seabet betting title:seabet betting design"Idea Tree"
seabet betting time:2020Year10month31日 19:00
Reporting location:School HeadquartersComputer Building313
seabet betting Introduction:
Understanding how an seabet betting works is not too difficult a task，But you need to understand how the seabet betting is designed，But it is very difficult。When we see the ingenious algorithms designed by others，In admiration，Often there are also cases with G. Polya Similar confusion:“How is such an exquisite seabet betting designed?Why didn’t I think of this seabet betting?”
This seabet betting will mainly introduce"Idea tree of seabet betting design", that is: First observe the structure of the problem，Then design a preliminary algorithm，Next, observe the behavior of the algorithm，Then iteratively improve the algorithm。We emphasize observing the structure of the problem，Emphasis on algorithm design based on problem structure，Emphasis on iterative improvement of algorithms based on the structure of the problem and the behavior of the algorithm—the process of solving the problem should not just try each algorithm seabet betting one by one，Not purely relying on inspiration，Instead, it should rely on the understanding of the problem structure;The deeper we understand the problem structure，The more helpful it is in designing the solution seabet betting。
About the speaker:
Bu Dongbo，Researcher at the Institute of Computing seabet betting, Chinese Academy of Sciences，Research interests include algorithm design、Bioinformatics (protein structure prediction), etc.。Main research results: To the classicsSAT seabet betting, accurately estimate randomness3SAT is located at4.21 Nearby; Developed "using artificial intelligence seabet betting to assist algorithm design"AIA System，In the classic course scheduling seabet betting, the change from “designing algorithms based on inspiration” to “learning algorithms from data” has been achieved;Designed a protein structure prediction algorithm that combines combinatorial optimization seabet betting and deep learning seabet betting FALCON, won international competitionCASP Third place;Proposed a new "reverse Monte Carlo" sampling seabet betting，Solved the long-standing problem of objective function design in optimization problems；Proposed a new neural network for learning conditional joint probabilities between residuesCopulaNet, significantly improves the accuracy of seabet betting, and the performance exceedsAlphaFold。
