On September 23, the 2018 Global Machine Learning Technology Conference hosted by Boolan was in full swing, and finally ushered in the most important guest of the conference, the most important founder of machine learning, and the Taishan Beidou in the field of machine learning-Michael · Professor Michael I. Jordan, let’s review the highlights of the conference~
The entire agenda starts with the opening speech of Boolan CEO Li Jianzhong. The speech expounded Boolan's vision of "carrying wisdom on a global scale". Under the witness of Professor Jordan, Li Jianzhong unveiled Boolan's employment training and education brand-Boda IT City to all the participants. He said that artificial intelligence competition is based on talents, and it is inseparable from the collaboration of top talents to promote the rapid development of the AI industry from the rise. There is a huge gap in the demand for AI talents in China, and top talents are even more in short supply. Boda IT City, through the integration of Boolan’s accumulated industry first-line expert lecturers, corporate resources and high-end corporate training experience, it has opened many ace professional courses such as artificial intelligence, product manager, Python development, big data, Java, Web front-end, etc. A training model that meets the needs of modern artificial intelligence and IT Internet talents, and strives to build the "Whampoa Military Academy" in the IT Internet industry.
After the speech, we ushered in a keynote speech by Michael I. Jordan, the father of machine learning: The principle of human-centered artificial intelligence: learning system meets economic system.
Professor Jordan reviewed several breakthroughs in the field of machine learning since the 1990s, and pointed out that the correct goal of human-centered artificial intelligence should not be to create human imitative AI. Our goal should be: to develop mutual and human Small-scale intelligence that can be well coordinated.
Regarding Jordan's keynote speech, the lecturer of this conference, Hong Liangjie, the director of data science at Etsy, a well-known e-commerce company in the United States, made a wonderful summary:
"Over the weekend, I listened to a keynote speech by Michael Jordan on the development prospects of AI. In general, Michael has two core points: First, the current artificial intelligence is to use machines to simulate human behavior as much as possible. The path may not be a good development path. He advocates the use of machines to drive the enhancement of human intelligence, which is what he calls "Intelligence Augmented." Second, the human brain and the market are two different types of intelligence. Big motivation. In the past, machine learning and artificial intelligence have ignored the huge premise that the essence of human society is an economic society and an economy. He put forward the important proposition of thinking about intelligence in the dimension of the market.
In general, Michael's speech is a high-level building, and it can be regarded as a clear stream in the current academic and industrial circles that are dazzled by the immediate prosperity. His idea of "artificial intelligence = data + algorithm + market" may be the major direction and driving force for the next wave of development. "
During the speech, Professor Jordan also announced the latest research results of his team, RAY-an open source platform for new AI applications. Distributed architecture Ray will conduct in-depth and extensive exploration of challenges in large-scale, real-world problem domains.
Special focus is on the decision-making aspects of AI, including reinforcement learning, bandits, model selection and hyperparameter tuning.
At the end of the speech, Professor Jordan listed some high-quality machine learning communities.
The whole speech is a high-level building, looking at the development and challenges of machine learning from the source. We will elaborate on it in a later article, so stay tuned.
After the speech was a fireside conversation, Professor Jordan had an in-depth communication with the audience. The audience at the conference are all AI elites from the Chinese Internet, which can be described as Crouching Tiger, Hidden Dragon. The audience kept raising high-level questions, and Professor Jordan also answered them one by one.
After lunch, enter the afternoon session of the lecture session. The 12 full strength sharing sessions covered the best practices and cutting-edge theories of machine learning in business monetization, search and recommendation, education, natural language processing, O2O, and new retail.
Teacher Zhou Hanning from Hulu, Huang Yiming from eBay and Teacher Du Chuanlong respectively introduced the application of smart recommendation technology in streaming media and e-commerce advertising recommendation.
And Teacher Su Hanjing from Tencent and Teacher Wang Xingxing from Meituan Dianping introduced the application of machine learning in business monetization, and explained to the audience how machine learning went from the ivory tower to the market.
After all, technology serves people. Only when it adapts to formal application scenarios can it be considered as a real landing. Focusing on cutting-edge engineering practices has always been the most important feature of the Boolan series of technical conferences. We have also invited the most prominent companies and technical experts in various fields to share the hottest best practices.
Specific to the actual application scenarios of machine learning technology, Mr. Wang Sheng introduced the application of machine learning in new retail based on the advanced experience of Xiaomi offline outlets; Mr. Guo Zhiyang introduced the application of data technology in the field of health insurance risk control; Cui Teacher Wei introduced how to use machine learning algorithms to achieve intelligent education for thousands of people and teach students in accordance with their aptitude; Teacher Wang Guanchun introduced how to use dialogue robots to promote a more natural way of human-computer interaction; Teacher Gu Xiang introduced how to use aviation big data To achieve predictive maintenance, use deep learning methods to achieve abnormal point detection.
In addition, Peng Yao from Qiniuyun and Yinkui Zhang from Xedge.ai brought wonderful speeches on tools, frameworks and optimization. The speeches at this conference can be said to cover all the main application areas of machine learning, from theory to practice, from architecture to tools. Teachers and professional audiences from different fields analyzed the questions raised at the beginning of the conference in a comprehensive and in-depth way through wonderful sharing, critical questions and in-depth dialogues: what should AI do?
The two-day conference ended in such a warm atmosphere. The so-called machine learning is not only the training of machines, but also the learning of humans. As Boolan founder Li Jianzhong said, machines have replaced humans in continuous evolution, expanding the boundaries of human possibilities, and this also makes us look forward to the future, to see how high artificial intelligence will take us?