CO
-
Team KAIST to Race at CES 2022 Autonomous Challenge
Five top university autonomous racing teams will compete in a head-to-head passing competition in Las Vegas
A self-driving racing team from the KAIST Unmanned System Research Group (USRG) advised by Professor Hyunchul Shim will compete at the Autonomous Challenge at the Consumer Electronic Show (CES) on January 7, 2022. The head-to-head, high speed autonomous racecar passing competition at the Las Vegas Motor Speedway will feature the finalists and semifinalists from the Indy Autonomous Challenge in October of this year. Team KAIST qualified as a semifinalist at the Indy Autonomous Challenge and will join four other university teams including the winner of the competition, Technische Universität München.
Team KAIST’s AV-21 vehicle is capable of driving on its own at more than 200km/h will be expected to show a speed of more than 300 km/h at the race.The participating teams are:1. KAIST2. EuroRacing : University of Modena and Reggio Emilia (Italy), University of Pisa (Italy), ETH Zürich (Switzerland), Polish Academy of Sciences (Poland) 3. MIT-PITT-RW, Massachusetts Institute of Technology, University of Pittsburgh, Rochester Institute of Technology, University of Waterloo (Canada)4.PoliMOVE – Politecnico di Milano (Italy), University of Alabama 5.TUM Autonomous Motorsport – Technische Universität München (Germany)
Professor Shim’s team is dedicated to the development and validation of cutting edge technologies for highly autonomous vehicles. In recognition of his pioneering research in unmanned system technologies, Professor Shim was honored with the Grand Prize of the Minister of Science and ICT on December 9.
“We began autonomous vehicle research in 2009 when we signed up for Hyundai Motor Company’s Autonomous Driving Challenge. For this, we developed a complete set of in-house technologies such as low-level vehicle control, perception, localization, and decision making.” In 2019, the team came in third place in the Challenge and they finally won this year.
For years, his team has participated in many unmanned systems challenges at home and abroad, gaining recognition around the world. The team won the inaugural 2016 IROS autonomous drone racing and placed second in the 2018 IROS Autonomous Drone Racing Competition. They also competed in 2017 MBZIRC, ranking fourth in Missions 2 and 3, and fifth in the Grand Challenge.
Most recently, the team won the first round of Lockheed Martin’s Alpha Pilot AI Drone Innovation Challenge. The team is now participating in the DARPA Subterranean Challenge as a member of Team CoSTAR with NASA JPL, MIT, and Caltech.
“We have accumulated plenty of first-hand experience developing autonomous vehicles with the support of domestic companies such as Hyundai Motor Company, Samsung, LG, and NAVER. In 2017, the autonomous vehicle platform “EureCar” that we developed in-house was authorized by the Korean government to lawfully conduct autonomous driving experiment on public roads,” said Professor Shim.
The team has developed various key technologies and algorithms related to unmanned systems that can be categorized into three major components: perception, planning, and control. Considering the characteristics of the algorithms that make up each module, their technology operates using a distributed computing system.
Since 2015, the team has been actively using deep learning algorithms in the form of perception subsystems. Contextual information extracted from multi-modal sensory data gathered via cameras, lidar, radar, GPS, IMU, etc. is forwarded to the planning subsystem. The planning module is responsible for the decision making and planning required for autonomous driving such as lane change determination and trajectory planning, emergency stops, and velocity command generation. The results from the planner are fed into the controller to follow the planned high-level command. The team has also developed and verified the possibility of an end-to-end deep learning based autonomous driving approach that replaces a complex system with one single AI network.
2021.12.17 View 8341 -
KAIST Plans to Open a New York Campus
President Lee signs an MOU with New York-based Big Continent Inc. Chairman Hee-Nam Bae on funding the New York campus
President Kwang Hyung Lee announced a plan to open a KAIST campus in New York with funding from New York-based entrepreneur Hee-Nam Bae. President Lee and Big Continent Inc. Chairman Hee-Nam Bae signed the MOU last week for the funding to open the campus in New York.
President Lee said it will take years to open up a campus in New York in order to conform with both Korean and US legal procedures. However, during a news conference in New York following the signing of the MOU with Chairman Bae, President Lee said this is the first step toward realizing KAIST’s new vision of a ‘Global Twin Strategy’ by making New York KAIST’s newest stronghold to target both domestic and global markets.
