1, landing, cross-border, efficiency, leadership meet to talk about big data;
2, big Data become the mainstream industry needs 5-10 years;
3, the data is the appearance, the essence is the question;
4, Doug Cutting:fuel for Change:data, EDH, Style catches On:ecosysystem, the data multi-tool, Cloudera;
1, white Paper, development trend;
2, integration, cross-border, foundation, breakthrough;
3, multi-disciplinary integration, security and privacy is worrying;
4, deep learning, visual analysis, open source system;
5, equipment monitoring monitoring and monitoring of the data become a hot spot;
6, Tianjin Binhai New Area;
7, Big Data Creative competition, student performance prediction;
1, Store anything, Object stonge, scalable;
2, Real time bidding;
3, how to do data analysis in a non-programmatic way;
4, Watson perceptual computing;
5, DeepQA depth problem solution;
1, how to use big data cock wire;
2, communication waiting for collaboration will waste more than one machine's performance;
3, Hadoop 90% is the time to wait, the node model;
4, pervasive software equipment, middle layer, middleware;
5, the risk is very big, the return is very low, not every company can play;
6, machine learning needs iterative iteration;
7, hard disk reading cost is very high;
8. The two-direction of data and model;
9. ML Machine Learning
10, Spark:faster MapR on Data Parallel;
11, Graphlab,petuum, semi-synchronous;
12, SAP based on the parallelization of the structure;
1, read the hearts of people;
2, Gartner Forecast;
3, perception, thinking, control;
4, the ability to learn is the essence of intelligence;
5, experience is data; Baidu experience;
6, Human (everyone) to connect 1000 devices-the internet of everything;
7, the Internet of everything-everything intelligence;
8, generalization error-"decomposition-" hypothesis-"model;
There is no perfect hypothesis;
9, the data is not perfect: noise, not whole, limited;
10, the algorithm is not perfect;
11, intelligence is very energy-consuming;
BDTC2014 China Big Data Technology conference minutes