Yuchao Liu

Enhancing recommender systems by incorporating social information

Although recommendation techniques have achieved distinct developments over the decades, the data sparseness problem of the involved user-item matrix still seriously influences the recommendation quality. Most of the existing techniques for recommender systems cannot easily deal with users who have very few ratings. How to combine the increasing amount of different types of social information such as user generated content and social relationships to enhance the prediction precision of the recommender systems remains a huge challenge. In this paper, based on a factor graph model, we formalize the problem in a semi-supervised probabilistic model, which can incorporate different user information, user relationships, and user-item ratings for learning to predict the unknown ratings. We evaluate the method in two different genres of datasets, Douban and Last.fm. Experiments indicate that our method outperforms several state-of-the-art recommendation algorithms. Furthermore, a distributed learning algorithm is developed to scale up the approach to real large datasets.

An uncertain control framework of cloud model

The mathematical representation of a concept with uncertainty is one of the foundations of Artificial Intelligence. Uncertain Control has been the core in VSC systems and nonlinear control systems, as the representation of Uncertainty is required. Cloud Model represents the uncertainty with expectation Ex, entropy En and Hyper-entropy He by combining Fuzziness and Randomness together. Randomness and fuzziness make uncertain control be a difficult problem, hence we propose an uncertain control framework of Cloud Model called UCF-CM to solve it. UCF-CM tunes the parameters of Ex, En and He with Cloud, Cloud Controller and Cloud Adapter to generate self-adaptive control in dealing with uncertainties. Finally, an experience of a representative application with UCF-CM is implemented by controlling the growing process of artificial plants to verify the validity and feasibility.

Cloud Model Approach for Lateral Control of Intelligent Vehicle Systems

Studies on intelligent vehicles, among which the controlling method of intelligent vehicles is a key technique, have drawn the attention of industry and the academe. This study focuses on designing an intelligent lateral control algorithm for vehicles at various speeds, formulating a strategy, introducing the Gauss cloud model and the cloud reasoning algorithm, and proposing a cloud control algorithm for calculating intelligent vehicle lateral offsets. A real vehicle test is applied to explain the implementation of the algorithm. Empirical results show that if the Gauss cloud model and the cloud reasoning algorithm are applied to calculate the lateral control offset and the vehicles drive at different speeds within a direction control area of ±7°, a stable control effect is achieved.

Path planning algorithm regarding rapid parking based on static optimization

To solve problems of automatic parking during the autonomous driving, a rapid optimization algorithm regarding parking has been proposed in this paper. The biggest challenge comes from the non-holonomic characteristics of automobiles and short-distance obstacles. The algorithm in question solves the obstacles avoidance by using Minkowski sum. Geometric path planning is derived from kinematics and connected with path parameters of Runge-Kutta discretization method. Since parking paths and relevant parameters can be obtained by virtue of Runge-Kutta discretization method, the optimization method for every discrete path is to calculate parking paths that are independent from parking lots. The static optimization method can be resolved by digitization in an effective way. Effects of the above mentioned algorithm will be evaluated by different stimulation scenarios. The stimulation experiments show that path planning algorithm regarding rapid parking based on static optimization is feasible for the following three parking scenarios-parallel parking, vertical parking and angle parking. All these three parking forms can be achieved on the premise of no modification in algorithm. Path planning can be completed within several milliseconds even in a narrow space.

Cloud computing beyond turing machines

Development of Internet technology and social network has greatly changed the traditional software engineering based on single Turing machine. Software development will be cooperated and completed on the network with collective intelligence. The interaction among human-machine and machine-machine becomes the kernel of Internet computing, while Turing model studied on Entscheidungsproblem based on an automatic computer theoretical model without interaction with people. Clusters or virtual clusters become the basic platform of cloud computing centers. And SaaS (Software as a Service), PaaS (Platform as a Service), IaaS (Infrastructure as a Service) become the common knowledge for software engineers. Furthermore, the research of network science has discovered lots of physical law about the distribution of information resources, such as the power law distribution of Web services. This paper focuses on this computing paradigm, and analyses the trend of software development style facing Internet computing. And the features of cloud computing such as virtualization, granular computing, soft computing and uncertainty are discussed too.

Comparative study of type-2 fuzzy sets and cloud model

The mathematical representation of a concept with uncertainty is one of foundations of Artificial Intelligence. Type-2 fuzzy sets study fuzziness of the membership grade to a concept. Cloud model, based on probability measure space, automatically produces random membership grades of a concept through a cloud generator. The two methods both concentrate on the essentials of uncertainty and have been applied in many fields for more than ten years. However, their mathematical foundations are quite different. The detailed comparative study will discover the relationship between each other, and provide a fundamental contribution to Artificial Intelligence with uncertainty.

From turing machine intelligence to collective intelligence

Almost all of the progress of artificial intelligence in the last 50 years has been based on the Turing model and Von Neumann architecture. Researchers have always tried to put the human intelligence into machines by the ways of algorithms, codes or symbols that could be understood and executed by machines, thus, we may be bounded to Turing model too tightly. In Internet and World Wide Web and developing cloud computing, network has changed the role from a single huge Turing machine or sum of some Turing machines to the collective intelligence, where the inputs or outputs of nodes in network are happening not only among computers, but also among people, such that Internet has been beyond Turing machine. Users in Internet who own similar interests may cluster naturally into scalable and boundless communities with uncertainty, where online interaction avoids the difficulty of common sense representation in traditional artificial intelligence. Furthermore, collective intelligence may emerge from the crowds interaction. Those would become the new research frontiers in intelligence science.

A qualitative requirement and quantitative data transform model

Development of Internet technology and human computing has changed the traditional software work pattern which faces the single computer greatly. Service oriented computing becomes the new tendency and everything is over services. The appearance of mobile Internet provides a more convenient way for people to join the Internet activity. Thus, the meeting between the requirements from a number of users and services is the main problem. Cloud model, a qualitative and quantitative model can be taken to bridge the qualitative requirement on the problem domain and quantitative data services on the solution domain.

On foundations of services interoperation in cloud computing

With a long-run accumulation of IT technologies, cloud computing becomes a new revolution after PC revolution in 1980s, the Internet revolution in 1990s, and the mobile Internet revolution in 2000s. Cloud computing is a paradigm of Internet computing, which takes software as a service oriented to a number of users with changing requirements from time to time, at the same time, takes the response of a requirement as an up-to-date best effort rather than a unique precise one. It is changing the way we share data, information and knowledge. Services support direct reusing of application programs instead of software deployment, and improve usability of resource sharing. Clouds can be regarded as an enabler for interoperation of large scale service provisioning. Therefore, assembling of software became services aggregation under the mature understanding of interoperability.