Dr. Mahmoud Ahmad Al-Khasawneh from Al-Madinah International University, Malaysia
Research Area: Image Encryption, Big Data, BlockChain and IOT
Title: Lifelong learning DNN
Abstract: The term AI is actually coined when Neural Network came into existence. Number of fields like image recognition, language translation, pattern discovery and especially game engineering get benefitted through DNN, these DNNs not only walked parallel with a brain but in some situations, ran far faster than the most intelligent human being. A million of weighted connections act like a brain, except in one key respect—they don't learn over time, as animals do. It takes a very long time to train a traditional DNN on a dataset, and, once that happens, it must be completely re-trained if even a single piece of new information is added. Once designed, programmed, and trained by developers, they do not adapt to new data or new tasks without being retrained, often a very time-consuming task.
Thus Real-time adaptability by AI systems became a hot topic in research. By the dawn of Lifelong-DNN a dramatic revolution was observed. LDNN is the only AI solution that allows for incremental learning and is the breakthrough that companies across many industries have needed to make deep learning useful for their businesses. Lifelong DNN technology allows for a massive reduction in the time it takes to train a neural network and all but eliminates the time it takes to add new information In several trials, the scientists revealed that previously trained NNs could adapt to new circumstances swiftly and efficiently without undergoing further training. The update allows for a "significant reduction in training time compared to traditional DNN; few seconds versus few hours - a reduction in overall data needs, and the ability for deep learning neural networks to learn without the risk of forgetting previous knowledge - with or without the cloud."
"In a few years, much of what we consider AI today won't be considered AI without lifelong learning."
Prof. Guandong Xu from University of Technology Sydney, Australia
Research Area: Data mining and data analytics; Web mining and recommender systems; Semantic computing and knowledge graph; Text mining and NLP; Social network analysis, Social media mining, Social Analytics; Predictive analytics and Fintech analytics
Title: AI-empowered Finance Data Analytics
Abstract: Finance Data management, including stock market, superannuation, insurance etc, is a typical big data problem, involving a lot of customer data and finance data. Traditional wealth data analytics is merely relying on quantitative approaches, resulting in inadequate insights for smart wealth management. In this talk, we will introduce our recent research progress on how to leverage customer unstructured and structured data to derive business insights for decision-making support and better customer management. In particular, cutting-edge methods, such as deep learning, representation learning, and interpretable machine learning have been adopted and explored in our research. Experiments conducted on real world datasets validated our developed models in terms of effectiveness and efficiency, and added business value.