We are in another cycle of excitement over Artificial Intelligence. Much of that is being driven by advances in deep learning and other areas of machine learning related to computer vision, speech detection and robotics. Some of these advances are truly important and bound to have large impacts on what software can do for the enterprise.  There is also a great deal of hype and misunderstanding of what AI is and where current techniques are likely to take us. This misunderstanding is likely to lead to difficulties in turning these advances into useful products and so result in overall malinvestment. Some key concepts not well appreciated include: the difference between perception and thinking, the practical difficulties posed by supervised learning, the poor ability of current modeling techniques to benefit from and incorporate human knowledge including the concept of causality. We will discuss these shortcomings and possible ways that the field might be able to achieve more success. 

In the wave of digital transformation, as a traditional chemical enterprise, what kind of challenges and opportunities will facing? How to take advantage of intelligence and big data?  As a typical traditional industry company,  what are the practical expectations in big data and artificial intelligence?

AI is math. Though we do not have to proof something when creating a deep neuro network, we are using statistical methods to learn the patterns from the data. Unsurprisingly, we've done some unseen assumption behind the model that will never lead to an intelligence could understand and generalize about the world. Today we are going to discuss about the limitations of machine intelligence based on some real world projects, from image recognition to natural language processing, about both data and models.

Running Machine Learning in your organization is more than clicking through an AI tutorial on your laptop or building a Proof-of-Concept. In this presentation we will talk about the journey of a company becoming AI-driven. This includes how to become data-driven, how to combine narrow AI-models to achieve complex tasks and last but not least how to deploy AI algorithms continuously based on the principles of Continuous Integration and Continuous Delivery (CI/CD). All this will be illustrated with several ThoughtWorks project examples.      

Daimler China has made great progress in the areas of big data in the last two years. Significant achievements in terms of data strategy, foundation, algorithms & modellings, as well as visualization are described, along with three tangible use-case examples ranging from the connected car, to making use of neural networks in business decisions as well as predicting sales.

This presentation is about my experience to create a unique data product to solve airfare prediction problem. 

The idea comes from a question several years ago: "How can I buy cheapest airline ticket when I decide when to departure?". Inspired by Farecast (Already closed after brought by Microsoft), I started to collect daily airfare 2 years ago and now I collected above 35 billion lines of records. Based on these data, I created a product called "aiflygo" (爱飞狗 in Chinese), which can give you real time airfare suggestion, recent price graph, history price search. You can use these data to help you decide when to buy a ticket.

This product has been put online as Wechat mini app (scan below code) for nearly one year. iFanr and reported this product (https://www.ifanr.com/minapp/956764, https://minapp.com/article/17486/), Qdaily also interviewed me about the price trend (https://www.qdaily.com/cards/50166.html). I also wrote some articles about my research (in Chinese)

* https://www.jianshu.com/p/31c7210c535a

* https://www.jianshu.com/p/366e839af3ea

* https://www.jianshu.com/p/3b22fb101189

How to manage enterprise data asset? Should we build the data platform top-down or bottom-up? Business driven or Technology driven?

Can we start data innovation projects considering poor data quality? How to quickly discover business value and delivery MVP products fast?

We need a new set of enterprise data architectures and methods to build data and algorithm-driven businesses. "

Minsheng Bank's intelligent unified identity and access control platform provides cross-security domain, cloud, group, multi-legal personnel lifecycle management. Provided SSO and authorization management services for hundreds of systems across the bank. The organization management service integrates the technologies of each layer of the defense-defense system through artificial intelligence and big data to identify the risks. The intelligent decision-making engine automatically selects the four-dimensional authentication means that the user knows, the user belongs to the user, the user is, and the user performs the combination. Provide comprehensive and comprehensive risk protection measures and means.

As the digital progresses, data security has become one of the most important considerations for many companies. You may have heard of the 3Rs Principle (Rotate, Repave, and Repair) from Cloud Native, and have learned the Zero Trust Principle promoted by technology leaders like Google. 

At the same time, more and more organizations turn to professional security practice. Data and privacy protection services unable to suit their particular businessl needs. That’s why real-time deep practice and data security protection have been new focuses of organization security. We will start by talking about real practice, and through sharing real cases, introduce ways to be safer and more effective in today’s world with serious data security challenges.

Data governance is a huge and complex work, there are obvious problems in the traditional method of data governance about cycle and performance. Metadata provides an entry point for data analyzing and an entry point for data governance with large and complex data. Metadata-driven data governance architecture provides a better way in data governance.

The humanitarian sector is fighting stagnancy when it comes to technology’s influence. The technology that is currently being talked about a lot in the technology sector sometimes doesn’t work in these high resource-constrained environments. The sector however still seems inundated with data, but the challenge is harnessing the same. What is the kind of money and skills needed to support the social sector, when it comes to ensuring that data can make a difference to the social ecosystem? The solutions vary from being tech based to policy based ones and more. This talk would outline some key problems in the current humanitarian data ecosystem and some suggestions/solutions that we arrived at through extensive research and interviews with some key influencers in this space.