Out of the Lab, a Scientist Dig out the Merit of AI.
| Summary | Dr. Shou-De Lin, One of Principal Investigators of the Ministry of ScienceTechnology AI Innovation Research Center Programbr / br / Title: Out of the Lab, a Scientist Dig out the Merit of AI.br / br / Quote:br / “It is worth giving up some things because of dream pursuing” Professor SHOU-DE, LIN at the department of computer scienceInformation Engineering in National Taiwan University, Chief Machine Learning Scientist in Appier, said “An escape from comfort zone to seek new challenges makes my life become more colorful.”br / br / Content:br / br / Given qualified for being as the freshman of National Taiwan University College of Medicine, Professor Lin chose the department of electrical engineering in NTU as the first priority in Joint College Entrance Examination (JCEE). Though the undergraduate education did not cultivate him the passion on the field of electrical engineering, Professor Lin said, however, he was still recommended for further study at the graduate institute of electronics engineering in NTU due to his talentsoutstanding academic performance.br / br / Put down the present positioncollet myself to be a scholar again. Then, committed to the research on AIbr / br / Professor Lin has the faith on “Don’t waste your lives drawing lines,” so he decided to give up studying the graduate schoolchose to join the military service first instead. He took advantage of onea half years in the military service for further thought on his future at the north of Taiwan - Donyin. Professor Lin tried to kill time by reading through all of the books at the island library. He was inspired much by several Artificial intelligencecomplex science books, among of which are significant for him to be aware that computer computing would influence a lot on the industry, the business activities, the governmental policies,the mankind’s future. One of the important models who set an excellent example for him is the founder of Algorithm, John Holland.br / br / After completing the compulsory military service, Professor Lin straightly committed on AI research at the University of Michigan, where John Holland was there. Professor Lin said “No one agreed with my decision about the shift of professional research field as the semiconductorcommunication technology was the main stream then.” Due to the lankness of softwareAI related proficiency, there were many things for him to study from the very beginning. Besides, he also needed to tackle with the cultural shocklanguage barrier to equip him to get accustomed to the new research areanew living environment.br / br / Professor Lin said “People would tend to use the strength of their own knowledgethe skills to compensate the language weaknessthe culture difference upon surrounded by non-mother tongue environment. However, the first two years in the U.S.A. was the most frustrated time as being an AI beginner like me because my opinions were always ignored.” br / br / In such circumstances, it is the first time for him with the feeling disregardedthis is not easy for him to overcome because he never suffered any kind of this difficulty within his learning journey from childhood till then. “However, I persist in the passionate on AI regardless no matter what kind of resistance encountered.” Professor Lin said. He failed in the application of PhD program of the department of AI in UM. He needed to make decision on whether directly entering into the job market in the U.S. with his diploma of M.E.E.shift to apply for the PhD program on AI in somewhere else. He chose the latter resolutely to continue the research on AI. He left UM for ISI (information science institute) in UCLA as an research assistant without pay though he was in poor economic condition. At that time, he can only afford $10-dollar daily rent for a sofa bed at his senior classmate’s home. After a period of time, he moved to an apartment with a very low rent. Originally, he planned to find someone to share the cost however, he could not find anyone. Given that he knew the reason was due to the high crime rate at the area afterwards, he had no choice in the hope to settle down to continue his research asap. Therefore, he kept staying there.br / br / Professor Lin also realized that English writing competency was essential to survive for being as a career at the research community. In order to enhancepolish the competency, he attended five different English writing classes at the first three years in the U.S. In order to avoid the embarrassment of receiving the order to rewrite all from his advisor again, he studied very hard in the hope to learn how to write a clearconcise essay. Contributed to hardworking, he can host an English writing class in NTU afterwards. He said “No pains no gains. Writing competency is related with working hard to keep practice not with talents.”br / br / The turning point for better happened when there was only thousands of his bank account deposit left. Being as a volunteer research assistant, he gained highly recognition from the laboratory host, who not only helped him to get the permission to study Ph.D. but also granted him a research assistant vacancy with pay. “Compared with the other many students received the scholarship smoothly, I went through a tough process to get the grant.” Furthermore, he won the best paper award among more than three hundred submissions in WI International Conference on Web Intelligence. This paper also was the first paper for him to be as the lead author.br / br / Participating in the U.S. National Security Program to locate terrorists by the use of AIbr / br / "A person should be settled in the most appropriate place to realize his full potential," he said. During his learning journey of the PhD program at the University of South Carolina, Lin was proficient in scholarly works. He craved the data mining research with the focus on how to spot abnormalsuspicious instances in the huge social network. Prior to receiving the PhD degree, he got recognition by the US Los Alamos National Laboratory, where the atomic bomb used during World War II was produced. He was invited to participate in a project to use data mining to find out terrorists. At that time, he was one of the few non-U.S. citizens who could enter the National Laboratory to participate in important national security projects.br / br / Getting an ideal job offer in the U.S., where there are many talented people, is generally regarded as making a dream come trueleading to a bright future. However, what Lin desired most at that time was to purse academic freedomdo researches he wanted. Based on the personal belief that “it is worth giving up what he had in the pursuit of what he wanted,” Lin chose not to set limits on himself,accordingly to quit the job of the National Laboratory. He decided to return to Taiwan in the hope that he could continue to jump out of the comfort zoneto explore different