AI in EECS
AI in EECS
Artificial intelligence and machine learning are principal components of EECS research. AI/ML research in each research group is summarized as follows.
Intelligent Computing Systems
As the paradigm of artificial intelligence (AI) has changed from knowledge-based reasoning to data-based learning, high-performance computing systems, trainable data acquisition, and differentiated algorithms become key factors in determining AI technology competitiveness.
The importance of big data platforms and analysis technologies, such as big data collection, storage, and processing, has been highlighted to secure learnable data in the information-oriented society, and high-performance computing technologies and specialized algorithm design technologies for each field are fundamental technology to lead AI commercialization. The Intelligent Computing Systems research team at EECS focuses on those researches: big data processing/analysis, high-performance computing systems, and AI core applications such as computer vision and image/signal processing.
Connected Smart Systems
When a huge amount of data is generated from the Internet of Things (IoT), sensors, satellite, mobile and self-driving vehicles, data must be transmitted, processed and stored in distributed places such as cloud/edge servers. In addition, in order to execute deep-learning algorithms that finds an important meaning from data, it is necessary to utilize vast amounts of data distributed throughout the network, such as mobile devices, IoT devices, cloud/edge computing servers, as well as distributed computing resources. Hence, distributed data processing and distributed machine learning are essential in the era of IoT, big data, cloud computing, and connected AI. The Connected Smart System is a system for efficiently managing large-scale data and computing resources in the network for distributed data processing and distributed machine learning. Connected Smart Systems team in DGIST EECS study AI and machine learning based 5G+/6G communication/network, cloud/edge computing systems, neural network learning systems, automobile HMI and vision application in connected autonomous vehicles, big data analysis and processing via system design of communication/network.
The recent advances in the domain of Artificial Intelligence (AI) are backed up by big data, novel algorithms (SW), and powerful processors (HW). Even the tech giants around the globe, e.g. Samsung, Google, Intel, Apple, etc., have designed their own AI chips to secure a good position in the promising AI business. According to McKinsey report, ~40% of total value in AI will be occupied by the semiconductor-related businesses.
In EECS at DGIST, we put great efforts into developing new devices and fabrications, circuits, and architectures for AI. Our goal is to create innovative hardware stacks to enable efficient processing of AI algorithms.
Bio-medical systems that gather data from various physiological signals are the key ingredients in realizing the vision of smart healthcare in our society. Access to such data can enable the early diagnosis of several physical and mental diseases and thus help in efficiently managing them. Given this possibility, several world-leading companies are actively extending their big data, cloud computing platform, and their range wearable mobile devices to drive this vision towards reality.
The EECS department at DGIST is actively developing Artificial Intelligence (AI)/Machine Learning (ML) techniques and hardware systems for the bio-medical applications such as: (i) Deep learning algorithms for various medical imaging data including non-invasive ultrasound technology, (ii) ML based study of brain and its functional connectivity, (iii) AI hardware for mimicking the functionality of brain along with several neuro-prosthetics, and (iv) wearable physiological signal sensors for ML based analysis. Students at EECS have interdisciplinary research and educational exposure at the world class infrastructure we have here at DGIST.
사이버 물리 시스템 / Cyber-Physical Systems
Cyber-Physical Systems (CPS) have emerged as a promising research paradigm that bridges the cyber-world of computing and communication with the physical world. In CPS, various physical devices connected to the cyber-world and offered new prediction/learning (AI) capabilities are capable of autonomously functioning and interaction with dynamically changing environments, improving safety, efficiency, and reliability. Examples of CPS include self-driving cars, smart transportation systems, smart factories, smart homes, smart healthcare systems, and artificial intelligence robots. As one of the core technologies driving Industry 4.0, CPS are stimulating significant changes in the quality of human life.
CPS are functionally composed of a mix of computation, communication, and control processes: 1) acquire physical world information through a variety of sensors tied to the internet, 2) recognize/analyze it through artificial intelligence, and 3) apply the processed results to the physical world through controllers/actuators.