Seminar of AI Research for Finance, E-Commerce, and Social Media Applications
- Online presentation via Zoom
- 17 August 2022 (1:30pm-5:50pm), Singapore Standard Time
- 17 August 2022 (12:30pm-4.50pm), Vietnam Standard Time
Organizers in Singapore:
- Nanyang Technological University, Singapore
- Temasek Foundation, Singapore
Partners in Vietnam:
- Asst. Prof. Luu Anh Tuan, School of Computer and Engineering, Nanyang Technological University, Singapore.
- Email: [email protected]
Artificial Intelligence (AI) has enjoyed tremendous success over recent years with many real-world applications and played crucial role in various aspects of human life, as well as an important factor to boost the economy. This series of research seminars involves research talks from well-known researchers introducing recent advanced AI technologies applied into the domains of Finance, E-Commerce, and Social Media.
Program and time table:
Date: August 17 (Wednesday), 2022, Vietnam Standard Time
|12:30 - 12:40||Opening and Introduction|
|12:40 - 13:20||Assoc. Prof. Quan Thanh Tho, HCMUT, Vietnam||SOBOG: SOcial BOt detection using Graph neural networks|
|13:20 - 14:00||Assoc. Prof. Erik Cambria, NTU, Singapore||Neuro-symbolic AI for Social Media Applications|
|14:00 - 14:40||Prof. Gao Cong, NTU, Singapore||The Data Issues in Evaluations of Recommendation Systems|
|14:40 - 15:20||Asst. Prof. Liu Ziwei, NTU, Singapore||Intuitive Neural Avatars|
|15:20 - 16:00||Assoc. Prof. Li Boyang, NTU, Singapore||Multimodal Computational Narrative Intelligence|
|16:00 - 16:40||Asst. Prof. Le Duy Dung, VinUni, Vietnam||Techniques for Efficient Retrieval of Top-K Recommendations|
|16:40 - 16:50||Closing|
SOBOG: SOcial BOt detection using Graph neural networks
Assoc. Prof. Quan Thanh Tho
Abstract: Over the past decades, online social networks such as Twitter and Facebook have become a significant part of people’s daily lives, particularly amid the ongoing global calamity - the COVID-19 pandemic. This gives room for social bot attacks that are designed to automatically replicate the behavior of real accounts. Most of these bots are employed for nefarious purposes such as disseminating false information, artificially amplifying the popularity of a person or movement, or spreading spam. Many studies have been conducted in an attempt to discover new strategies for identifying social bot accounts. To deal with large-scale attacks from social bots, Machine Learning (ML) has emerged as a noticeable path of bot detection problem that can handle massive amounts of data.
Various approaches using different ML were reported for social bot detection. Recently, the emerging Graph Neural Network (GNN) has been increasingly attracting much attention due to its suitable nature to handle social media data. However, when handling
large-scale data, this technique expectedly suffers from high computational complexity. To address this issue, we propose SOBOG (SOcial BOt detection using Graph neural networks), which breaks the original huge graph into account-based closure
subgraphs for more efficient processing. Experimenting with real Twitter benchmarking datasets, our model gained promising performance, as compared to other related works.
Bio: Dr. Quan Thanh Tho is an Associate Professor in the Faculty of Computer Science and Engineering, Ho Chi Minh City University of Technology (HCMUT), Vietnam. He received his B.Eng. degree in Information Technology from HCMUT in 1998 and received Ph.D degree in 2006 from Nanyang Technological University, Singapore. His current research interests include formal methods, program analysis/verification, the Semantic Web, machine learning/data mining and intelligent systems. Currently, he is the Vice Dean for Academic Affairs of the Faculty Committee.
