본문 바로가기
대메뉴 바로가기
KAIST
Newsletter Vol.25
Receive KAIST news by email!
View
Subscribe
Close
Type your e-mail address here.
Subscribe
Close
KAIST
NEWS
유틸열기
홈페이지 통합검색
-
검색
KOREAN
메뉴 열기
signing
by recently order
by view order
Deep Learning Framework to Enable Material Design in Unseen Domain
Researchers propose a deep neural network-based forward design space exploration using active transfer learning and data augmentation A new study proposed a deep neural network-based forward design approach that enables an efficient search for superior materials far beyond the domain of the initial training set. This approach compensates for the weak predictive power of neural networks on an unseen domain through gradual updates of the neural network with active transfer learning and data augmentation methods. Professor Seungwha Ryu believes that this study will help address a variety of optimization problems that have an astronomical number of possible design configurations. For the grid composite optimization problem, the proposed framework was able to provide excellent designs close to the global optima, even with the addition of a very small dataset corresponding to less than 0.5% of the initial training data-set size. This study was reported in npj Computational Materials last month. “We wanted to mitigate the limitation of the neural network, weak predictive power beyond the training set domain for the material or structure design,” said Professor Ryu from the Department of Mechanical Engineering. Neural network-based generative models have been actively investigated as an inverse design method for finding novel materials in a vast design space. However, the applicability of conventional generative models is limited because they cannot access data outside the range of training sets. Advanced generative models that were devised to overcome this limitation also suffer from weak predictive power for the unseen domain. Professor Ryu’s team, in collaboration with researchers from Professor Grace Gu’s group at UC Berkeley, devised a design method that simultaneously expands the domain using the strong predictive power of a deep neural network and searches for the optimal design by repetitively performing three key steps. First, it searches for few candidates with improved properties located close to the training set via genetic algorithms, by mixing superior designs within the training set. Then, it checks to see if the candidates really have improved properties, and expands the training set by duplicating the validated designs via a data augmentation method. Finally, they can expand the reliable prediction domain by updating the neural network with the new superior designs via transfer learning. Because the expansion proceeds along relatively narrow but correct routes toward the optimal design (depicted in the schematic of Fig. 1), the framework enables an efficient search. As a data-hungry method, a deep neural network model tends to have reliable predictive power only within and near the domain of the training set. When the optimal configuration of materials and structures lies far beyond the initial training set, which frequently is the case, neural network-based design methods suffer from weak predictive power and become inefficient. Researchers expect that the framework will be applicable for a wide range of optimization problems in other science and engineering disciplines with astronomically large design space, because it provides an efficient way of gradually expanding the reliable prediction domain toward the target design while avoiding the risk of being stuck in local minima. Especially, being a less-data-hungry method, design problems in which data generation is time-consuming and expensive will benefit most from this new framework. The research team is currently applying the optimization framework for the design task of metamaterial structures, segmented thermoelectric generators, and optimal sensor distributions. “From these sets of on-going studies, we expect to better recognize the pros and cons, and the potential of the suggested algorithm. Ultimately, we want to devise more efficient machine learning-based design approaches,” explained Professor Ryu.This study was funded by the National Research Foundation of Korea and the KAIST Global Singularity Research Project. -Publication Yongtae Kim, Youngsoo, Charles Yang, Kundo Park, Grace X. Gu, and Seunghwa Ryu, “Deep learning framework for material design space exploration using active transfer learning and data augmentation,” npj Computational Materials (https://doi.org/10.1038/s41524-021-00609-2) -Profile Professor Seunghwa Ryu Mechanics & Materials Modeling Lab Department of Mechanical Engineering KAIST
2021.09.29
View 9073
KAIST's efforts begin to become the first Korean university establishing a "strategic technology management system."
KAIST completed the signing of business agreement with the Korea Strategic Trade Institute to establish a strategic technology management system on the 22nd of September. The agreement between KAIST and Korea Strategic Trade Institute (under the Ministry of Knowledge Economy) encompasses 1) the establishment of processes for strategic technology management on campus, 2) development and accommodation of management system on par with major countries, and 3) protection and management through continued education and promotion. Strategic technology management is necessary to prevent the illegal distribution of technologies developed in Korea to those countries and organizations of concern. The need for the management system arose due to the fact that technology transfer has become venerable to illegal export of strategic technologies. The agreement between the two parties offer protection to KAIST when exporting strategic technologies as it necessitates the permission of the government prior to the technology transfer.
2011.09.27
View 7637
New Korean Robot Responds to Non-Verbal Commands
An online newspaper covering the latest headline news based on MS Windows, Internet and technology trends, Infopackets, posted an article on the research result of KAIST researchers: a robot prototype developed, which is able to respond to human’s non verbal communication. Robots now can read human gestures and react to their subtle commands by “designing the robot’s main system to mimic the actions of a human in the same manner in which human brains function.” For details of the article, please click the link: http://www.infopackets.com/news/technology/science/2011/20110125_new_korean_robot_responds_to_non_verbal_commands.htm
2011.01.26
View 8340
KAIST Undergraduates Open Four-Day International Conference
The ICISTS-KAIST, an annual international conference organized by KAIST undergraduate students, opened on Thursday (Aug. 20) at the KAIST"s main campus in Daejeon. The 2009 ICISTS (International Conference for the Integration of Science and Technology into Society) drew around 200 experts and students from 44 countries. Since its inception in 2005 to promote discourse on important science and technology issues affecting modern society, the conference has served as an opportunity for academic networking among students in various parts of the world. The four-day conference consists of lectures, open discussions among lecturers and students, field trips to help students to understand actual applications, and team projects. This year"s conference offers three workshops under the themes of "Climate Change: Merging Technology and Policy for Green Solutions"; "Human-Computer Interaction: Designing Computer System for Intuitive Human Access"; and "Nano Clinic: Breakthrough in Conquering Disease." Lectures by invited experts in various scientific fields will help broaden students" perspectives particularly from interdisciplinary viewpoints, said an organizer of the conference.
2009.08.28
View 13310
KAIST, Hynix Agree to Cooperate in Fostering Skilled Manpower
KAIST and Hynix Semiconductor, the world’s second largest producer of dynamic random access memory (DRAM) chips, have agreed to promote bilateral cooperation in fostering highly skilled manpower for the semiconductor industry. A signing ceremony was held on Jan. 21 in Seoul. The agreement marked an expansion of the scope of cooperation between the two organizations into the system IC industry. Since 1995, KAIST and Hynix have cooperated in fostering human resource specialized in the memory semiconductor area, bringing up a total of 250 highly skilled personnel in the area so far. Under the new agreement, Hynix will provide financial support, including scholarships, to KAIST for the next five years. The number of students subject to the Hynix-financed program will be increased to 20 a year from the current 10. New material engineering and physics will be added to the areas covered by the cooperation program.
2008.01.29
View 10704
<<
첫번째페이지
<
이전 페이지
1
>
다음 페이지
>>
마지막 페이지 1