“New York is the center of the world’s commerce, culture, and new technologies. If we want to grow big, we should go to one of the biggest cities in the world and New York is the place. I highly encourage our students and faculty go into the world and never be satisfied enjoying the top position in Korea. The next place to investigate will be Silicon Valley,” said President Lee.
“We still have many issues to resolve domestically. We need to discuss more details first with the Board of Trustees and the Korean government,” he added.
The New York campus will aim to become an enterprise-type university to help KAIST create global value. "Our goal is to make sure that Korean businesses gain competitiveness in the global market and can become listed on the NASDAQ. We plan to open majors related to AI, financial engineering, and cultural technologies. We will recruit students from both the US and KAIST to study at our New York campus.” President Lee said.
Chairman Bae, a self-made entrepreneur who immigrated to the US in 1981, also leads the Global Leadership Foundation in the US. “President Lee and I have already toured several candidate sites for the campus in the New York region and we will make a final decision on the best site to purchase,” said Chairman Bae.
Chairman Bae added that he has always dreamed of fostering young global talents who will take on global challenges with pioneering minds. He believes KAIST shares this global vision.
The New York campus will be the first KAIST campus for global students funded by someone from the private sector. This is also a major step forward for KAIST, which was founded by a six million dollar USAID loan in 1971. KAIST announced its plans to establish Kenya KAIST in 2018 with funding from the Korea Eximbank’s 95 million USD development cooperation fund loan to the Kenyan government. KAIST will provide turn-key-based education consultancy featuring curriculum design and the construction of facilities for Kenya’s first advanced science and technology institute. The campus will be located in the Konza Techno City near Nairobi and plans to open in 2023.
2021.12.13 View 6947 -
Startup Elice Donates 300 Million KRW to School of Computing
Elice hopes to create a virtuous circle that bridges the educational gap
Elice, a student startup from the School of Computing has committed to donate 300 million KRW to KAIST. Jae-Won Kim, CEO of the coding education company, established the startup with his colleagues in 2015. Since then, more than 100 companies, including 17 of Korea’s top 20 companies such as SK and LG have used Elice' digital coding platform to educate employees. More than 200,000 employees have completed the online training with completion rates over 80%.
Kim said during the donation ceremony that he hopes to fund the renovation of the School of Computing building and that he will continue to work on expanding platforms that will help make communication between educators and students more interactive. He explained, “We are making this contribution to create a virtuous circle that bridges the educational gap and improves the quality of education."
President Kwang Hyung Lee was pleased to welcome the student startup’s donation, saying, "Software talent is one of the most precious resources we should foster for the nation’s future. I am thrilled to see that a startup that was founded here on the KAIST campus has grown into a great company that provides excellent coding education for our society.”
Professor Alice Oh, who was the advisor for Kim and his colleagues when they launched the startup, joined the ceremony along with the founding members from KAIST including CPO Su-In Kim, CTO Chong-Kuk Park, and team leader Chang-Hyun Lee.
2021.12.13 View 3890 -
A Team of Three PhD Candidates Wins the Korea Semiconductor Design Contest
“We felt a sense of responsibility to help the nation advance its semiconductor design technology”
A CMOS (complementary metal-oxide semiconductor)-based “ultra-low noise signal chip” for 6G communications designed by three PhD candidates at the KAIST School of Electrical Engineering won the Presidential Award at the 22nd Korea Semiconductor Design Contest.
The winners are PhD candidates Sun-Eui Park, Yoon-Seo Cho, and Ju-Eun Bang from the Integrated Circuits and System Lab run by Professor Jaehyouk Choi. The contest, which is hosted by the Ministry of Trade, Industry and Energy and the Korea Semiconductors Industry Association, is one of the top national semiconductor design contests for college students.
Park said the team felt a sense of responsibility to help advance semiconductor design technology in Korea when deciding to participate the contest. The team expressed deep gratitude to Professor Choi for guiding their research on 6G communications.
