Theses and Dissertations

Permanent URI for this collection

https://hdl.handle.net/10217/100389

Browse

Now showing 1 - 20 of 300

Open Access
Embodied multimodal referring expressions generation
(Colorado State University. Libraries, 2024) Alalyani, Nada H., author; Krishnaswamy, Nikhil, advisor; Ortega, Francisco, committee member; Blanchard, Nathaniel, committee member; Wang, Haonan, committee member
Using both verbal and non-verbal modalities in generating definite descriptions of objects and locations is a critical human capability in collaborative interactions. Despite advancements in AI, embodied interactive virtual agents (IVAs) are not equipped to intelligently mix modalities to communicate their intents as humans do, which hamstrings naturalistic multimodal IVA. We introduce SCMRE, a situated corpus of multimodal referring expressions (MREs) intended for training generative AI systems in multimodal IVA, focusing on multimodal referring expressions. Our contributions include: 1) Developing an IVA platform that interprets human multimodal instructions and responds with language and gestures; 2) Providing 24 participants with 10 scenes, each involving ten equally-sized blocks randomly placed on a table. These interactions generated a dataset of 10,408 samples; 3) Analyzing SCMRE, revealing that the utilization of pointing significantly reduces the ambiguity of prompts and increases the efficiency of IVA's execution of humans' prompts; 4) Augmenting and synthesizing SCMRE, resulting in 22,159 samples to generate more data for model training; 5) Finetuning LLaMA 2-chat-13B for generating contextually-correct and situationally-fluent multimodal referring expressions; 6) Integrating the fine-tuned model into the IVA to evaluate the success of the generative model-enabled IVA in communication with humans; 7) Establishing the evaluation process which applies to both humans and IVAs and combines quantitative and qualitative metrics.
Embargo
Interaction and navigation in cross-reality analytics
(Colorado State University. Libraries, 2024) Zhou, Xiaoyan, author; Ortega, Francisco, advisor; Ray, Indrakshi, committee member; Moraes, Marcia, committee member; Batmaz, Anil Ufuk, committee member; Malinin, Laura, committee member
Along with immersive display technology's fast evolution, augmented reality (AR) and virtual reality (VR) are increasingly being researched to facilitate data analytics, known as Immersive Analytics. The ability to interact with data visualization in the space around users not only builds the foundation of ubiquitous analytics but also assists users in the sensemaking of the data. However, interaction and navigation while making sense of 3D data visualization in different realities still need to be better understood and explored. For example, what are the differences between users interacting in augmented and virtual reality, and how can we utilize them in the best way during analysis tasks? Moreover, based on the existing work and our preliminary studies, improving the interaction efficiency with immersive displays still needs to be solved. Therefore, this thesis focuses on understanding interaction and navigation in augmented reality and virtual reality for immersive analytics. First, we explored how users interact with multiple objects in augmented reality by using the "Wizard of Oz" study approach. We elicited multimodal interactions involving hand gestures and speech, with text prompts shown on the head-mounted display. Then, we compared the results with previous work in a single-object scenario, which helped us better understand how users prefer to interact in a more complex AR environment. Second, we built an immersive analytics platform in both AR and VR environments to simulate a realistic scenario and conducted a controlled study to evaluate user performance with designed analysis tools and 3D data visualization. Based on the results, interaction and navigation patterns were observed and analyzed for a better understanding of user preferences during the sensemaking process. ii Lastly, by considering the findings and insights from prior studies, we developed a hybrid user interface in simulated cross-reality for situated analytics. An exploratory study was conducted with a smart home setting to understand user interaction and navigation in a more familiar scenario with practical tasks. With the results, we did a thorough qualitative analysis of feedback and video recording to disclose user preferences with interaction and visualization in situated analytics in the everyday decision-making scenario. In conclusion, this thesis uncovered user-designed multimodal interaction including mid-air hand gestures and speech for AR, users' interaction and navigation strategies in immersive analytics in both AR and VR, and hybrid user interface usage in situated analytics for assisting decision-making. Our findings and insights in this thesis provide guidelines and inspiration for future research in interaction and navigation design and improving user experience with analytics in mixed-reality environments.
