scheduling has the best known capacity augmentation bound of 2 for parallel tasks with implicit deadlines. In this paper, we explore the soft real-time performance of federated scheduling and address the average-case workloads instead of the worst-case values. In particular, we consider stochastic tasks -- tasks for which execution time …

Inferring Memory Map Instructions, Paul T. Scheid, Ari J. Spilo, Ron K. Cytron

All Computer Science and Engineering Research

We describe the problem of inferring a set of memory map instructions from a reference trace, with the goal of minimizing the number of such instructions as well as the number of unreferenced but mapped storage locations. We prove the related decision problem NP-complete. We then present and compare the results of two heuristic approaches on some actual traces.

Capacity Augmentation Bound Of Federated Scheduling For Parallel Dag Tasks, Jing Li, Abusayeed Saifullah, Kunal Agrawal, Christopher Gill

All Computer Science and Engineering Research

We present a novel federated scheduling approach for parallel real-time tasks under a general directed acyclic graph (DAG) model. We provide a capacity augmentation bound of 2 for hard real-time scheduling; here we use the worst-case execution time and critical-path length of tasks to determine schedulability. This is the best known capacity augmentation bound for parallel tasks. By constructing example task sets, we further show that the lower bound on capacity augmentation of federated scheduling is also 2 for any m > 2. Hence, the gap is closed and bound 2 is a strict bound for federated scheduling. The federated scheduling …

Performance Modeling Of Virtualized Custom Logic Computations, Michael J. Hall, Roger D. Chamberlain

All Computer Science and Engineering Research

Virtualization of custom logic computations (i.e., by sharing a fixed function across distinct data streams) provides a means of reusing hardware resources, particularly when resources are limited. This is common practice in traditional processors where more than one user can share processor resources. In this paper, we virtualize a custom logic block using C-slow techniques to support fine-grain context-switching. We then develop and present an analytic model for several performance measures (throughput, latency, input queue occupancy) for both fine-grained and coarse-grained context switching (to a secondary memory). Next, we calibrate the analytic performance model with empirical measurements. We then validate …