We study the problem of maximizing non-monotone submodular functions subject to a p-independence system constraint. Although the submodularity ratio has been well-studied in maximizing set functions under monotonic sc...
详细信息
The progress of Artificial Intelligence (AI) has been the most remarkable in the 21st Century. However, how far AI must have gone, it works in healthcare remains a big challenge for great minds across the globe. Epile...
详细信息
The progress of Artificial Intelligence (AI) has been the most remarkable in the 21st Century. However, how far AI must have gone, it works in healthcare remains a big challenge for great minds across the globe. Epilepsy is a neurological disorder that occurs in a human being at any stage in life. Thus, frequent to rare seizures are faced by an individual having epilepsy and sometimes lead to death. The electroencephalogram (EEG) signals help diagnose this ailment. However, often long EEG signal takes a day or more even for trained neurologists to detect this disorder and may even lead to human errors. Thus, it becomes imperative to develop a robust and computationally efficient system that can help detect this ailment without causing delays and human error. This paper aims to extract the time-frequency plane from time domain signals using Short-time Fourier transform (stFT) and develop a neural network model using CNN to classify seizures. This paper has achieved the highest classification accuracy of 100% for some of the experiments and 95-99% accuracy for a few of them, thus helping better classify health and seizures with less computation required.
Artificial intelligence will play a vital role for autonomous and cooperative driving in future intelligent transportation systems (ITS). Whereas, it needs sufficient computing resource, data resource and elaborated m...
详细信息
ISBN:
(纸本)9781665432078
Artificial intelligence will play a vital role for autonomous and cooperative driving in future intelligent transportation systems (ITS). Whereas, it needs sufficient computing resource, data resource and elaborated models to provide powerful artificial intelligence for enhanced driving and cooperating in ITS. Current schemes, such as computation offloading, distributed data caching, and remote model training, usually focus on part of this problem, and cannot substantially solve the multidimensional burden in such intelligent system. In this paper, we propose a multi-dimensional offloading (MDO) scheme to realize efficient and joint offloading of computation, data cache and decision model in ITS. Specifically, an optimization model of MDO is presented and divided into two sub-problems to get the optimized resource allocation with extremely low latency, which is crucial for a time-sensitive system. Simulations are conducted to verify the performance of our proposed scheme, and the simulation results show MDO can decrease the time latency considerably compared with traditional schemes.
High-performing human teams leverage intelligent and efficient communication and coordination strategies to collaboratively maximize their joint utility. Inspired by teaming behaviors among humans, I seek to develop c...
详细信息
ISBN:
(纸本)9781450392136
High-performing human teams leverage intelligent and efficient communication and coordination strategies to collaboratively maximize their joint utility. Inspired by teaming behaviors among humans, I seek to develop computational methods for synthesizing intelligent communication and coordination strategies for collaborative multi-robot systems. I leverage both classical model-based control and planning approaches as well as data-driven methods such as Multi-Agent Reinforcement Learning (MARL) to provide several contributions towards enabling emergent cooperative teaming behavior across both homogeneous and heterogeneous (including agents with different capabilities) robot teams. In future work, I aim to investigate efficient ways to incorporate humans' teaming strategies for robot teams and directly learn team coordination policies from human experts.
Locality Preserving Projection (LPP) is a dimensional reduction method that has been widely used in various fields. While traditional LPP only uses a single projection matrix to reduce the dimension and preserve the l...
详细信息
Since the beginning of the 21st century, in training, life and work, people have become more and more aware of the increasing frequency of communication between different languages. Whether they are idiosyncrasies or ...
详细信息
Since the beginning of the 21st century, in training, life and work, people have become more and more aware of the increasing frequency of communication between different languages. Whether they are idiosyncrasies or forms of communication, there are rigorous issues and increasing demands on the accuracy of translation communication. The main purpose of this paper is to design and research a system for intelligent recognition of English translation based on machine learning algorithms. Rapid advances in computing power, growth and adoption of the Internet, and multilingual knowledge bases in both countries and the United Nations have given us millions of bilingual institutions. More and more researchers are devoted to computational engineering training with success. Experiments show that the NP-sequencing pattern library plays the most important role, and its translation results are improved by 3 percentage points compared to the baseline system Moses. And about 2876 English terms in the whole test corpus can find the same syntactic structure in the pattern library, accounting for about 88.7% of the test corpus.
It could be of great difficulty and cost to directly apply complex deep neural network to mobile devices with limited computing and endurance abilities. This paper aims to solve such problem through improving the comp...
详细信息
As AI technology advances and its use increases, efforts to incorporate machine learning for malware detection are increasing. However, for malware learning, a standardized data set is required. Because malware is uns...
详细信息
As AI technology advances and its use increases, efforts to incorporate machine learning for malware detection are increasing. However, for malware learning, a standardized data set is required. Because malware is unstructured data, it cannot be directly learned. In order to solve this problem, many studies have attempted to convert unstructured data into structured data. In this study, the features and limitations of each were analyzed by investigating and analyzing the method of converting unstructured data proposed in each study into structured data. As a result, most of the data conversion techniques suggest conversion mechanisms, but the scope of each technique has not been determined. The resulting data set is not suitable for use as training data because it has infinite properties.
Today’s global processes of post-industrial society and digital economy development are closely related engineering education. Vigorous growth of technology and communications tools implies mobility and reducing nati...
详细信息
With the advancement of science and technology in recent years, the era of big data has gradually matured, and the era of intelligence has gradually penetrated into ordinary life. The improvement of technology around ...
详细信息
暂无评论