Embargo
Towards automated security and privacy policies specification and analysis
(Colorado State University. Libraries, 2024) Alqurashi, Saja Salem, author; Ray, Indrakshi, advisor; Ray, Indrajit, committee member; Malaiya, Yashwant, committee member; Simske, Steve, committee member
Security and privacy policies, vital for information systems, are typically expressed in natural language documents. Security policy is represented by Access Control Policies (ACPs) within security requirements, initially drafted in natural language and subsequently translated into enforce- able policy. The unstructured and ambiguous nature of the natural language documents makes the manual translation process tedious, expensive, labor-intensive, and prone to errors. On the other hand, Privacy policy, with its length and complexity, presents unique challenges. The dense language and extensive content of the privacy policies can be overwhelming, hindering both novice users and experts from fully understanding the practices related to data collection and sharing. The disclosure of these data practices to users, as mandated by privacy regulations such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA), is of utmost importance. To address these challenges, we have turned to Natural Language Processing (NLP) to automate extracting critical information from natural language documents and analyze those security and privacy policies. Thus, this dissertation aims to address two primary research questions: Question 1: How can we automate the translation of Access Control Policies (ACPs) from natural language expressions to the formal model of Next Generation Access Control (NGAC) and subsequently analyze the generated model? Question 2: How can we automate the extraction and analysis of data practices from privacy policies to ensure alignment with privacy regulations (GDPR and CCPA)? Addressing these research questions necessitates the development of a comprehensive framework comprising two key components. The first component, SR2ACM, focuses on translating natural language ACPs into the NGAC model. This component introduces a series of innovative contributions to the analysis of security policies. At the core of our contributions is an automated approach to constructing ACPs within the NGAC specification directly from natural language documents. Our approach integrates machine learning with software testing, a novel methodology to ensure the quality of the extracted access control model. The second component, Privacy2Practice, is designed to automate the extraction and analysis of the data practices from privacy policies written in natural language. We have developed an automated method to extract data practices mandated by privacy regulations and to analyze the disclosure of these data practices within the privacy policies. The novelty of this research lies in creating a comprehensive framework that identifies the critical elements within security and privacy policies. Thus, this innovative framework enables automated extraction and analysis of both types of policies directly from natural language documents.
Open Access
SMOKE+: a video dataset for automated fine-grained assessment of smoke opacity
(Colorado State University. Libraries, 2024) Seefried, Ethan, author; Blanchard, Nathaniel, advisor; Sreedharan, Sarath, committee member; Roberts, Jacob, committee member
Computer vision has traditionally faced difficulties when applied to amorphous objects like smoke, owing to their ever-changing shape, texture, and dependence on background conditions. While recent advancements have enabled simple tasks such as smoke detection and basic classification (black or white), quantitative opacity estimation in line with the assessments made by certified professionals remains unexplored. To address this gap, I introduce the SMOKE+ dataset, which features opacity labels verified by three certified experts. My dataset encompasses five distinct testing days, two data collection sites in different regions, and a total of 13,632 labeled clips. Leveraging this data, we develop a state-of-the-art smoke opacity estimation method that employs a small number of Residual 3D blocks for efficient opacity estimation. Additionally I explore the use of MAMBA blocks in a video based architecture, exploiting their ability to handle spatial and temporal data in a linear fashion. Techniques developed during the SMOKE+ dataset creation were then refined and applied to a new dataset titled CSU101, designed for educational use in Computer Vision. In the future I intend to expand further into synthetic data, incorporating techniques into Unreal Engine or Unity to add accurate opacity labels.
