教学课程资源库 更多>>

关键词： Subspace constraints   Software   Strips   Time complexity   Software testing   Computer science   Software testing   random testing   adaptive random testing   restricted random testing   exclusion zone   largest available zone  

摘要： Adaptive random testing (ART) has been proposed to enhance the effectiveness of random testing (RT) through more even spreading of the test cases. In particular, restricted random testing (RRT) is an ART algorithm based on the intuition of skipping all the candidate test cases that are within the neighborhoods (or zones) of previously executed test cases. RRT has higher effectiveness than RT in terms of failure detection but incurs a higher time cost. In this paper, we aim to further reduce the time costs for RRT and improve the effectiveness for RT and ART methods. We propose a proactive technique known as & x201C;RRT by largest available zone & x201D;(RRT-LAZ). Like RRT, RRT-LAZ first defines an exclusion zone around every executed test case in order to determine the available zones. Unlike the original RRT, RRT-LAZ then compares all the available zones to proactively pick the largest one, from which the next test case is randomly generated. Both simulation analyses and empirical studies have been employed to investigate the efficiency and effectiveness of RRT-LAZ in relation to RT and related ART algorithms. The results show that RRT-LAZ has significantly lower time costs than RRT. Furthermore, RRT-LAZ is more effective than RT and related ART methods for block failure patterns in low-dimensional input spaces. In general, since RRT-LAZ employs a proactive technique instead of a passive one in generating next cases, it is much more cost-effective than RRT. RRT-LAZ is also more cost-effective than RT and other ART methods that we have studied.

关键词： Information science   Computer science  

摘要： Augmented reality has come a long way in the past decades in the manner of reduced costs and increased reliability. Despite these developments, there is still a difficulty bridging the gap between practical use and pure theory. This work explores if the current state of augmented reality is at a stage that can be utilized to implement solutions for heavy machinery maintenance routines by exploring the quality of education from augmented reality applications, capturing said maintenance routines for our own system, then replicating them in an augmented reality application with consumer grade equipment.

关键词： Computer science  

摘要： Increasingly, decisions are based on insights and conclusions derived from data-driven analyses. In this dissertation, we address important technical challenges that contribute to improving the quality of such analyses. First, since they crucially depend on the quality of data, we propose PODS (Predictable Outliers in Data-trendS), a method that derives explanations for data outliers by identifying meaningful relationships between them. These explanations help users determine whether outliers correspond to bad data (which could be cleaned) or, rather, to important features (which should be accounted for). Second, we are interested in improving the quality of predictive Machine Learning models that generate poor predictions. In these cases, the problem may be lack of data, and the solution might involve adding more informative, high-quality data to the model. To this end, we propose IDA (Identifying Datasets for Augmentation), a method that identifies datasets to augment the input data of a model based on their likelihood of improving the model’s quality. In order to help data scientists derive high-quality results that they can trust, we envision a new generation of methods that explore dataset features and relationships to support more sound data-driven analyses, helping users reach more informed conclusions and make decisions confidently. By proposing PODS and IDA, we take a step towards this larger vision.

关键词： Computer science  

摘要： Program comprehension is a sub-field of software engineering that seeks to understand how developers understand programs. Comprehension acts as a starting point for many software engineering tasks such as bug fixing, refactoring, and feature creation. The dissertation presents a series of empirical studies to understand how developers comprehend software in realistic settings. The unique aspect of this work is the use of eye tracking equipment to gather fine-grained detailed information of what developers look at in software artifacts while they perform realistic tasks in an environment familiar to them, namely a context including both the Integrated Development Environment (Eclipse or Visual Studio) and a web browser (Google Chrome). The iTrace eye tracking infrastructure is used for certain eye tracking studies on large code files as it is able to handle page scrolling and context switching. The first study is a classroom-based study on how students actively trained in the classroom understand grouped units of C++ code. Results indicate students made many transitions between lines that were closer together, and were attracted the most to if statements and to a lesser extent assignment code. The second study seeks to understand how developers use Stack Overflow page elements to build summaries of open source project code. Results indicate participants focused more heavily on question and answer text, and the embedded code, more than they did the title, question tags, or votes. The third study presents a larger code summarization study using different information contexts: Stack Overflow, bug repositories and source code. Results show participants tended to visit up to two codebase files in either the combined or isolated codebase session, but visit more bug report pages, and spend longer time on new Stack Overflow pages they visited, when given either these two treatments in isolation. In the combined session, time spent on the one or two codebase files they viewed