challenges, especially the one which allows him to enrich the experience, to establish the social network,to make contributions for society.br / br / Return to Taiwan to nurture talentsto establish a machine learning application teambr / br / In 2007, Lin returned to Taiwanmade great efforts to share with his students in-depth knowledgeexperience in the data mining field. It is worth to mention that he founded Machine DiscoverySocial Network Mining Lab while teaching at the Department of Computer ScienceInformation Engineering at the National Taiwan University. The NTU team has attended in the highly competitive international competition, the ACM KDD Cup, which is regarded as the world cup competition for data miningknowledge discovery. His team won the championship for five consecutive years,this brilliant record has not yet been broken. In 2013, Lin decided to take a back seatturned his attention to host data miningknowledge discovery competitions, hoping to contribute several years of experience to run competitions. As organizers for four major competitions including the ACM KDD Cup, he has recently analyzed the vital information on how to host a data miningknowledge discovery competitionwill publish it by a dissertation.br / br / Lins long-term research focuses on solving the challenges in the application of machine learning practice. Generally acknowledged as the leading scholar in Taiwanwith collaboration experience for more than 50 companies, Lin has been awarded by many well-known research associations all over the world, including three “Microsoft Research Award”, “Google Research Award”, “IBM Research Award”other outstanding achievements. In addition to recognition by industry fields, Lin is the first Taiwanese researcher in the field of Electrical EngineeringComputer Sciences (EECS) that received US Airforce AOARD Research Grant Award (sponsored for five years). He has also received many awards in the academic field, such as the “Ta-You Wu Scholar Award”“Foundation for The Advancement of Outstanding Scholarship Innovation Award”.br / br / While receiving increased famereputation, Lin’s passion for education has never decreased. Besides, he keeps continuing great efforts on talents cultivation. Currently, his students are from a variety of background, such as researchdevelopment engineers from softwarehardware companies in Taiwanoverseas. There are also professors from well-known schools such as UCSD, as well as start-ups founders. In light of the rapid development of AI in recent years, he has been assigned by Department of InformationTechnology Education, the Ministry of Education to plan the establishment of an artificial intelligence talent cultivation office since 2017. Within onea half years being as the director of that office, he has designed the Taiwan AI course map, AI carnival, AI science websitethe first AI Cupother initiatives, to arise Taiwan students’ enthusiasm toward AI learning.br / br / From the academia to the industry practice—the philosophymission of applied scientistsbr / br / Lin has collaborated with many companies till now. In the process of collaboration, he gradually found that the AI model he designed for paper publication, though seeming to be technically correct, still has a gap between theoryapplication in the practical field. This gap is not due to poor accuracy but due to many details needed to be resolved during actual practice. For example, the data collection may be very fragmentedbiased, the model may skew with the data over time,insufficient name entity tagging is a common problem that needs to be solved. “The key to the future AI battle is not the optimization of the model,the improvement of the accuracy, but rather who can solve these seemingly trivial problems in practical application,develop the core technology to launch products,” he asserted. Therefore, scientists must step out of the lab, to clarify the problems encountered in the practical field of AI further, to find outlets for landing applications,to keep up with the trend development as AI researchtechnology are on stages of evolutionrevolution.br / br / “With twenty years in the academic field, hundreds of academic papers published,hundreds of mastersdoctoral students cultivated, I started asking myself again, what is the next challenge?” Lin decided to jump out of the comfort zone againto think carefully about how to bring AI into the real industry. br / br / In 2019, hehis colleagues co-founded Transcend Technology, an Edge AI startup, with the focus on the explorationanalysis of street view data. In 2020, he joined on secondment to a well-known AI unicorn startup company, Appier, as the chief machine learning scientist, hoping that he could cross the AI academicindustrial arenas by shortening the actual application fault. br / br / Since 2019, in response to the “third wave of artificial intelligence revolution," Lin hoped that the use of AI can be more widespreadmore accessible through his participation in the “AI Innovation Research Center Project of the Ministry of ScienceTechnology”conducted “Advanced Technologies for Resource-Constrained Deep Learning.” br / br / Through Identifying AIs weaknesses from an industrial perspectiveusing R & D as a basis to lay out the future industrial needs of "artificial intelligence popularization, he continues researchesapplies the research results to the industry. He solves various types of pain points with visionshopes to help Taiwans AI startups to play an important role in the world stage.br / br / Real meaningvalue of innovative technology— it can be usedbr / br / “Machine learning tells us that as long as we set goalswalk toward them, the good results will come out.” Lin noted. Encouraging his students to join start-up companies, he later indulges himself in the industry as the technical advisorconsultanteven rolled up his sleeves to work together with colleagues to start a business. It is the best real story that can reflect a scientist embraces the spiritbravery to apply, to experiment, to innovate,to challenge, step by step to make the dream come true. br / br / While moremore applications will come outbenefit to humans, there will be grand challenges arisen in the upcoming era of AI. It is necessary to have experts like Lin who has long been involved in AI theoreticallypractically to relay his valuable experience to next generation to cultivate talents, to explore sciencestechnology positioning,more importantly, to establish a transparent decision-making mechanism that can be understood by the public. Let AI play real value,help humans to create a safeintelligent future world.br / br / |
||
|---|---|---|---|
| Keyword | Artificial Intelligence Cooperation Competitiveness AI indicators | ||
| Download | 走出研究室的科學家 要幫AI找到真正價值.pdf | ||
| Research Project | |||
| Research Team | |||
More like this
Provide the latest information of AI research centers and applied industries
-
O'Intelligent Inc.