Open Access
Toward robust embedded networks in heavy vehicles - machine learning strategies for fault tolerance
(Colorado State University. Libraries, 2024) Ghatak, Chandrima, author; Ray, Indrakshi, advisor; Malaiya, Yashwant, committee member; Kokoszka, Piotr, committee member
In the domain of critical infrastructure, Medium and Heavy Duty (MHD) vehicles play an integral role in both military and civilian operations. These vehicles are essential for the efficiency and reliability of modern logistics. The operations of modern MHD vehicles are heavily automated through embedded computers called Electronic Control Units (ECUs). These ECUs utilize arrays of sensors to control and optimize various vehicle functions and are critical to maintaining operational effectiveness. In most MHD vehicles, this sensor data is predominantly communicated using the Society of Automotive Engineering's (SAE) J1939 Protocol over Controller Area Networks (CAN) and is vital for the smooth functioning of MHD vehicles. The resilience of these communication networks is especially crucial in adversarial environments where sensor systems are susceptible to disruptions caused by physical (kinetic) or cyber-attacks. This dissertation presents an innovative approach using predictive machine learning algorithms to forecast accurate sensor readings in scenarios where sensor systems become compromised. The study focuses on the SAE J1939 networks in MHD vehicles, utilizing real-world data from a Class 6 Kenworth T270 truck. Three distinct machine-learning methods are explored and evaluated for their effectiveness in predicting missing sensor data. The results demonstrate that these models can nearly accurately predict sensor data, which is essential in preventing the vehicle from entering engine protection or limp modes, thereby extending operational capacity under adverse conditions. Overall, this research highlights the potential of machine learning in enhancing the resilience of networked cyber-physical systems, particularly in MHD vehicles. It underscores the significance of predictive algorithms in maintaining operational feasibility and contributes to the broader discussion on the resilience of critical infrastructure in hostile settings.
Open Access
Robust gesture detection for multimodal problem solving
(Colorado State University. Libraries, 2024) VanderHoeven, Hannah G., author; Blanchard, Nathaniel, advisor; Krishnaswamy, Nikhil, advisor; Cleary, Anne M., committee member
Throughout various collaborative problem solving (CPS) tasks, multiple different communicative modalities may be used by participants as they communicate with each other to work towards some goal. The ability to recognize and act on these modalities is vital for a multimodal AI agent to effectively interact with humans in a meaningful way. Potential modalities of interest might include, speech, gesture, action, pose, facial expression, and object positions in three dimensional space. As AI becomes move commonplace in various collaborative environments, there is a lot of potential to use an agent to help support learning, training and understanding of how small groups work together to complete CPS tasks. Designing a well rounded system to best understand small group interactions, multiple different modalities need to be supported. Gesture is one of many important features to consider in multimodal design. Robust gesture recognition is a key component of multimodal language understanding in addition to human-computer interaction. Most vision based approaches for gesture recognition focus on static standalone gestures that are identifiable in a single video frame. In CPS tasks, more complex gestures made up of multiple "phases" are more likely to exist. For instance deixis, or pointing, as it is used to indicate objects and referents in a scene. In this thesis, I present a novel method for robust gesture detection based on gesture phase semantics. This method is competitive with many state of the art computer vision approaches while being faster to train on annotated data. I also present various applications of this method to utilize pointing detection in a real-world collaborative task, and I discuss the importance of robust gesture detection as an important feature in multimodal agent design in further depth.
Open Access
Streamlining decentralized ledger: enhancing historical data access in Hyperledger Fabric
(Colorado State University. Libraries, 2024) Bachinin, Andrei, author; Ray, Indrakshi, advisor; Malaiya, Yashwant K., committee member; Ray, Indrajit, committee member; Simske, Steven J., committee member
This thesis presents design and implementation of Indexing Solution (IS) that aims to enhance query retrieval process in Hyperledger Fabric (HLF). In order to address limitations of HLF, we introduce new indexing algorithms: the Version-Based Index (VBI) and the Block-Based Index (BBI). We also introduce several previously unsupported query APIs: GetHistoryForVersionRange (GHFVR), GetHistoryForBlockRange (GHVBR), GetHistoryForKeyRange (GHFKR). All the work we propose is designed to integrate and work seamlessly with HLF, ensuring full backward compatibility with existing architecture. Our experiments demonstrates that proposed solution significantly outperforms original HLF regarding query execution time, reducing it from 6.621 seconds to 0.019 seconds for certain queries. While VBI and BBI introduce a space index overhead, it remains comparable to the space overhead shown in HLF. Furthermore, we adopt parallel data insertion that mitigates a slower data insertion observed in both VBI and BBI. Comparison with traditional database systems (LevelDB and MySQL) and other blockchain solution from literature (Lineage Chain and vChain+) highlights significant advantage of our IS in terms of querying capabilities, paving the way for broader application of HLF.