Neuro-symbolic AI for Social Media Applications
Assoc. Prof. Erik Cambria
Abstract: With the recent developments of deep learning, AI research has gained new vigor and prominence. However, machine learning still faces three big challenges: (1) it requires a lot of training data and is domain-dependent; (2)
different types of training or parameter tweaking leads to inconsistent results; (3) the use of black-box algorithms makes the reasoning process uninterpretable. At SenticNet, we address such issues in the context of NLP via sentic computing, a neurosymbolic
approach that aims to bridge the gap between statistical NLP and the many other disciplines necessary for understanding human language such as linguistics, commonsense reasoning, and affective computing. Sentic computing is both top-down and bottom-up:
top-down because it leverages symbolic models such as semantic networks and conceptual dependency representations to encode meaning; bottom-up because it uses subsymbolic methods such as deep neural networks and multiple kernel learning to infer syntactic
patterns from data. We apply sentic computing to areas such as social media marketing, social media monitoring, financial forecasting, cyber issue detection, and more.
Bio: Erik Cambria is the Founder of SenticNet, a Singapore-based company offering B2B sentiment analysis services, and an Associate Professor at NTU, where he also holds the appointment of Provost Chair in Computer Science and Engineering. Prior to joining NTU, he worked at Microsoft Research Asia (Beijing) and HP Labs India (Bangalore) and earned his PhD through a joint programme between the University of Stirling and MIT Media Lab.
His research focuses on neurosymbolic AI for explainable natural language processing in domains like sentiment analysis, dialogue systems, and financial forecasting. He is recipient of several awards, e.g., IEEE Outstanding Career Award, was listed among the AI's 10 to Watch, and was featured in Forbes as one of the 5 People Building Our AI Future. He is an IEEE Fellow, Associate Editor of many top-tier AI journals, e.g., INFFUS and IEEE TAFFC, and is involved in various international conferences as program chair and SPC member.
The Data Issues in Evaluations of Recommendation Systems
Prof. Gao Cong
Abstract: The talk will first give an overview of the research on recommendation systems done by the team of the speaker. Then the rest of the talk will focus on the data issues in recommendation systems. We take an in-depth look
at a fundamental but often neglected aspect of the evaluation procedure, i.e. the datasets themselves. To do so, we adopt a systematic and comprehensive approach to understand the datasets used for implicit feedback based top-K recommendation.
We look at the characteristics of these datasets to understand their similarities and differences. Finally, we conduct an empirical study to determine whether the choice of datasets used for evaluation can influence the observations and/or conclusions
obtained. Our findings suggest that greater attention needs to be paid to the selection process of datasets used for evaluating recommender systems in order to improve the robustness of the obtained results.
Bio: Professor Cong Gao obtained his PhD in Computer Science from the National University of Singapore (NUS) in 2004. He received both his Master of Engineering and Bachelor of Engineering degrees in Information Systems from Tianjin University, China in 1999 and 1996 respectively. Professor Cong was a Postdoctoral Research Fellow with the University of Edinburgh from 2004 to 2006 and a Researcher with Microsoft Research Asia from 2006 to 2008. He joined the Department of Computer Science in Aalborg University, Denmark, as an Assistant Professor in 2008. He became an Assistant Professor with the School of Computer Science and Engineering (SCSE) in 2010 and was promoted to Associate Professor with Tenure in 2015 and to Professor in 2018. He has a courtesy appointment with the School of Physical and Mathematical Sciences (SPMS) since May 2016.
Professor Cong`s research in data science, including data management and data mining, has made significant impact in the areas of geospatial and textual data management, point of interest recommendation and user preference modelling, and mining social networks and social media. Professor Cong has been an Associate Editor of ACM Transactions on Database Systems, the most prestigious journal for data management, since 2016. He is also currently in the editorial board of Proceeding of Very Large Database (PVLDB) and International Journal of Social Network Mining. He was an Associate Editor of PVLDB from 2013 to 2014. He has been involved in organising several conferences as the PC Vice Chair, PC Co-Chair, and Publication Chair. Notably, he is the PC Vice Chair for IEEE International Conference on Data Engineering.
