Decision Intelligence

HUYNH Laboratory
Professor：HUYNH, Nam Van

E-mail： E-mai
［Research areas］
Data Science, Artificial Intelligence, Operations Research, Decision Support Systems
［Keywords］
Machine learning, Data analytics, Argumentation, Optimization, Uncertainty management, Decision analysis

Skills and background we are looking for in prospective students

Our lab gladly welcomes highly motivated students who have acquired a solid background and skills through their undergraduate/graduate program, and who have a willingness to study actively and cooperatively.

What you can expect to learn in this laboratory

Students are expected to develop valuable knowledge and problem-solving skills that can be applied in the areas of practical applications of data science and decision analysis. These problem-solving skills include the ability of 1) identifying and formulating decision problems; 2) acquiring and modelling knowledge for solving them; and 3) generating, evaluating, validating and implementing solutions to these problems.

【Job category of graduates】 Academic jobs in universities, IT industry, Consulting and marketing companies, etc.

Research outline

With the research vision that Data Science paired with (data-driven) Knowledge Management will provide the core of Intelligent Decision-Support Systems, the strategic mission of our laboratory is to develop both fundamental research and applied research relating to the creation, integration, reasoning and use of knowledge from data in intelligent decision support systems.

In particular, our current research interests lie in data analytics and machine learning (ML), AI reasoning, uncertainty management, operations research and decision making. The synergy of these research areas will allow us to provide adaptivity at all levels of intelligent decision support systems in today’s era of big data, from acquiring, fusing and reasoning to using knowledge for supporting decision-making.
As graphically illustrated in the figure, we explore fundamental topics in these research areas combined with practical applications in fields ranging from e-commerce and marketing intelligence to finance, industrial and business management. Such an approach essentially helps to guide theoretical research, ensuring that it is relevant and significant, while providing novel and advanced solutions to practical problems and research questions. Some specific topics are as follows:

• Multi-source learning and knowledge fusion: This research is concerned with the development of a novel methodology for learning, reasoning and fusion of knowledge discovered from multiple sources of data in a distributed environment for decision support in intelligent systems. We particularly focus on a new integrated approach that combines advanced ML techniques with evidential reasoning based on Dempster-Shafer theory of evidence for the development of multi-source learning frameworks capable of appropriately handling uncertainty and conflict/inconsistency.

• Interpretable ML: Inspired by the generality of Argumentation in AI reasoning and its dialectical nature as how people convince each other to draw conclusions by exchange of arguments, our research aims to establish an argumentation-based approach for developing a novel dialectical framework for explanations and evolvement of learning systems, articulated along the following research challenges: 1) Explainable by interrogation: how to generate the most faithful explanation of a given (black-box) learning model; 2) Explainable by design: how to design inherently interpretable models without the cost of sacrificing accuracy for interpretability; 3) Evolvement of learning systems: how black-box models with their explanations and inherently interpretable models collaborate and push each other to evolve their capabilities for both accuracy and interpretability.

Key publications

D.-V. Vo, J. Karnjana, V.-N. Huynh. An integrated framework of learning and evidential reasoning for user profiling using short texts, Information Fusion 70 (2021), 27-42.
D.-H. Nguyen, V.-N. Huynh. Revealed preference in argumentation: Algorithms and applications. Intern. J. of Approximate Reasoning 131 (2021), 214-251.
T. Nguyen-Mau, V.-N. Huynh. An LSH-based k-representatives clustering method for large categorical data. Neurocomputing 463 (2021), 29-44.
V.-D. Nguyen, V.-N. Huynh, S. Sriboonchitta. Integrating community context information into a reliably weighted collaborative filtering system using soft ratings. IEEE Transactions on Systems, Man, and Cybernetics: Systems 50 (2020), 1318-1330.

Teaching policy

The guiding principle of research in the laboratory is that a well-developed decision support system must be based on a sound theoretical foundation. Our education and research strategy is therefore trying to balance the advancement of theoretical research and its applicability in practice. Students will be exposed to relevant practical problems during development and improvement of their systems thinking and modelling competencies, which in turn helping them get insights into problems, inspired and guided theoretical research (making it relevant, innovative, effective and significant) in knowledge modelling and decision making. This strategy is done through regularly-held lab meetings and activites.

［Website］ URL：https://www.jaist.ac.jp/~huynh/