Embargo
The shape of sound: rendering interactive six-degrees-of-freedom audio in software
(Colorado State University. Libraries, 2024) Rehberg, Daniel, author; Ortega, Francisco Raul, advisor; Rajopadhye, Sanjay, committee member; Malinin, Laura, committee member
Six-degrees-of-freedom (6DoF) audio is an area of growing interest in interactive software, but it has faced several challenges: it does not easily conform to object-based rendering when achieved with arrays of ambisonics microphones; prior studies rely on subjective metrics which do not clearly indicate how this additional audio interaction might aid a human in a localization task (an indication of enhanced spatial awareness of a sound event); and the ambisonics technique requires specialized equipment and recording space, as well as audio engineering expertise for setup and calibration to work properly. These factors limit the accessibility of 6DoF audio to be implemented in research experiments or within commercial products like videogames. My work has involved taking an interdisciplinary approach to design, prototype, and validate (with human subjects) an inherently object-based 6DoF rendering method. This method exploits computational geometry techniques and follows a rendering paradigm inspired by the programmable graphics pipeline to create 3D audio meshes which can be transformed in real time to dynamically render monaural audio samples – meaning the output of the method can still be input into contemporary audio filtering and spatialization functions/tools, like a head-related transfer function. This work includes two studies performed with human subjects as well as a breakdown of the rendering method and its prototype implementation. The results of the human-subject studies indicate clear advantages to localizing a spatial sound in 3D space compared to the contemporary three-degrees-of-freedom approach.
Open Access
Using eye gaze to automatically identify familiarity
(Colorado State University. Libraries, 2024) Castillon, Iliana, author; Blanchard, Nathaniel, advisor; Sreedharan, Sarath, committee member; Cleary, Anne M., committee member
Understanding internal cognitive states, such as the sensation of familiarity, is crucial not only in the realm of human perception but also in enhancing interactions with artificial intelligence. One such state is the experience of familiarity, a fundamental aspect of human perception that often manifests as an intuitive recognition of faces or places. Automatically identifying cognitive experiences could pave the way for more nuance in human-AI interaction. While other works have shown the feasibility of automatically identifying other internal cognitive states like mind wandering using eye gaze features, the automatic detection of familiarity remains largely unexplored. In this work, we employed a paradigm from cognitive psychology to induce feelings of familiarity. Then, we trained machine learning models to automatically detect familiarity using eye gaze measurements, both in experiments with traditional computer use (e.g., eye tracker attached to monitor) and in virtual reality settings, in a participant independent manner. Familiarity was detected with a Cohen's kappa value, a measurement of accuracy corrected for random guessing, of 0.22 and 0.21, respectively. This work showcases the feasibility of automatically identifying feelings of familiarity and opens the door to exploring automated familiarity detection in other contexts, such as students engaged with a learning task while interacting with an intelligent tutoring system.
