"BDTC lecturer" machine study Professor Biopo on academic accomplishment

Biopo (Eric Xing) is currently Professor of computer science at Carnegie Mellon University. His main research interests focus on the development of machine learning and statistical methodology and large-scale computing systems and architectures, in the hope of solving the problem of automated learning, reasoning, and decision making in a potential world of high dimensional, multiple peaks and dynamics in complex systems. Biopo has a Ph. D. In molecular biology from the University of Rochester and a PhD in computer science from the University of California, Berkeley.

In December 2014 12-14th, the 2014 China Large Data Technology conference (CCF), held in Beijing, Biopo will discuss the way of machine learning.

Biopo Ongoing research work includes: 1. Statistical learning Basics, it includes theories and algorithms for estimating time or spatial variation coefficients, sparse structured input/output model and nonparametric Bayesian model; 2. In a distributed system or cloud, build a framework of parallel machine learning based on large model and large data scale; 3. The calculation and statistical analysis of gene regulation, genetic variation and disease correlation were 4. Apply statistical learning to social networks, data mining, and computer vision.

Biopo has published more than 200 peer-reviewed papers so far. He is the associate editor of the Journal of the American Statistical Association (JASA), The Applied Statistics Yearbook (AOAS), IEEE Journal of Analytical and Machine Intelligence (PAMI) and PLoS Journal of Computational Biology (the PLoS Journal of Computational), Journal of Machine Learning (MLJ) and the executive editor of the Journal of Machine Learning Research (JMLR). He is a member of the U.S. Department of Defense Advanced Research Program (DARPA) Information Science and Technology Advisory Group, the National Science Foundation (NSF) Career Award, Alfred P. Sloan Scholar Award, the American Air Force Youth Scholar Award, and the IBM Open Collaboration Research Scholar Award.

Whatever you do, study, work, or lifestyle, his choice is based on his interest, follow his heart. Because like physics, when Biopo selected the Department of Physics, and because like biology, Biopo ran to the dean's Office to register elective courses, a biological double degree, graduation his credit is about 1.7 times times the average classmate.

Biopo's dream is to become a scientist, after graduating from university, Biopo chose to go abroad for further study. Biopo entered a PhD in molecular biology at a university in New Jersey, USA, in 1994. His studies went very well, usually taking 6 years to complete, and Biopo 4.5 got his PhD.

Biology is an experimental science, he felt that doing scientific research also requires rigorous mathematical thinking, training a rigorous methodology. Based on this idea, Biopo a postdoctoral fellow in molecular biology and went to the University of Berkeley to study computer Science (applied mathematics).

After getting another ph. D., Biopo decided to enter the university to pursue research and education to fulfill his dream. He took 12 interviews and received 8 offers, one from Carnegie Mellon University, when 500 people applied for the job and only 4 were admitted.

Biopo has said in an interview earlier that scientific research is often boring and the combination of products is not close enough, such as illusory imagination of the future. There are some misconceptions. Many people, based on one-sided understanding or limited experience, a simple antithesis of theoretical research and applied research, resulting in a superficial exclusionary idea; for example, some "basic research" scholars despise "applied research" scholars who think the latter are not strictly orthodox, or that some engineers in the industry despise university researchers, Think the latter is not practical. It's hard to be a leading scholar and engineer with this idea.

In his opinion, computer science research is the least boring research, and excellent computer research often embodies the close combination of theoretical and practical problems. It is not a pure derivation formula, design model, proof theorem; Rigorous computer science needs to deal with math, which may be boring for some people, but it's still fun and exciting for others.

He also said that to do scientific research, can not be made into competitions, but a kind of the unknown and perfect self pursuit. As the game is utilitarian in nature, is not to do first-class results, but also need to have an essential question of curiosity, hope to understand and solve this problem. Faster than others, beating others is only a by-product.

Biopo's position is a researcher, and has his own profound opinion on the quality and accomplishment that the researcher should have. He thought:

1. Honesty. Research is a self-examination process, not like exams, do homework, someone to change you, you need to be responsible for your own research.

2. Have a passion. The researchers will be too excited to sleep because of the problem and always think of some problems.

3. Can endure hardship. Doing research, especially interdisciplinary research, needs to do things that others have not done, it is very difficult, it is likely a long process, there is no progress, and every day to have a high intensity of work.

4. Optimistic mood, and stable. When a researcher's research advanced the time, caused his article to be difficult to publish, is difficult to accept by the society, he must be able to carry on, to all kinds of criticism can critically absorb.

5. Aggressive and courageous. Dare to do other people have not touched the difficult subject. It may be better to follow the study, but in order to make a big contribution, we must dare to try higher risk.

6. Perseverance, to be able to focus on long-term research. There are a lot of attractions in the world now, and the people who do the research must find the motivation to focus on their research work.

Regarding the scientific accomplishment, Biopo also has carried on the summary, has the following points: The insight, the innovation spirit, the unceasing study, is good at the exchange, very strong curiosity, simultaneously handles many questions the ability. (Zebian/Zhonghao)

More lecturers and schedule information please pay attention to the 2014 China Data Technology Conference (and the second CCF large data academic conference) official website. Another, now buy BDTC tickets to enjoy a maximum of 1500 yuan discount, the event until October 17. Advance Advance

Free Subscription "CSDN cloud Computing (left) and csdn large data (right)" micro-letter public number, real-time grasp of first-hand cloud news, to understand the latest big data progress!

CSDN publishes related cloud computing information, such as virtualization, Docker, OpenStack, Cloudstack, and data centers, sharing Hadoop, Spark, Nosql/newsql, HBase, Impala, memory calculations, stream computing, Machine learning and intelligent algorithms and other related large data views, providing cloud computing and large data technology, platform, practice and industry information services.