教学课程资源库 更多>>

关键词： Artificial intelligence   Computer science  

摘要： Classical AI planning aims to solve the problem of searching for a sequence of steps that transitions an agent from some state of the world to some goal. Hierarchical planning is a particular case of AI planning that assumes the existence of a decomposition model that represents how to break down tasks or goals into sequences of actions. While hierarchical planning allows us to reduce the search space compared to classical planning, it has been shown that hand authoring these decomposition models require domain-expert knowledge and time, and can be error-prone for complex application domains. This is also the case for hierarchical plan recognition, which aims to solve the problem of inferring the goals and plans of agent(s) given a sequence of actions and a decomposition model. Our goal is to study learning techniques for automatically constructing decomposition models from training data for hierarchical planning and plan recognition. Specifically, we focus on learning a decomposition model represented using Combinatory Categorial Grammars (CCGs). This dissertation describes a study on supervised and unsupervised CCG learning techniques for hierarchical planning and plan recognition, a novel CCG plan recognition algorithm, a theoretical analysis of supervised CCG learning algorithms, and the application of learned CCGs for playing computer games. In this dissertation, we provide a background of relevant research to the dissertation, describe our study on CCG learning, and outline future steps.

关键词： Neurosciences   Applied mathematics   Artificial intelligence   Physiology   Computer science   Propagation   Deep learning   Back propagation   Assignment problem   Optimization   Neural networks   Algorithms  

摘要： Synaptic plasticity is the primary physiological mechanism underlying learning in the brain. It is dependent on pre- and post-synaptic neuronal activities, and can be mediated by neuromodulatory signals. However, to date, computational models of learning that are based on pre- and post-synaptic activity and/or global neuromodulatory reward signals for plasticity have not been able to learn complex tasks that animals are capable of. In the machine learning field, neural network models with many layers of computations trained using gradient descent have been highly successful in learning difficult tasks with near-human level performance. To date, it remains unclear how gradient descent could be implemented in neural circuits with many layers of synaptic connections. The overarching goal of this thesis is to develop theories for how the unique properties of neurons can be leveraged to enable gradient descent in deep circuits and allow them to learn complex tasks. The work in this thesis is divided into three projects. The first project demonstrates that networks of cortical pyramidal neurons, which have segregated apical dendrites and exhibit bursting behavior driven by dendritic plateau potentials, can in theory leverage these physiological properties to approximate gradient descent through multiple layers of synaptic connections. The second project presents a theory for how ensembles of pyramidal neurons can multiplex sensory and learning signals using bursting and short-term plasticity, in order to approximate gradient descent and learn complex visual recognition tasks that previous biologically inspired models have struggled with. The final project focuses on the fact that machine learning models implementing gradient descent assume symmetric feedforward and feedback weights, and presents a theory for how the spiking properties of neurons can enable them to align feedforward and feedback weights in a network. As a whole, this work aims to bridge the gap between pow

关键词： Norms   Expected values   Society   Computer science   Artificial intelligence  

摘要： Norms help regulate a society. Norms can naturally emerge from interactions in a society, including conflicts between existing norms and how those conflicts are resolved. We study explicit norms, those that are represented in some structured form by the agents. Previous studies on emergence have not tackled explicit norms. We propose an approach for the emergence of explicit norms by developing agents who provide and reason about explanations for norm violations in deciding what sanctions to produce and identifying alternative norms. These agents use a genetic algorithm to create and evolve a set of norms and reinforcement learning to learn the value of these norms. We find that providing and evaluating explanations leads to the emergence of norms that provide better cohesion and goal satisfaction for member agents. Our results are stable for societies with differing attitudes of generosity.

关键词： Neural networks   Neurons   Logistics   Convergence   Machine learning   Licenses   Computer science   Artificial neural network   deep neural network   learning challenges   activation function  

摘要： In artificial neural networks, activation functions play a significant role in the learning process. Choosing the proper activation function is a major factor in achieving a successful learning performance. Many activation functions are sufficient universal approximators, but their performance is lacking. Thus, many efforts have been directed toward activation functions to improve the learning performance of artificial neural networks. However, the learning process involves many challenges, such as saturation, dying, and exploding/vanishing the gradient problems. The contribution of this work resides in several axes. First, we introduce two novel activation functions: absolute linear units and inverse polynomial linear units. Both activation functions are augmented by an adjustable parameter that controls the slope of the gradient. Second, we present a comprehensive study and a taxonomy of various types of activation functions. Third, we conduct a broad range of experiments on several deep neural architecture models with consideration of network type and depth. Fourth, we evaluate the proposed activation functions' performance in image and text classification tasks. For this purpose, several public benchmark datasets are utilized to evaluate and compare the performance of the proposed functions with that of a group of common activation functions. Finally, we deeply analyze the impact of several common activation functions on deep network architectures. Results reveal that the proposed functions outperform most of the popular activation functions in several benchmarks. The statistical study of the overall experiments on both classification categories indicates that the proposed activation functions are robust and superior among all the competitive activation functions in terms of average accuracy.