关键词： Computer science  

摘要： The administration of routine interventions, from breathing support to pain management, constitutes a major part of inpatient care. Thoughtful treatment is crucial to improving patient outcomes and minimizing costs, but these interventions are often poorly understood, and clinical opinion on best protocols can vary significantly. Through a series of case studies of key critical care interventions, this thesis develops a framework for clinician-in-loop decision support. The first of these explores the weaning of patients from mechanical ventilation: admissions are modelled as Markov decision processes (MDPs), and model-free batch reinforcement learning algorithms are employed to learn personalized regimes of sedation and ventilator support, that show promise in improving outcomes when assessed against current clinical practice. The second part of this thesis is directed towards effective reward design when formulating clinical decisions as a reinforcement learning task. In tackling the problem of redundant testing in critical care, methods for Pareto-optimal reinforcement learning are integrated with known procedural constraints in order to consolidate multiple, often conflicting, clinical goals and produce a flexible optimized ordering policy. The challenges here are probed further to examine how decisions by care providers, as observed in available data, can be used to restrict the possible convex combinations of objectives in the reward function, to those that yield policies reflecting what we implicitly know from the data about reasonable behaviour for a task, and that allow for high-confidence off-policy evaluation. The proposed approach to reward design is demonstrated through synthetic domains as well as in planning in critical care. The final case study considers the task of electrolyte repletion, describing how this task can be optimized using the MDP framework and analysing current clinical behaviour through the lens of reinforcement learning, before going

关键词： Linguistics   Feature extraction   Support vector machines   Bibliographies   Electronic mail   Computer science   Business   Online product reviews   spam reviews   spam review detection   linguistic features   spammer behavioral features  

摘要： Online reviews regarding different products or services have become the main source to determine public opinions. Consequently, manufacturers and sellers are extremely concerned with customer reviews as these have a direct impact on their businesses. Unfortunately, to gain profits or fame, spam reviews are written to promote or demote targeted products or services. This practice is known as review spamming. In recent years, the spam review detection problem has gained much attention from communities and researchers, but still there is a need to perform experiments on real-world large-scale review datasets. This can help to analyze the impact of widespread opinion spam in online reviews. In this work, two different spam review detection methods have been proposed: (1) Spam Review Detection using Behavioral Method (SRD-BM) utilizes thirteen different spammer's behavioral features to calculate the review spam score which is then used to identify spammers and spam reviews, and (2) Spam Review Detection using Linguistic Method (SRD-LM) works on the content of the reviews and utilizes transformation, feature selection and classification to identify the spam reviews. Experimental evaluations are conducted on a real-world Amazon review dataset which analyze 26.7 million reviews and 15.4 million reviewers. The evaluations show that both proposed models have significantly improved the detection process of spam reviews. Specifically, SRD-BM achieved 93.1% accuracy whereas SRD-LM achieved 88.5% accuracy in spam review detection. Comparatively, SRD-BM achieved better accuracy because it works on utilizing rich set of spammers behavioral features of review dataset which provides in-depth analysis of spammer behaviour. Moreover, both proposed models outperformed existing approaches when compared in terms of accurate identification of spam reviews. To the best of our knowledge, this is the first study of its kind which uses large-scale review dataset to analyze different spammers'

关键词： Artificial intelligence   Computer science  

摘要： Machine learning has been playing an important role in our lives for quite a while now. It has been used for a multitude of applications ranging from voice recognition to disease detection. In traditional machine learning, a learning algorithm is presented with a large set of data and it is the task of the algorithm to find a model that explains the data best. This process, even though useful, can be very time consuming. With the help of a teacher the time required to train a learner can be reduced drastically. Training a machine learner with the help of a teacher is known as optimal control for machine learning (also known as machine teaching), and is applicable in a multitude of domains including education and security. In this thesis, we will explore some such usefulness of optimal control. We first provide evidence that optimal control can be helpful to teach humans, both for learning perceptual fluency in chemistry and the pronunciation of written words. Our research also verifies that an artificial neural network can be useful as a cognitive model for humans. Then we present a new type of attack on machine learners called training set camouflage. In this attack a malicious agent can train a publicly available learner on a sensitive task with the help of a benign looking training set. The malicious agent does not raise any suspicion while passing this training set. Finally, we show a comparative analysis of the different metrics used in adversarial machine learning. In particular, through human experiments we show that even though none of the metrics currently used in practice is suitable, pixel 3-norm provides the best approximation. The research works presented in this thesis show how a teacher can play an important and useful role to train a machine learning algorithm.

