教学课程资源库 更多>>

关键词： Energy   Materials science   Artificial intelligence   Computer science   Mathematics   Mining   Optics  

摘要： Utilizing computational methods for uncovering new materials with desirable properties is a powerful tool for materials discovery. The grand goal of these approaches is to efficiently search through all of material space in order to find the best candidates for any application. Compared to traditional “trial-and-error” based materials discovery approaches, computational materials discovery allows for: i) a diverse range of materials can be explored, beyond the already successful spaces, ii) easier optimization over multiple performance metrics, and iii) statistically motivated insights into the materials space can be derived. However, discovering materials in this manner is a very difficult optimization problem that can involve assessing hundreds of thousands of candidate materials according to multiple selection criteria. Understanding how to best utilize these materials datasets is therefore an open problem that requires the development of screening methodologies. In this dissertation, I explore the use of data-driven materials discovery methods for discovering new assembly-free van der Waals heterostructures and high brightness photocathode materials. Additionally, I will discuss the development new machine-learning methodologies to improve the efficiency and success of computationalbased material discovery efforts. Van der Waals heterostructures are layered materials that consist of vertical stacks of weakly bound two-dimensional (2D) layers. Since these 2D layers are held together through van der Waals interactions, a nearly endless number of unique van der Waals heterostructure can be synthesized, thus giving rise to a massive space for material design. However, the experimental synthesis of these van der Waals heterostructures can be challenging. One of the most common methods for synthesizing van der Waals heterostructures involves the slow and imprecise stacking of individual 2D layers that have been obtained from the mechanical exfoliation of a bulk layere

关键词： Artificial intelligence   Electrical engineering   Computer science  

关键词： Electrical engineering   Artificial intelligence   Public health   Neurosciences   Computer science  

摘要： Scalp electroencephalograms (EEGs) are used in a broad range of health care institutions to monitor and record electrical activity in the brain. EEGs are essential in diagnosis of clinical conditions such as epilepsy, seizure, coma, encephalopathy, and brain death. Manual scanning and interpretation of EEGs is time-consuming since these recordings may last hours or days. It is also an expensive process as it requires highly trained experts. Therefore, high performance automated analysis of EEGs can reduce time to diagnosis and enhance real-time applications by identifying sections of the signal that need further review. Automatic analysis of clinical EEGs is a very difficult machine learning problem due to the low fidelity of a scalp EEG signal. Commercially available automated seizure detection systems suffer from unacceptably high false alarm rates. Many signal processing methods have been developed over the years including time-frequency processing, wavelet analysis and autoregressive spectral analysis. Though there has been significant progress in machine learning technology in recent years, use of automated technology in clinical settings is limited, mainly due to unacceptably high false alarm rates. Further, state of the art machine learning algorithms that employ high dimensional models have not previously been utilized in EEG analysis because there has been a lack of large databases that accurately characterize clinical operating conditions. Deep learning approaches can be viewed as a broad family of neural network algorithms that use many layers of nonlinear processing units to learn a mapping between inputs and outputs. Deep learning-based systems have generated significant improvements in performance for sequence recognitions tasks for temporal signals such as speech and for image analysis applications that can exploit spatial correlations, and for which large amounts of training data exists. The primary goal of our proposed research is to develop deep le

关键词： Artificial intelligence   Computer science   Transportation  

摘要： Oil and gas pipelines are critical infrastructures that support modern human life by supplying oil and gas to many petrochemical-based facilities such as power plants, fertilizer plants, refueling stations, and others. The failure of pipelines can be devastating. Such failures cause environmental harm, property damage, and can kill or injure biological life. Proper inspections and regular maintenance are required to keep pipelinesin good working condition. Currently, most oil and gas pipeline inspections are done manually and incur tremendous costs. In many cases, inspections are done only to meet government regulations, making these pipelines vulnerable to catastrophic disasters. This thesis investigates the usability of unmanned aerial vehicles (UAV) and artificial intelligence (AI) in oil and gas pipeline inspections. A drone was used to collect video of the target pipeline and geographic information system (GIS) data. The video was segmented using a modified DeepLabV3+ segmentation network and then analyzed to identify pipeline alignment issues, missing crossbar issues, and pipeline corrosion issues. The experimental results demonstrated that the proposed pipeline inspection system can efficiently segment pipelines and supporting components in the drone videos, collected from Lamar University testbed. In addition, the system can detect pipeline alignment issues, missing crossbar issues, and pipeline corrosion issues successfully. The results obtained in this work can be used to improve oil and gas pipeline inspection practices.