Open Access
Transformer, diffusion, and GAN-based augmentations for contrastive learning of visual representations
(Colorado State University. Libraries, 2024) Armstrong, Samuel, author; Pallickara, Sangmi, advisor; Pallickara, Shrideep, advisor; Ghosh, Sudipto, committee member; Breidt, F. Jay, committee member
Generative modeling and self-supervised learning have emerged as two of the most prominent fields of study in machine learning in recent years. Generative models are able to learn detailed visual representations that can then be used to generate synthetic data. Modern self-supervised learning methods are able to extract high-level visual information from images in an unsupervised manner and then apply this information to downstream tasks such as object detection and segmentation. As generative models become more and more advanced, we want to be able to extract their learned knowledge and then apply it to downstream tasks. In this work, we develop Generative Contrastive Learning (GCL), a methodology that uses contrastive learning to extract information from modern generative models. We define GCL's high-level components: an encoder, feature map augmenter, decoder, handcrafted augmenter, and contrastive learning model and demonstrate how to apply GCL to the three major types of large generative models: GANs, Diffusion Models, and Image Transformers. Due to the complex nature of generative models and the near-infinite number of unique images they can produce, we have developed several methodologies to synthesize images in a manner that compliments the augmentation-based learning that is used in contrastive learning frameworks. Our work shows that applying these large generative models to self-supervised learning can be done in a computationally viable manner without the use of large clusters of high-performance GPUs. Finally, we show the clear benefit of leveraging generative models in a contrastive learning setting using standard self-supervised learning benchmarks.
Open Access
Automating the derivation of memory allocations for acceleration of polyhedral programs
(Colorado State University. Libraries, 2024) Ferry, Corentin, author; Rajopadhye, Sanjay, advisor; Derrien, Steven, advisor; Wilson, Jesse, committee member; Pasricha, Sudeep, committee member; McClurg, Jedidiah, committee member; Sadayappan, Ponnuswamy, committee member; de Dinechin, Florent, committee member; Collange, Caroline, committee member
As processors compute power keeps increasing, so do their demands in memory accesses: some computations will require a higher bandwidth and exhibit regular memory access patterns, others will require a lower access latency and exhibit random access patterns. To cope with all demands, memory technologies are becoming diverse. It is then necessary to adapt both programs and hardware accelerators to the memory technology they use. Notably, memory access patterns and memory layouts have to be optimized. Manual optimization can be extremely tedious and does not scale to a large number of processors and memories, where automation becomes necessary. In this Ph.D dissertation, we suggest several automated methods to derive data layouts from programs, notably for FPGA accelerators. We focus on getting the best throughput from high-latency, high-bandwidth memories and, for all kinds of memories, the lowest redundancy while preserving contiguity. To this effect, we introduce mathematical analyses to partition the data flow of a program with uniform and affine dependence patterns, propose memory layouts and automation techniques to get optimized FPGA accelerators.
Open Access
Deep learning for bioinformatics sequences: RNA basecalling and protein interactions
(Colorado State University. Libraries, 2024) Neumann, Don, author; Ben-Hur, Asa, advisor; Beveridge, Ross, committee member; Blanchard, Nathaniel, committee member; Reddy, Anireddy, committee member
In the interdisciplinary field of bioinformatics, sequence data for biological problems comes in many different forms. This ranges from proteins, to RNA, to the ionic current for a strand of nucleotides from an Oxford Nanopore Technologies sequencing device. This data can be used to elucidate the fundamentals of biological processes on many levels, which can help humanity with everything from drug design to curing disease. All of our research focuses on biological problems encoded as sequences. The main focus of our research involves Oxford Nanopore Technology sequencing devices which are capable of directly sequencing long read RNA strands as is. We first concentrate on improving the basecalling accuracy for RNA, and have published a paper with a novel architecture achieving state-of-the-art performance. The basecalling architecture uses convolutional blocks, each with progressively larger kernel sizes which improves accuracy for the noisy nature of the data. We then describe ongoing research into the detection of post-transcriptional RNA modifications in nanopore sequencing data. Building on our basecalling research, we are able to discern modifications with read level resolution. Our work will facilitate research into the detection of N6-methyladeosine (m6A) while also furthering progress in the detection of other post-transcriptional modifications. Finally, we recount our recently accepted paper regarding protein-protein and host-pathogen interaction prediction. We performed experiments demonstrating faulty experimental design for interaction prediction which have plagued the field, giving the faulty impression the problem has been solved. We then provide reasoning and recommendations for future work.