关键词： Educational Applications   Gender Bias   Linguistic Complexity   Machine Learning   Natural Language Processing   Electrical engineering   Computer science   Educational tests & measurements   To Be Assigned  

摘要： There is a growing community of research focusing on educational applications of natural language processing (NLP). The applications tend to focus on analysis of student writing for scoring and feedback, and analysis of language learning. There has been less focus on analysis of language use in educational content, like assessment questions and textbooks, which is largely an expert driven process. This work examines this space, presenting automated tools for analysis of language use in K-16 science, technology, engineering and mathematics (STEM) education, and demonstrates the utility of automatically extracted features in studying student performance. This work also serves to bridge research in educational measurement and machine learning, providing a machine learning framework for analysis of factors that contribute to the difficulty of science assessment items. Within the broader umbrella of language use, this work focuses on two aspects: language difficulty (or linguistic complexity), and gender representation. Linguistic complexity has been studied from both the expert driven educational perspective and in the context of machine learning and NLP based tools. For the latter, models have shown a high agreement with expert annotation for longer documents, however, have not been shown to work well for shorter, informational texts. This work presents a discourse aware hierarchical neural model for classification of linguistic complexity quantified as grade level, demonstrated to work accurately for shorter texts, achieving state-of-the- art performance. Unlike most existing NLP based methods, the performance of our model is also validated for the downstream task of predicting student performance, where we find an impact both for K-12 and college level STEM assessments. The model for classification also generalizes to other text classification problems. Educational measurement research for prediction of difficulty of assessments questions is important in the context

关键词： Finance   Artificial intelligence   Computer science  

摘要： Restaurant servers are an example of an insider threat to the security of restaurant financial data. This paper applies machine learning to detect the digital representation of malevolent behavior of restaurant employees. The results of this research could be used to notify restaurant owners in real time when fraud is being committed. This paper applies machine learning (ML) techniques including neural networks, support vector machines, Random Forest, and Adaboost, to detecting insider fraud in restaurant point-of-sales data. By developing engineered features, and applying undersampling and oversampling class balancing techniques, and statistically generated data we show that ML techniques can improve fraud detection performance. In particular, detection with a Random Forest model using cross validation can be increased 55% by oversampling the minority class to the same size as the majority class. And results with a Neural Net model trained to detect fraud on the first year the restaurant opened, and tested on data from the following year can be improved by 50% by decreasing the majority class to be the same size as the minority class. We show that with statistically generated data, the performance of preprocessing features matches the performance of engineered features and achieve 99.9% F1-Scores.

关键词： Information technology   Information science   Computer science  

摘要： Conversational search systems often fail to recognize the information need of the user, especially for exploratory and complex tasks where the question is non-factoid in nature. In any conversational search environment, spoken dialogues by the user communicate the search intent and the information need of the user to the system. In response, the system performs specific, expected search actions. This is a domain-specific natural language understanding problem where the agent must understand the user’s utterances and act accordingly. Prior literature in intelligent systems suggests that in a conversational search environment, spoken dialogues communicate the search intent and the information need of the user. The meaning of these spoken utterances can be deciphered by accurately identifying the speech or dialogue acts associated with them. However, only a few studies in the information retrieval community have explored automatic classification of speech acts in conversational search systems, and this creates a research gap. Also, during spoken search, the user rarely has control over the search process as the actions of the system are hidden from the user. This eliminates the possibility of correcting the course of search (from the user’s perspectives) and raises concerns about the quality of the search and the reliability of the results presented. Previous research in human-computer interaction suggests that the system should facilitate user-system communication by explaining its understanding of the user’s information problem and the search context (referred to as the system’s model of the user). Such explanations could include the system’s understanding of the search on an abstract level and the description of the search process undertaken (queries and information sources used) on a functional level. While these interactions could potentially help the user and the agent to understand each other better, it is essential to evaluate if explicit clarifications are nec