关键词： Diseases   Computational modeling   Predictive models   Neural networks   Databases   Computer science   Encoding   Microbe-disease associations   neural network   pairwise loss   Bayesian personalized ranking  

摘要： Identifying the microbe-disease associations is conducive to understanding the pathogenesis of disease from the perspective of microbe. In this paper, we propose a deep matrix factorization prediction model (DMFMDA) based on deep neural network. First, the disease one-hot encoding is fed into neural network, which is transformed into a low-dimensional dense vector in implicit semantic space via embedding layer, and so is microbe. Then, matrix factorization is realized by neural network with embedding layer. Furthermore, our model synthesizes the non-linear modeling advantages of multi-layer perceptron based on the linear modeling advantages of matrix factorization. Finally, different from other methods using square error loss function, Bayesian Personalized Ranking optimizes the model from a ranking perspective to obtain the optimal model parameters, which makes full use of the unobserved data. Experiments show that DMFMDA reaches average AUCs of 0.9091 and 0.9103 in the framework of 5-fold cross validation and Leave-one-out cross validation, which is superior to three the-state-of-art methods. In case studies, 10, 9 and 9 out of top-10 candidate microbes are verified by recently published literature for asthma, inflammatory bowel disease and colon cancer, respectively. In conclusion, DMFMDA is successful application of deep learning in the prediction of microbe-disease association.

关键词： 师生关系   生命共生   师生之“乐  

摘要： 孔子提出的“有朋自远方来，不亦乐乎”之命题，以“乐”指称了师生关系的理想形态。“乐”是指以“悦”为前提的、师生双方在交往活动中的交互性情感体验，本质上表达了师生“道”之相遇。儒家师生之“乐”表征了一种朝向生命共生的师生关系形态，师生双方同时兼具主体与客体之双重属性，以“道”为中心形成由互动而联结的生命共生体。就价值取向而言，师生之“乐”彰明了教师之“乐”的主导性，以“师”与“道”的内在联结为前提与根据，从而确证一种基于共识的教育性。就对他者性的确证方式而言，师生双方与“道”的真实联结，使二者的互动关系具体展开为道德义务的双向对待，儒家师生关系由此表现为对称的伦理关系。

关键词： Computer science   Educational technology   Science education  

关键词： Deep learning   Electricity   Power   Signal processing   Neural networks   Internet of Things   Batteries   Algorithms   Privacy   Time series   Computer science   Electrical engineering   Energy   Engineering   Web studies  

摘要： The modern electricity network, known as the smart grid (SG), aims at improving the efficiency, reliability, and security of electric power systems by using intelligent transmission, control, and distribution networks. One of the main components of the SG are smart meters (SMs), which are devices that enable data exchange between users and utility provider (UP) by recording fine-grained power consumption measurements of households. This fine-grained electricity consumption data is used for billing, load forecasting, energy theft detection, and several other applications for improving the grid operation. However, the SM data can be shared with a third-party, which can potentially infer sensitive information about users, including the behavioral patterns or even the type of appliances used in the dwelling. Such a users' privacy violation is one of the main reasons why UPs are not willing to share SM datasets with third-parties (e.g. research groups at Universities). Moreover, this privacy issue can be considered from consumers' points of view, i.e. it concerns the users that their electricity data is eavesdropped or even they might not trust UPs. These two views of privacy concern with regards to SM data led to emerging two main families of SM privacy-preserving models in the literature, called data manipulation and demand load shaping. In this thesis, we deal with the aforementioned privacy matters in SM data sharing, where machine learning based privacy-preserving frameworks are developed to mitigate the privacy concerns. Firstly, within the data manipulation framework, we develop an information-theoretic based causal privacy-preserving model implemented using recurrent neural networks, which, unlike all the previously proposed models, incorporates the temporal correlation of the SM data without imposing any explicit assumptions on the generating model of the power measurement. Generating of the release data for such a model is studied using two mechanisms including

关键词： Advanced placement Computer Science   Computer science education   Equity   Location   School size   Socio-economic  

摘要： The purpose of this study was to examine the extent to which differences in high school location, socioeconomic status (SES), and size influenced a student’s participation in Advanced Placement Computer Science (AP CS) in the state of Louisiana. A non-experimental, causal-comparative research design was used in this study in which archival data was retrieved from the State of Louisiana Department of Education and the National Center for Education Statistics (NCES) for the 2018/2019 school year. Descriptive and inferential statistical procedures were conducted to determine the difference in Advanced Placement Computer Science (AP CS) with respect to school location, SES, and size. Findings revealed that school location and SES were statistically insignificant. School location and SES did not influence students’ participation in AP CS. School size was found to be statistically significantly related to AP CS participation; large high schools enrolled more students in AP CS than smaller schools. Additionally, results from this research indicated that access to AP CS did not imply AP CS participation. For instance, even though more urban schools offered AP CS courses, more students in urban schools did not enroll in AP CS when compared to students in rural schools.