Open Access
Towards fair and efficient distributed intelligence
(Colorado State University. Libraries, 2024) Gorbett, Matt, author; Ray, Indrakshi, advisor; Shirazi, Hossein, committee member; Simske, Steve, committee member; Jayasumana, Anura, committee member
Artificial Intelligence is rapidly advancing the modern technological landscape. Alongside this progress, the ubiquitous presence of computational devices has created unique opportunities to deploy intelligent systems in novel environments. For instance, resource constrained machines such as IoT devices have the potential to enhance our world through the use of Deep Neural Networks (DNNs). However, modern DNNs suffer from high computational complexity and are often relegated to specialized hardware, a bottleneck which has severely limited their practical use. In this work, we contribute to improving these issues through the use of neural network compression. We present new findings for both model quantization and pruning, two standard techniques for creating compressed and efficient DNNs. To begin, we examine the efficacy of neural network compression for time series learning, an unstudied modality in model compression literature. We construct a generalized Transformer architecture for multivariate time series which applies both binarization and pruning to model parameters. Our results show that the lightweight models achieve comparable accuracy to dense Transformers of the same structure on time series forecasting, classification, and anomaly detection tasks while significantly reducing the computational burden. Next, we propose two novel algorithms for neural network compression: 1) Tiled Bit Networks (TBNs) and 2) Iterative Weight Recycling (IWR). TBNs present a new form of quantization to tile neural network layers with sequences of bits to achieve sub-bit compression of binary-weighted models. The method learns binary vectors (i.e. tiles) to populate each layer of a model via tensor aggregation and reshaping operations; during inference, TBNs use just a single tile per model layer. TBNs perform well across a diverse range of architecture (CNNs, MLPs, Transformers) and tasks (classification, segmentation) while achieving up to 8x reduction in size compared to binary-weighted models. The second algorithm, IWR, generates sparse neural networks from randomly initialized models by identifying important parameters within neural networks for reuse. The approach enables us to prune 80% of ResNet50's parameters while still achieving 70.8% accuracy on ImageNet. Finally, we examine the feasibility of deploying compressed DNNs in practical applications. Specifically, we deploy Sparse Binary Neural Networks (SBNNs), TBNs, and other common compression algorithms on an embedded device for performance assessment, finding a reduction in both peak memory and storage size. By integrating algorithmic and theoretical advancements into a comprehensive end-to-end methodology, this dissertation contributes a new framework for crafting powerful and efficient deep learning models applicable in real-world settings.
Open Access
Throughput optimization techniques for heterogeneous architectures
(Colorado State University. Libraries, 2024) Derumigny, Nicolas, author; Pouchet, Louis-Noël, advisor; Rastello, Fabrice, advisor; Hack, Sebastian, committee member; Rohou, Erven, committee member; Malaiya, Yashwant, committee member; Ortega, Francisco, committee member; Pétrot, Frédéric, committee member; Wilson, James, committee member; Zaks, Ayal, committee member
Moore's Law has allowed during the past 40 years to exponentially increase transistor density of integrated circuits. As a result, computing devices ranging from general-purpose processors to dedicated accelerators have become more and more complex due to the specialization and the multiplication of their compute units. Therefore, both low-level program optimization (e.g. assembly-level programming and generation) and accelerator design must solve the issue of efficiently mapping the input program computations to the various chip capabilities. However, real-world chip blueprints are not openly accessible in practice, and their documentation is often incomplete. Given the diversity of CPUs available (Intel's / AMD's / Arm's microarchitectures), we tackle in this manuscript the problem of automatically inferring a performance model applicable to fine-grain throughput optimization of regular programs. Furthermore, when order of magnitude of performance gain over generic accelerators are needed, domain-specific accelerators must be considered; which raises the same question of the number of dedicated units as well as their functionality. To remedy this issue, we present two complementary approaches: on one hand, the study of single-application specialized accelerators with an emphasis on hardware reuse, and, on the other hand, the generation of semi-specialized designs suited for a user-defined set of applications.
