教学课程资源库 更多>>

关键词： Biomedical engineering   Bioengineering   Computer science   Biophysics   Artificial intelligence  

摘要： Biological soft tissue comprises many parts of the human body. Disorders that occur in soft tissue are wide-ranging in type and degree and require biomechanical studies for better understanding. This study aims to investigate the yield point, which is a specific biomechanical property derived from elastic curves from extension testing of soft tissue specimens. When a soft tissue specimen is stretched, there exists a certain point where further extension will cause plastic deformation, or irreversible damage to the tissue; this is referred to as the yield point. The gold standard method of identifying this point in biological soft tissues is through visual inspection, which is unreliable and prone to subjectivity. An objective methodology for determining the yield point can enable findings relating to two purposes – demarking the elastic region for determining elastic parameters and for determining the yield strength, a conceivable metric of a specimen’s susceptibility to failure. Identifying the yield point is not straightforward for biological soft tissues. The definition of the yield point is mixed and further aggravated by the nonlinear force-extension response and micromechanics of gradual collagen recruitment. For this reason, approaches for yield point identification used for classic structural engineering materials are not viable. This study proposes a machine learning-based methodology to objectively identify the yield point. A decision tree algorithm is trained to identify unique feature splits and patterns in soft tissue biomechanical behavior. Additionally, this study investigates the viability of two previously developed methods of identifying the yield point and evaluates the collective methodologies as tools for soft tissue examination.

关键词： Computer science   Artificial intelligence   Datasets   Sentiment analysis   Support vector machines   Bias   Classification  

摘要： Interactive websites generate terabytes of data on a daily basis. This data can be used in multiple analytical applications to teach computers more about human behavior. Text classification is such an application. Multiple freely available user-generated text data can be used to teach computers to identify the sentiments behind a user’s on-screen interactions without the need of any human intervention. Sentiment analysis is an interesting problem, solving which would theoretically get a computer closer to passing the Turing test. Through this thesis, we test the ability of a classifier to accurately identify user sentiments. However, we do not focus on standard classification settings and the aim is to train the classifier in such a way that it would also be effective in identifying sentiment behind user generated text generated from a completely new social media platform. To be able to do this, we must first identify behavioral bias based on user interactions in two different social media sites as well as websites that accept user reviews. This bias must then be mitigated in order to obtain an unbiased classifier that can then be used to identify user sentiments on any social media platform. For the research in this thesis, such user-generated text is obtained from the social media sites Reddit and Twitter. We also obtain product review data related to both books and wine. Various natural language processing techniques are then employed to process the data and extract similar and dissimilar trends. Vectorized user text would be used to train sentiment classifiers. Finally, classification bias would be identified and mitigated in order to obtain classifiers that can identify human sentiments in real-time with an improved accuracy with limited dependency on source information.

关键词： Computer science   Artificial intelligence  

关键词： Artificial intelligence   Computer science  

关键词： Electrical engineering   Energy   Computer science   Sustainability  

摘要： The electric power system is undergoing dramatic transformations due to the emergence of renewable resources and demand-side revolutions. However, in order to face the increasing level of system complexity and uncertainty, we need to come up with algorithms that are able to operate the power grid in a safe, reliable and sustainable manner. Such algorithms need to be compatible with infrastructure advancements as well as societal needs. We argue in this thesis that data-driven methods combined with physical knowledge and optimization theories provide a set of powerful tools that can achieve performance guarantees in a diverse set of energy system applications. This dissertation covers the uncertainty modeling, optimal control, and decision-making for power systems with high penetration of renewables. Leveraging the availability of sensing, actuation and data platforms, we are able to construct specifically-design data-driven algorithms by incorporating domain knowledge such as time series characterization, function convexity, and optimization problem structures. In particular, we present three algorithmic contributions in this work: we present a generative model to forecast possible future scenarios of renewable generators; we illustrate an optimal control framework which is built upon on input convex neural network; we formulate a machine learning task for the optimal power flow problem, and we come up with a generalizable and robust learning-based optimization solver to enable real-time decision-making under the environment of stochastic electricity demand. The development of such algorithms is built upon the foundation of control, optimization and machine learning theories. In addition to these algorithmic contributions, a central focus of this thesis is the application of such data-driven tools in many challenging energy system tasks, while extending the state of the art. The developed methods have enabled more robust electric grid planning and energy storage ope

关键词： Decision trees   Deformation   Acoustics   Vibration   Support vector machines   Computer science   Mechanics   Physics  