“Our colleagues from other labs and seniors who already graduated helped us a great deal, so we owe them a lot,” explained Park. Cho added that their hard work finally got recognized and that acknowledgement pushes her to move forward with her research. Meanwhile, Bang said she is delighted to see that many people seem to be interested in her research topic.
Research for 6G is attempting to reach 1 tera bps (Tbps), 50 times faster than 5G communications with transmission speeds of up to 20 gigabytes. In general, the wider the communication frequency band, the higher the data transmission speed. Thus, the use of frequency bands above 100 gigahertz is essential for delivering high data transmission speeds for 6G communications.
However, it remains a big challenge to make a precise benchmark signal that can be used as a carrier wave in a high frequency band. Despite the advantages of CMOS’s ultra-small and low-power design, it still has limitations at high frequency bands and its operating frequency. Thus, it was difficult to achieve a frequency band above 100 gigahertz.
To overcome these challenges, the three students introduced ultra-low noise signal generation technology that can support high-order modulation technologies. This technology is expected to contribute to increasing the price competitiveness and density of 6G communication chips that will be used in the future.
5G just got started in 2020 and still has long way to go for full commercialization. Nevertheless, many researchers have started preparing for 6G technology, targeting 2030 since a new cellular communication appears in every other decade.
Professor Choi said, “Generating ultra-high frequency signals in bands above 100 GHz with highly accurate timing is one of the key technologies for implementing 6G communication hardware. Our research is significant for the development of the world’s first semiconductor chip that will use the CMOS process to achieve noise performance of less than 80fs in a frequency band above 100 GHz.”
The team members plan to work as circuit designers in Korean semiconductor companies after graduation. “We will continue to research the development of signal generators on the topic of award-winning 6G. We would like to continue our research on high-speed circuit designs such as ultra-fast analog-to-digital converters,” Park added.
2021.11.30 View 6827 -
Renault 5 EV and Canoo’s Pickup Truck Win the 2021 FMOTY Awards
KAIST Future Mobility of the Year Awards recognize the most innovative concept cars of the year
The Renault 5 EV from France and a pickup truck from the US startup Canoo won the 2021 Future Mobility of the Year Awards (FMOTY) hosted by the Cho Chun Shik Graduate School of Green Transportation at KAIST. The awards ceremony was held at Renault Samsung Motors in Seoul on November 25.
KAIST began the FMOTY in 2019 to advance future car technology and stimulate growth in the industry. The award recognizes the most innovative ideas for making the most futuristic concept car and improving the technological and social value of the industry.
The awards ceremony was attended by KAIST President Kwang Hyung Lee, the dean of the Cho Chun Shik Graduate School of Green Transportation In Gwun Jang, CEO of Renault Samsung Motors Dominique Signora, and CEO of Canoo Tony Aquila. President Lee said, “The new world order will be impacted by new technology developers who envision the future. Their innovation and creative ideas will open a new world of sustainable future transportation.”
Out of the 46 concept cars revealed at global motor exhibitions between last year and the first quarter of this year, models demonstrating transport technology useful for future society and innovative service were selected in the categories of passenger cars and commercial vehicles.
Sixteen automotive journalists from 11 countries, including the chief editor of Car Magazine in Germany Georg Kacher and editorial director of BBC Top Gear Charlie Turner, participated as judges.
This year’s award for the best concept car for a passenger vehicle went to an electric vehicle, the Renault 5 EV. The compact electric car was highly regarded for its practicality and environmental friendliness. A pickup truck by Canoo, an American EV manufacturing start-up, won the award in the commercial vehicle category. The pickup features an innovative design allowing for a variety of functions topped with a competitive price and it received overwhelming support from the judges.
While Hyundai Motors swept both prizes at the awards last year and demonstrated the potential of Korean concept cars, Canoo’s win in the commercial vehicle section as a young American venture company brought attention to the changing dynamics in the automotive market. This shows that young EV start-ups can compete with existing car companies as the automotive paradigm is shifting from those with internal combustion engines to EVs.
The awards organizers said that the Cho Chun Shik Graduate School of Green Transportation will continue to hold the FMOTY to lead the fast-changing global mobility market. For more information, please visit www.fmoty.org.