Open Access
Typed synthesis of fast multiplication algorithms for post-quantum cryptography
(Colorado State University. Libraries, 2024) Scarbro, William, author; Rajopadhye, Sanjay, advisor; McClurg, Jedidiah, committee member; Achter, Jeffrey, committee member
Multiplication over polynomial rings is a time consuming operation in many post-quantum cryptosystems. State-of-the-art implementations of multiplication for these cryptosystems have been developed by hand using an algebraic framework. A similar class of algorithms, based on the Discrete Fourier Transform, have been optimized across a variety of platforms using program synthesis. We demonstrate how the algebraic framework used to describe fast multiplication algorithms can be used in program synthesis. Specifically, we extend and then abstract this framework for use in program synthesis, allowing AI search techniques to find novel, high performance implementations of polynomial ring multiplication across platforms.
Open Access
Pruning visual transformers to increase model compression and decrease inference time
(Colorado State University. Libraries, 2024) Yost, James E., author; Whitley, Darrell, advisor; Ghosh, Sudipto, committee member; Betten, Anton, committee member
We investigate the efficacy of pruning a visual transformer during training to reduce inference time while maintaining accuracy. Various training techniques were explored, including epoch-based training, fixed-time training, and training to achieve a specific accuracy threshold. Results indicate that pruning from the inception of training offers significant reductions in inference time without sacrificing model accuracy. Different pruning rates were evaluated, demonstrating a trade-off between training speed and model compression. Slower pruning rates allowed for better convergence to higher accuracy levels and more efficient model recovery. Furthermore, we examine the cost of pruning and the recovery time of pruned models. Overall, the findings suggest that early-stage pruning strategies can effectively produce smaller, more efficient models with comparable or improved performance compared to non-pruned counterparts, offering insights into optimizing model efficiency and resource utilization in AI applications.
Open Access
Quality assessment of protein structures using graph convolutional networks
(Colorado State University. Libraries, 2024) Roy, Soumyadip, author; Ben-Hur, Asa, advisor; Blanchard, Nathaniel, committee member; Zhou, Wen, committee member
The prediction of protein 3D structure is essential for understanding protein function, drug discovery, and disease mechanisms; with the advent of methods like AlphaFold that are capable of producing very high quality decoys, ensuring the quality of those decoys can provide further confidence in the accuracy of their predictions. In this work we describe Qε, a graph convolutional network that utilizes a minimal set of atom and residue features as input to predict the global distance test total score (GDTTS) and local distance difference test score (lDDT) of a decoy. To improve the model's performance, we introduce a novel loss function based on the ε-insensitive loss function used for SVM-regression. This loss function is specifically designed for the characteristics of the quality assessment problem, and provides predictions with improved accuracy over standard loss functions used for this task. Despite using only a minimal set of features, it matches the performance of recent state-of-the-art methods like DeepUMQA. The code for Qε is available at https://github.com/soumyadip1997/qepsilon.
Open Access
Scalable and efficient tools for multi-level tiling
(Colorado State University. Libraries, 2008) Renganarayana, Lakshminarayanan, author; Rajopadhye, Sanjay, advisor
In the era of many-core systems, application performance will come from parallelism and data locality. Effective exploitation of these require explicit (re)structuring of the applications. Multilevel (or hierarchical) tiling is one such structuring technique used in almost all high-performance implementations. Lack of tool support has limited the use of multi-level tiling to program optimization experts. We present solutions to two fundamental problems in multi-level tiling, viz., optimal tile size selection and parameterized tiled loop generation. Our solutions provide scalable and efficient tools for multi-level tiling. Parameterized tiled code refers to tiled loops where the tile sizes are not (fixed) compile-time constants but are left as symbolic parameters. It can enable selection and adaptation of tile sizes across a spectrum of stages through compilation to run-time. We introduce two polyhedral sets, viz., inset and outset, and use them to develop a variety of scalable and efficient multi-level tiled loop generation algorithms. The generation efficiency and code quality are demonstrated on a variety of benchmarks such as stencil computations and matrix subroutines from BLAS. Our technique can generate tiled loop nests with parameterized, fixed or mixed tile sizes, thereby providing a one-size-fits all solution ideal for inclusion in production compilers. Optimal tile size selection (TSS) refers to the selection of tile sizes that optimize some cost (e.g., execution time) model. We show that these cost models share a fundamental mathematical property, viz., positivity, that allows us to reduce optimal TSS to convex optimization problems. Almost all TSS models proposed in the literature for parallelism, caches, and registers, lend themselves to this reduction. We present the reduction of five different TSS models proposed in the literature by different authors in a variety of tiling contexts. Our convex optimization based TSS framework is the first one to provide a solution that is both efficient and scalable to multiple levels of tiling.
