Approximate algorithms enable NP problems to be solved. Traditional approximation algorithm theory blindly pursues an approximation algorithm to reduce complexity in design. It may spend plenty of time while the effect is disappointing. We explor for a quantum leap from traditional approximate computing and pursue data-driven approximate algorithms and theories. The ultimate results are to provide small and medium-sized enterprises with accurate and efficient queries on big data despite limited hardware resources.
Research Areas
Based on the idea of transforming big data into small data, we research the design method of data-driven approximation algorithm and design the approximation algorithm. Finally, we extend the data-driven approximation algorithm theory to the optimization of different types of query languages/problems. In the case of limited big data resources, real-time analysis can be provided.