2021.11.26 View 5791 -
Professor Sung-Ju Lee’s Team Wins the Best Paper and the Methods Recognition Awards at the ACM CSCW
A research team led by Professor Sung-Ju Lee at the School of Electrical Engineering won the Best Paper Award and the Methods Recognition Award from ACM CSCW (International Conference on Computer-Supported Cooperative Work and Social Computing) 2021 for their paper “Reflect, not Regret: Understanding Regretful Smartphone Use with App Feature-Level Analysis”.
Founded in 1986, CSCW has been a premier conference on HCI (Human Computer Interaction) and Social Computing. This year, 340 full papers were presented and the best paper awards are given to the top 1% papers of the submitted. Methods Recognition, which is a new award, is given “for strong examples of work that includes well developed, explained, or implemented methods, and methodological innovation.”
Hyunsung Cho (KAIST alumus and currently a PhD candidate at Carnegie Mellon University), Daeun Choi (KAIST undergraduate researcher), Donghwi Kim (KAIST PhD Candidate), Wan Ju Kang (KAIST PhD Candidate), and Professor Eun Kyoung Choe (University of Maryland and KAIST alumna) collaborated on this research.
The authors developed a tool that tracks and analyzes which features of a mobile app (e.g., Instagram’s following post, following story, recommended post, post upload, direct messaging, etc.) are in use based on a smartphone’s User Interface (UI) layout. Utilizing this novel method, the authors revealed which feature usage patterns result in regretful smartphone use.
Professor Lee said, “Although many people enjoy the benefits of smartphones, issues have emerged from the overuse of smartphones. With this feature level analysis, users can reflect on their smartphone usage based on finer grained analysis and this could contribute to digital wellbeing.”
2021.11.22 View 5446 -
Seegene Opens Covid-19 Testing Mobile Station on Campus
Seegene donates testing reagents for 40,000 people with results available in three hours
Seegene, a molecular diagnostic testing company, donated enough testing reagents for 40,000 COVID-19 tests for the KAIST community and set up a mobile testing station run by the Seegene Medical Foundation on October 28.
The entire COVID-19 diagnosis process, including specimen collection, PCR testing, and results analysis, can be conducted at the mobile testing unit developed by Seegene. The on-site testing station will help the campus get ready to return to normal, especially as the government is transitioning toward its ‘living with Covid-19’ policy, which eases a range of social distancing restrictions. Any KAIST community member can get a Covid-19 test on campus and receive the results within three hours. The station can conduct up to 7,500 tests per day.
This is an extension of the agreement between KAIST and Seegene made in July for research collaboration. The two institutions will work together on various research projects including ultrafast PCR testing, sample collection, and cloud-based data transmission and analysis.
Prior to this donation, according to an administrative order from Daejeon City, KAIST opened a temporary COVID-19 testing center in collaboration with Seegene and conducted COVID-19 tests at the KAIST Clinic over four days starting from September 28. All students living on campus were tested, and all 2,775 tested negative.
Seegene CEO Jong-Yoon Chun said, "KAIST and Seegene signed an agreement for collaborative research on molecular diagnosis in July prior to this donation, and we are happy to maintain a connection with KAIST.” He added, “We hope that this donation will help students return to their ordinary university lives.”
Vice President for Planning and Budget Bowon Kim said, "As KAIST is currently planning to conduct offline lectures in preparation for ‘living with COVID-19’, Seegene’s donation will be particularly helpful.” He added, “The two institutions will continue to cooperate, leading to not only the short-term stabilization of the campus, but also collaborative research for the vitalization of molecular diagnosis technology and the bio industry.”
2021.11.03 View 4776 -
Minister of Justice Meets with KAIST International Community Members
The Office of Immigration Service in Daejeon will dispatch On-Site Immigration Service Team at the campus once a week
Minister of Justice Beom Kye Park met with international students and faculty members on October 29 at the KAIST campus and promised the government’s prompt and flexible revision of the process for acquiring legal residency and Korean citizenship for talents in the fields of science and technology.
During the meeting to discuss immigration difficulties with foreign students, researchers, and faculty at KAIST, many KAIST international students expressed their wishes to continue their research in Korea after graduation and asked for legal support to acquire permanent residence status.