Open Access
Improving software maintainability through aspectualization
(Colorado State University. Libraries, 2009) Mortensen, Michael, author; Ghosh, Sudipto, advisor; Bieman, James M., advisor
The primary claimed benefits of aspect-oriented programming (AOP) are that it improves the understandability and maintainability of software applications by modularizing cross-cutting concerns. Before there is widespread adoption of AOP, developers need further evidence of the actual benefits as well as costs. Applying AOP techniques to refactor legacy applications is one way to evaluate costs and benefits. Aspect-based refactoring, called aspectualization, involves moving program code that implements cross-cutting concerns into aspects. Such refactoring can potentially improve the maintainability of legacy systems. Long compilation and weave times, and the lack of an appropriate testing methodology are two challenges to the aspectualization of large legacy systems. We propose an iterative test driven approach for creating and introducing aspects. The approach uses mock systems that enable aspect developers to quickly experiment with different pointcuts and advice, and reduce the compile and weave times. The approach also uses weave analysis, regression testing, and code coverage analysis to test the aspects. We developed several tools for unit and integration testing. We demonstrate the test driven approach in the context of large industrial C++ systems, and we provide guidelines for mock system creation. This research examines the effects on maintainability of replacing cross-cutting concerns with aspects in three industrial applications. We study several revisions of each application, identifying cross-cutting concerns in the initial revision, and also cross-cutting concerns that are added in later revisions. Aspectualization improved maintainability by reducing code size and improving both change locality and concern diffusion. Costs include the effort required for application refactoring and aspect creation, as well as a small decrease in performance.
Open Access
Exploring the bias of direct search and evolutionary optimization
(Colorado State University. Libraries, 2008) Lunacek, Monte, author; Whitley, Darrell, advisor
There are many applications in science that yield the following optimization problem: given an objective function, which set of input decision variables produce the largest or smallest result? Optimization algorithms attempt to answer this question by searching for competitive solutions within an application's domain. But every search algorithm has some particular bias. Our results show that search algorithms are more effective when they cope with the features that make a particular application difficult. Evolutionary algorithms are stochastic population-based search methods that are often designed to perform well on problems containing many local optima. Although this is a critical feature, the number of local optima in the search space is not necessarily indicative of problem difficulty. The objective of this dissertation is to investigate how two relatively unexplored problem features, ridges and global structure, impact the performance of evolutionary parameter optimization. We show that problems containing these features can cause evolutionary algorithms to fail in unexpected ways. For example, the condition number of a problem is one way to quantify a ridge feature. When a simple unimodal surface has a high condition number, we show that the resulting narrow ridge can make many evolutionary algorithms extremely inefficient. Some even fail. Similarly, funnels are one way of categorizing a problem's global structure. A single-funnel problem is one where the local optima are clustered together such that there exists a global trend toward the best solution. This trend is less predicable on problems that contain multiple funnels. We describe a metric that distinguishes problems based on this characteristic. Then we show that the global structure of the problem can render successful global search strategies ineffective on relatively simple multi-modal surfaces. Our proposed strategy that performs well on problems with multiple funnels is counter-intuitive. These issues impact two real-world applications: an atmospheric science inversion model and a configurational chemistry problem. We find that exploiting ridges and global structure results in more effective solutions on these difficult real-world problems. This adds integrity to our perspective on how problem features interact with search algorithms, and more clearly exposes the bias of direct search and evolutionary algorithms.