摘要： Today's electrical machines are finding many applications in new environments where lowernoise levels are demanded, especially in electric vehicles. It is, therefore, imperative thatthe designs must satisfy certain acoustic noise requirements. In the past, experimental electric motors were developed in the early design processto assess their noise performance before full scale production. This approach could bean unpleasant experience for motor designers, as the mitigation solutions often neededsubstantial modification to their prototype models. These prototypes can be very expensive, and forcing in engineering changes to an existing machine can be a nightmare. For thisreason, the use of computer software in noise and vibrations analysis has attracted a lot ofattention in recent years. This thesis uses computer-aided predictive techniques such as the finite element methodto set the background for the computational discussions and exploration of some of theresearch gaps in the literature. It focuses on the interior permanent magnet synchronousmotor due to its extensive application, particularly in the automotive industry. In this work, the effect of non-sinusoidal current waveforms on the acoustic noise field ofthe electric motor drive has been analyzed. It is discovered that sampling the airgap spaceat different instants yields inconsistent electromagnetic force results. Therefore, this thesisproposes an efficient way of extracting the electromagnetic forces in the airgap space toaccount for very critical harmonics that may be ignored in conventional techniques such asthe single-time sampling. It also defines two new acoustic quantities to accurately estimatethe resulting acoustic noise for different switching frequencies of the power electronic circuit. Also, since the thermal behavior of the electric motor is mostly linked to the lossesthat result from the current supply system, the analysis incorporates the role of temperaturein defining the acoustic field of the

关键词： Microwave integrated circuits   Mutual information   Entropy   Feature extraction   Computer science   Time complexity   Licenses   Data mining   association detection   information entropy   maximum information coefficient   upper confidence bound  

摘要： The maximum information coefficient (MIC) is a novel and widely-using measure of association detection in large datasets. The most outstanding feature of MIC is that it has both generality and equability. However, MIC can only deal with two variables and cannot precisely estimate coupling associations of multiple variables. In this paper, we propose an extension of MIC to deal with multi-variable datasets, called the multi-variable maximum information coefficient (MMIC). Some inherited and novel properties of MMIC are proved, including generality, equability, monotonicity, and subadditivity. We design an algorithm based on greedy stepwise strategy and upper confidence bound (UCB) for an approximate calculation of MMIC. The tests of MMIC on generated datasets and examples on real datasets are carried out to detect known and novel associations.

关键词： Mathematics   Theoretical mathematics   Computer science  

摘要： This thesis deals with factorizations of integers into primes, with the Quadratic Sieve, which is one of the most important modern algorithms for factorization, and with a computer program to implement the Quadratic Sieve. It uses well-established fields of mathematics, in particular number theory and linear algebra

关键词： Computer science   Engineering   Agriculture   Applied Mathematics   Energy   Fluid mechanics   Hydraulic engineering   Mathematics   Mechanics   Physics  

摘要： The vortex pump is a kind of impurity pump. Characteristics of vortex pumps are their recessed, semi-open impeller and their enlarged gap between the impeller and the casing. The gap and impeller give the pump a good non-clogging performance. The pump is widely used in transporting sewage water or mixed liquid containing fiber or large particle. However, vortex pumps are relatively inefficient. Therefore, it is quite necessary to improve the efficiency of the vortex pump. But the flow inside the vortex pump is complicated. The mechanism of its internal flow has not been yet well comprehended. It remains unclear how exactly vortex pumps can be optimized. The current designing theory is based on the centrifugal pump, which is different from three-dimensional flow in the vortex pump. Therefore, the deep study on the internal flow mechanism of the vortex pump is needed. This dissertation investigates the influences of the impeller geometry to the characteristics and the flow mechanism of a vortex pump. The current experimental measurements can hardly see the detail of the entire flow field. In this study, the computational fluid dynamic is used. The three-dimensional model of the vortex pump is built and meshed with high-quality structural grids. Based on CFX, Navier-Stokes equations solved with k-ε turbulence model. The internal flow field of the vortex pump is investigated. The performance curve and the distribution of velocity and pressure are acquired. The impellers with different design parameters, such as blade outlet angle, blade number, blade winglet, is calculated and the results are validated with test data. The influence of design parameters on the characteristic curve is compared. Their mechanism is further discussed. The conclusions are summarized, which will help in the designing theory.

关键词： Computer engineering   Computer science   Artificial intelligence  

摘要： Deep learning, while an important machine learning technique, is susceptible to adversarial example attacks. Adversarial examples generated by adding perturbations on clean images/video frames can lead to mispredictions of deep neural networks. Moreover, deep learning/machine learning can be used to deceive humans by generating adversarially falsified media e.g., deepfake attacks. The thesis work will study the above two attack scenarios, i.e., machine-centric adversary and human-centric adversary, with targets to fool ML decisions and human decisions, respectively. We aim to build a generalizable and scalable supervised learning system for classifying attack attributes behind the machine-centric attacks as well as the human-centric attacks. We start from building an integrated Attack Toolchain Library (ATL) with a broad coverage of both machine-centric and human-centric adversaries, as well as through an integrated user interface for great flexibility and extensibility to serve our downstream tasks. Based on the developed ATL, we further design a meta-classifier pipeline architecture for predicting attack attributes. The proposed overall meta-classifier shows effectiveness in dealing with false alarms and data distribution shift, and generalization to both machine-centric and human-centric attacks.