International faculty members including PhD and Master’s candidate and postdoc fellows attended the meeting along with KAIST President Kwang Hyung Lee and Assistant Vice President of the International Office Scott Knowles.
Currently, there are 1,100 international members on campus: 421 undergraduates, 236 Master’s student, 266 PhD candidates, 79 researchers, and 67 faculty members.
President Lee said, “It is prerequisite to nurture the outstanding talents who earned their degrees in Korea for raising our national competitiveness. We would like to ask the government to ease the current system to embrace those excellent talents. That will definitely be necessary for securing new talents as well as for invigorating the domestic industry and R&D sector, which will lead to attracting the next excellent groups of talented students from abroad.”
Minister Park said that the government now needs more inclusive immigration policies granting legal residency and citizenship to the highly talented group. He added that the ministry will make every effort to help our degree holders acquire the relevant legal status to settle down here.
Meanwhile, the Office of Immigration Services in Daejeon set up the ‘On-Site Immigration Service’ at the campus and provided one-on-one consultation services for KAIST international community regarding extension of stays and alien registration affairs.
The On-Site Immigration Service will continue at the campus once a week for convenience of KAIST international community in the very near future.
2021.10.29 View 3876 -
Diva Sumi Jo to Join the KAIST Faculty
Visiting Distinguished Professor Jo will enrich KAIST’s scholarship and inspire futuristic art and technology research
Soprano Sumi Jo will join the KAIST faculty from the spring 2022 semester. Named as a visiting distinguished professor in the Graduate School of Culture Technology, she will give special leadership lectures. Her tenure will be through September 2024.
Jo joined the appointment ceremony held online at KAIST on October 14 from Portugal and expressed her high expectations for teaching KAIST students from next year. “I am very grateful for this opportunity to meet students at KAIST, the birthplace of advanced science and technology in Korea,” she said.
KAIST President Kwang Hyung Lee, who has stressed the importance of humanities and the arts in convergence studies of science and technology, lauded her joining the faculty as a big asset who will enrich KAIST’s scholarship. “Soprano Sumi Jo rose to stardom on the global music stage with her unrivaled talent and effort. I truly believe her experience and passion will inspire our students to expand their horizon of thought and knowledge,” said President Lee.
Distinguished Professor Jo will also participate in convergence research at the Graduate School of Culture Technology with KAIST professors and many other experts. The Sumi Jo Performing Arts Research Center at the Graduate School of Culture Technology will conduct research on the converging of imaging and audio processing technologies that will enhance virtual artists’ performances. Distinguished Professor Jo explained, “The world is changing so fast. I look forward to working on culture technology research at KAIST that will raise our life quality.”
Professor Juhan Nam from the Graduate School of Culture Technology said, “We look forward to working closely with her and her team to develop research themes that envision futuristic art combined with technology such as the metaverse and non-fungible tokens (NFTs).
Coloratura soprano Jo was born in Seoul and educated at Seoul National University and the Conservatorio Santa Cecilia in Italy. Among her teachers were Carolo Bergonzi and Giasnnelas Borelli. Following her graduation from the Conservatorio Santa Cecilia in 1985, she swept major international competitions in Seoul and Europe. In 1986, she was unanimously awarded the first prize in the Carlo Alberto Cappelli International Competition in Verona which is open only to the first-prize winners of major competitions.
Since her debut in the role of Gilda in Verdi’s Rigolleto in Italy in 1986, she has performed on the world's biggest stages along with noted maestros such as Herbert von Karajan, Georg Solti, Zubin Mehta, and James Levine. Distinguished Professor Jo, one of the most sought-after sopranos in the world, released more than 40 albums.
2021.10.15 View 4936 -
KAIST Entrepreneurial Partnership to Accelerate Startups and Venture Ecosystem
KAIST will launch the KAIST Entrepreneurial Partnership (KEP) program, which connects faculty members who own technology with those who want to launch startup.
The program encourages open innovation startups using strategies tailored to market-client demand requirements. This is also one of efforts to help realize ‘one startup per lab,’ initiated by President Kwang Hyung Lee’s new innovation strategy. KEP also aims to introduce the best technologies developing at KAIST to startups and to raise the success rate of technology commercialization.