AI-enabled Service Assurance Platform for 5G Vertical Application: The platform can help customers quickly import 5G vertical applications by providing overall customized solutions, including equipment evaluation, network deployment, application importing, network maintenance, and operation optimization. The platform can bridge the gap between the telecom industries and the vertical application industries and provide a total solution for the industries to import 5G vertical applications.
-
AAAI Technologies, Inc.
AI Deep Compression Toolchain: To enhance the computing power and lower the power consumption of edge devices, this toolchain delivers a 120x AI model compression ratio from original floating-point inference models, compress into fixed-point and bit-accurate edge models with a smaller than 1% accuracy loss. We also provide an open-source image bank and fully automatic labeling tool, available for free application and use at https://www.aicreda.com/.
-
Autonomous UAV Inspection System
With our autonomous UAV systems, we can effectively reduce many high-risk and high-cost inspections which require high manpower, such as inspections of dams, river, coastline patrols, etc.. Operators only need to setup and launch the system to allow the UAV to complete the designated tasks independently.
-
Development of TheorySystems of Robot Learning from Human Demonstration (LfD)-Development of Learning from Human Demonstration Robot
This project proposes a learning from demonstration (LfD) system that allows robots to be not only taught by human via demonstration but also adjusted by themselves.
-
5G C-V2I Enabled Intelligent Real-time Trajectory PredictionWarning System
This study proposes a 5G C-V2I (Cellular Vehicle-to-Infrastructure) enabled intelligent trajectory predictionwarning system, which can be implemented in a framework including RSUs (Road Side Units) with radar detection ability5G edge computing servers. This study exploits artificial intelligence to predict instant trajectories of vehicles at crossroads. The resulting augmented-awareness navigation information is then broadcasted to road users through 5G C-V2I with low latency. In practical applications, road users can obtain real-time dynamics of surrounding vehicles so that their level of safety can be effectively enhanced.
-
An integrated manufacturing platform, the law of sciencetechnology,industrial ecosystem - smart productionintelligent precision manufacturing with digital decision, AI modeling, big data governancekernel technologies
This project aims to develops a completely manufacturing flexible decision framework by applying big data analytics and AI techniques that integrates data from different decision-making units such as APC, APS, capacity planning, and inventory management planning to increase the decision quality and resilient capability in uncertain risks based on the perspective of factory operation.
-
An integrated manufacturing platform, the law of sciencetechnology,industrial ecosystem - smart productionintelligent precision manufacturing with digital decision, AI modeling, big data governancekernel technologies
A self-improving thermal error compensation model is established by federated learning to improve the machine tool performance under different ambient temperatures. Each machine can share and update the latest model parameter to improve the initial model from different ambient temperature variation trails, but still keep the private ambient information of each machine.
-
Defect detection and re-polishing system for highly reflective polished and ground parts.
Develop manufacturing technology and value for difficult-to-machine superalloy materials, fully automated machining technology, improve efficiency and machining quality stability, and combine automation and intelligence for unmanned autonomous operation of machining and inspection.
-
Integration of an ICU Visualization Dashboard (i-Dashboard) as a Platform to Facilitate Multidisciplinary Rounds
Multidisciplinary rounds (MDRs) are scheduled, patient-focused communication mechanisms among multidisciplinary providers in the intensive care unit (ICU). The surgical ICU team of National Cheng Kung University Hospital has developed and integrated i-Dashboard as a platform to facilitate MDRs. i-Dashboard is a custom-developed visualization dashboard that supports (1) key information retrieval and reorganization, (2) time-series data, and (3) display on large touchscreens during MDRs. The i-Dashboard increases the efficiency in data gathering and enhances communication accuracy and information exchange in MDRs.
-
Super-fast Convergence for Radiance Fields Reconstruction
The NeRF-based technique describes a super-fast convergence approach to reconstructing the per-scene radiance field from a set of images that capture the scene with known poses.
