This is a paid-for submitted press release. CCN does not endorse, nor is responsible for any material included below and isn’t responsible for any damages or losses connected with any products or services mentioned in the press release. CCN urges readers to conduct their own research with due diligence into the company, product or service mentioned in the press release.
On May 25th at Tsinghua University, MATRIX Beijing Research Institute, MATRIX Bayesian Research Fund, HuaZhang Co. of China Machine Press, and BIT GAME FOUNDATION PTE.LTD, jointly presented “Artificial Intelligence in the Post-java Era,” in the multi-function hall on the 2nd floor of the FIT Building. Bruce Eckel, President of Mindview and author of “Thinking in Java”, Wu Jinxin, CTO of BIT.GAME, Tan Zhongyi, the head of the Open Source Advancement Team and senior engineer at Baidu, and Professor Steve Deng, a scientist in artificial intelligence, presented their views on the subject
Prior to the symposium, Mr. Bruce Eckel visited the laboratory of Professor Steve Deng, to discuss emerging programming languages, and the future impact of developments in artificial intelligence on scientific progress
Photo: Bruce Ecke with Professor Steve Deng and student
Bruce Eckel: On how Kotlin can improve efficiencies and productivity?
It is now difficult to make additional changes to Java. Many elements of the language are unchangeable and other factors contribute to this inertia. Kotlin emerged precisely to enable a seamless switchover and integration with Java, leading Bruce to assert that Kotlin will succeed Java. Kotlin is a language based on a dynamic mixture of objects and functions optimized for concurrent production; as well as a better user experience throughout the programming process. Bruce shared his method for how to code in a Kotlin style, as well as using Kotlin logic to analyze questions. Kotlin is already quietly infiltrating some Java projects, and will help accelerate Java development.
Bruce Eckel is pictured above.
Wu Jinxin: Microservices architecture in practice
Wu Jinxin, CTO of BIT GAME, discussed the history of evolving architecture for enterprise services, pointing to the current effectiveness of microservices for developing enterprise applications. Generally speaking, this approach is labeled “microservices for large systems,” meaning the deployment of smaller modular services to achieve greater speed and agility.
Tan Zhongyi: an AI deep learning framework for PaddlePaddle
Tan Zhongyi, a senior R&D engineer at Baidu, shared insight into deep leaning frameworks, including the use of libraries to ready-made models that assist in writing deep learning programs. The framework for deep learning primarily consists of the following functions. Firstly, to more easily describing and generating computational graphs; secondly, creating computer generated gradients; and finally, cloaking the details of GPUs and CPUs. In deep learning, the three elements include data, algorithm, and computing, as well as the framework for deep learning.
Tan Zhongyi also went on to introduce PaddlePaddle.
Steve Deng: Toward Machines that Learn and Think Like People
Steve Deng started with a review of the history of deep learning, then presented a comparison of weak and strong artificial intelligence, sharing advances his team has made in the areas of strong artificial intelligence as well as future prospects in these fields.
Steve Deng shared that using deep learning technology, machines will be able to outperform humans in areas we are uniquely skilled in. In certain fields, the best human minds already can’t compare to artificial intelligence. However, deep learning or AI technologies have the characteristic of handling matters very rigidly. Each network as well as each tool is optimized to address a single problem, but cannot make corresponding adjustments in response to environmental changes –this is called weak artificial intelligence. Steve Deng believes that the field can aspire to artificial general intelligence, the ability to adaptively master various matters, which he calls strong artificial intelligence.
After discussing real-world applications of Bayesian reasoning, and the working principle behind Bayesian algorithms, Steve Deng said that his team is working on engineering a machine that can think like a human being. Having designed the underlying structure for Bayesian computing, each logic unit is assigned a set of digits, and thus translated into a particular function. Meanwhile, randomly generated elements are added to each unit, after which data can be combined together to generate binomial classifications and be combined in an arbitrary distribution. On the other hand, Steve Deng uses a probabilistic programming language, characterized by an improved random probabilistic distribution for each variable (as opposed to a single value-fixed variable) making it compatible with Bayesian reasoning.
In closing, Steve Deng discussed the use of modified open-source GPUs for accelerated performance to thwart adversarial nets at deep level. Take the use of whole brain simulation for cognitive development-the application of strong artificial intelligence can stimulate machines to think and learn like human beings, making a significant step forward toward artificial general intelligence.
Audience members at the standing room only event interacted eagerly, bringing the first round of the symposium organized by the MATRIX Bayes Research Fund to a successful conclusion.
Going forward, the MATRIX Bayes Research Fund will continue advancing the concepts of the assetization of data and digitization of assets, as well as promoting technological innovation, facilitating the emergence of new industries and enterprises; as well as providing a platform to showcase, exchange, and collaborate on leading-edge technologies.