The program will match KAIST faculty and student entrepreneur candidates with parties enrolled in the new Entrepreneur in Residence and Entrepreneurial Partner programs. Each team will be given a six-month test period with funding support.
KAIST will invite entrepreneurial experts from both technology and management fields to support the program participants. Around 30 experts with experience in developing new businesses, startups, and investments in large corporations or venture companies will be recruited as entrepreneurial partners. They will offer support for research and business development (R&BD), technology marketing, attracting venture investment from corporations, mergers and acquisitions, and business openings.
A survey showed that KAIST members who are interested in starting a business are experiencing difficulties finding an entrepreneurial expert (72.2%), with the complicated startup approval procedures (33.3%), and their lack of knowledge on entrepreneurship and funding (27.8%). The KEP program hopes to encourage KAIST faculty members and students who have well-developed business ideas and the appropriate technology but lack the capabilities to realize and develop them into a business.
Associate Vice President of Startups Young-Tae Kim said, “We will develop KEP into KAIST’s distinct entrepreneurial support system and produce exemplary outcomes of faculty and student startups. We will spread the startup DNA and lead the building of a virtuous cycle between entrepreneurship and the venture ecosystem.”
2021.10.14 View 4710 -
Industrial Liaison Program to Provide Comprehensive Consultation Services
The ILP’s one-stop solutions target all industrial sectors including conglomerates, small and medium-sized enterprises, venture companies, venture capital (VC) firms, and government-affiliated organizations.
The Industrial Liaison Center at KAIST launched the Industrial Liaison Program (ILP) on September 28, an industry-academic cooperation project to provide comprehensive solutions to industry partners. The Industrial Liaison Center will recruit member companies for this service every year, targeting all industrial sectors including conglomerates, small and medium-sized enterprises, venture companies, venture capital (VC) firms, and government-affiliated organizations.
The program plans to build a one-stop support system that can systematically share and use excellent resource information from KAIST’s research teams, R&D achievements, and infrastructure to provide member companies with much-needed services.
More than 40 KAIST professors with abundant academic-industrial collaboration experience will participate in the program. Experts from various fields with different points of view and experiences will jointly provide solutions to ILP member companies. To actively participate in academic-industrial liaisons and joint consultations, KAIST assigned 10 professors from related fields as program directors.
The program directors will come from four different fields including AI/robots (Professor Alice Oh, School from the School of Computing, Professor Young Jae Jang from the Department of Industrial & Systems Engineering, and Professor Yong-Hwa Park from Department of Mechanical Engineering), bio/medicine (Professor Daesoo Kim from Department of Biological Sciences and Professor YongKeun Park from Department of Physics), materials/electronics (Professor Sang Ouk Kim from the Department of Materials Science and Engineering and Professors Jun-Bo Yoon and Seonghwan Cho from the School of Electrical Engineering), and environment/energy (Professor Hee-Tak Kim from the Department of Biological Sciences and Professor Hoon Sohn from the Department of Civil and Environmental Engineering).
The transdisciplinary board of consulting professors that will lead technology innovation is composed of 30 professors including Professor Min-Soo Kim (School of Computing, AI), Professor Chan Hyuk Kim (Department of Biological Sciences, medicine), Professor Hae-Won Park (Department of Mechanical Engineering, robots), Professor Changho Suh (School of Electrical Engineering, electronics), Professor Haeshin Lee (Department of Chemistry, bio), Professor Il-Doo Kim (Department of Materials Science and Engineering, materials), Professor HyeJin Kim (School of Business Technology and Management), and Professor Byoung Pil Kim (School of Business Technology and Management, technology law)
The Head of the Industrial Liaison Center who is also in charge of the program, Professor Keon Jae Lee, said, “In a science and technology-oriented generation where technological supremacy determines national power, it is indispensable to build a new platform upon which innovative academic-industrial cooperation can be pushed forward in the fields of joint consultation, the development of academic-industrial projects, and the foundation of new industries.
He added, “KAIST professors carry out world-class research in many different fields and faculty members can come together through the ILP to communicate with representatives from industry to improve their corporations’ global competitiveness and further contribute to our nation’s interests by cultivating strong small enterprises
2021.09.30 View 5563 -
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 9331