关键词： Computer science   Electrical engineering   Datasets   Fourier transforms   Experiments   Neural networks   Decomposition   Breakdowns   Algorithms   Ablation   Energy consumption   Artificial intelligence   Computer engineering  

关键词： Computer science   Urban planning  

摘要： The complex interrelationship between the built environment and social problems is often described but frequently lacks the data and analytical framework for advances leading to widespread application. First, this study addresses this gap by presenting a machine learning (ML) approach to study whether street-level built environment visuals can be used to classify locations with high-crime and lower-crime activities. In training the model, spatialized expert narratives are used to classify locations (and therefore place types) as being linked to high crime. Google Street View (GSV) images are then extracted into semantic categories (e.g., road, sky, greenery, and building) through a deep learning image segmentation algorithm. From these local visual representatives are generated and used to train a classification model to identify similar potential high crime environments. The model is applied to multiple cities in the US. Results of the first part of the study show our model can predict high-and lower-crime areas with more than 98\% in the first test city, and above 95\% accuracy in the second test city. However, when the urban environment changes across regions, so the predictive capacity of the model falls. Second, this study presents a method to compute urban visual diversity using an Artificial Intelligence (AI)-based technique and shows its relationship with the social phenomenon. The developed framework show computational method to find single category and multiple category indices. A process to find important features from GSV images that can help to compute geospatial visual diversity. The second part of the study shows the reliability and validity of the method using Inter-Rater Reliability Indices (IRR). The results of the second part of the study indicate GSV images can be used to compute geospatial visual diversity and assess relationships between social phenomena and visual diversity indices.

关键词： Mining   Geotechnology   Computer science   Deep learning   Neural networks  

摘要： Autonomous technologies have the potential to transform mining by improving efficiency, safety, and productivity. The transition to fully autonomous mining requires an interoperable system that is autonomous at all stages of a mining operation including geotechnical hazard management. Deep learning is a viable option to develop autonomous technologies based on the improvements in computing power and sensor technology. However, the success of deep learning development depends on having a large and diverse training set. Geotechnical hazard management requires on-site data collection. The restrictions on on-site data collection such as accessibility, time, and financial constraints cause small and unbalanced datasets that limit the performance of deep learning systems and thereby prevent the increased use of autonomous technologies. This thesis presents a methodology that includes transfer learning, synthetic data generation, and an image-feature based data sampling technique to train a convolutional neural network (CNN) using a small dataset collected on-site. The methodology is implemented in a case study to develop an autonomous roof fall hazard detection system. The case study is a large-opening limestone mine in the Midwestern United States that has frequent roof fall problems caused by high horizontal stresses. The typical hazard management approach for this type of roof fall hazard relies heavily on visual inspections and expert knowledge. Therefore, images depicting hazardous and non-hazardous roof conditions are collected based on expert labeling. Using the transfer learning approach, a pre-trained network provided the suitable low-level features to train the final fully connected layer of a CNN. The network classification accuracy was improved by expanding the training set with synthetic images rendered from a 3-D model generated with digital photogrammetry. Also, a data sampling technique that uses Gabor magnitude responses was introduced. The objective of t

关键词： Computer engineering   Computer science   Artificial intelligence  

摘要： For Type 1 diabetes patients, they will have AP systems to inject insulin from outside to regulate their blood glucose level. Exercise will promote glucose intake and this effect will last up to 4 to 8 hours even after exercise. But for patients their insulin levels aren’t able to change rapidly. Thus, it is necessary to detect exercise for the AP system automatically so that it can regulate the insulin injected. In this paper, first, we proposed a new glucose model which considers the effect of exercise on glucose. This model is able to follow the ground truth of glucose trajectory. Based on this new glucose model and another machine learning model we perform exercise detection. We use confusion matrix to evaluate our detection accuracy, and we find that there is a trade-off between detection accuracy and time delay of detection. The exercise detection is also influenced by meals intake, which is carbs intake. Finally, we compare the performance of our model with others and we find our method outperforms than others.

关键词： Computer science   Artificial intelligence  

摘要： The development of in-memory technologies has fueled the emerging of in-memory computing systems. Simultaneously, with novel memory technologies such as high-bandwidth memory (HBM) and non-volatile memory (NVM), hybrid memory systems are expected to be more commonly used in Cloud Computing platforms, which opens a new field about memory management in both academia and industry communities. Memory capacity is always a critical bottleneck for any applications running in Cloud Computing platforms. Data explosion is also posing an unprecedented requirement for computing capacity to handle the ever-growing data volume, velocity, variety, and veracity. Thus, in-memory computing systems are increasingly looking inward at huge caches of under-processed or trash-away data as resources to be mined. The key purpose of managing data in any memory system is to keep more useful data with high memory utilization without compromising applications' performance, especially for machine and deep learning applications running in clouds. However, there are numerous challenges in realizing the above goal, including sharing memory among applications, managing cached data, data migration on hybrid memory systems, and control strategy for a unified hybrid memory pool. In this context, we concentrate on developing an efficient hybrid memory system and memory management strategies for in-memory computing on Cloud Computing platforms. To achieve this, we propose to develop a hybrid memory system that includes fast and relatively slow memory hardware and memory management strategies for applications running in cloud environments based on optimization formulations, feedback control, and machine/deep learning methods. In order to realize a runtime system that automatically optimizes data management on hybrid memory, we will (1) propose a new shared in-memory cache layer among parallel executors that are co-hosted on the same computing node, which aims to improve the overall hit rate of data blocks

关键词： Transportation   Computer science   Artificial intelligence