Intuitive Neural Avatars
Asst. Prof. Liu Ziwei
Abstract: Sensing, understanding and synthesizing the neural avatars intuitively have been a long-pursuing goal of computer vision and graphics, with extensive real-life applications. It is at the core of embodied intelligence. In
this talk, I will discuss our work in intuitive neural avatars from four different aspects: versatile perception, generalizable understanding, intuitive generation and creative animation. Our approach has shown its effectiveness and generalizability
on a wide range of tasks.
Bio: Prof. Ziwei Liu is a Nanyang Assistant Professor at School of Computer Science and Engineering (SCSE) in Nanyang Technological University, with MMLab@NTU. Previously, he was a research fellow (2018-2020) in CUHK (with Prof. Dahua Lin) and a post-doc researcher (2017-2018) in UC Berkeley (with Prof. Stella X. Yu). His research interests include computer vision, machine learning and computer graphics.
Ziwei received his Ph.D. (2013-2017) from CUHK / Multimedia Lab, advised by Prof. Xiaoou Tang and Prof. Xiaogang Wang. He is fortunate to have internships at Microsoft Research and Google Research. His works include Burst Denoising, CelebA, DeepFashion, Fashion Landmarks, DeepMRF, Voxel Flow, Long Tail, Compound Domain, and Wildlife Conservation. His works have been transferred to products, including Microsoft Pix, SenseGo, and Google Clips.
Multimodal Computational Narrative Intelligence
Assoc. Prof. Li Boyang
Abstract: Narrative Intelligence refers to the capability to craft, tell, understand, and respond appropriately to stories. It is a central component of human intelligence and a holy grail for computational mimicry. In this short talk,
I will discuss efforts and challenges in creating Computational Narrative Intelligence, especially those that start from both visual and textual data.
Bio: Dr. Li Boyang, Albert is a Nanyang Associate Professor at the School of Computer Science and Engineering, Nanyang Technological University. Before joining SCSE, he held a visiting position at the Alibaba-NTU Singapore Joint Research Institute. Prior to that, he was a Senior Research Scientist at Baidu Research USA from 2018 to 2019, and a Research Scientist and Group Leader at Disney Research Pittsburgh from 2015 to 2017. He received his Ph.D. degree in Computer Science from Georgia Institute of Technology in 2014, and his B.Eng. degree from Nanyang Technological University in 2008. He published more than 45 peer-reviewed papers in top-tier journals and conferences, and holds two US patents. His research work has been covered by influential technology media outlets such as Engadget, TechCrunch, New Scientist, and National Public Radio.
Techniques for Efficient Retrieval of Top-K Recommendations
Asst. Prof. Le Duy Dung
Abstract: The current scale of e-commerce systems requires recommender systems to improve the efficiency of both the preference elicitation phase and the recommendation retrieval phase. While there have been numerous
works on designing preference elicitation algorithms that can handle millions of users and items, efficient recommendation retrieval has only gained more attention recently due to the demand for online personalized recommendation of large scale systems.
In this presentation, I will provide an overview of recent advances for efficient retrieval of top-k recommendation, especially in a matrix factorization framework. This talk is based on our JAIR survey paper on “Efficient retrieval of of matrix
factorization-based top-k recommendations: A survey of recent approaches”
Bio: Dung D. Le is currently an Assistant Professor of Computer Science at College of Engineering and Computer Science, VinUniversity, Vietnam. He is a former senior data scientist in Ads and Personalization team, Grab Holdings Inc. and a former research scientist in School of Information Systems, Singapore Management University (SMU). He earned his PhD in Data Science and Engineering from SMU, working with Associate Professor Hady W. Lauw. In his candidature, he has been recognized with SMU Presidential Doctoral Fellowship Award (AY2018 and AY2019) and SMU PhD Student Life Award (AY2018 and AY2019). Formerly, he earned his Degree of Engineer in Mathematics and Informatics from Hanoi University of Science and Technology, in 2014.