International workshop on Artificial Intelligence 2022
Workshop: September 26-27, 2022
Venue: Japan Advanced Institute of Science and Technology
Workshop Program
September 26, 2022
- 10:00-10:10:
- Prof. Nguyen Le Minh: Opening session
- 10:10-10:40:
- Prof. Tomoko Matsui (ISM): Spatio-temporal data analysis methods and their
applications
Abstract: I'll introduce statistical methods and recent applications for spatiotemporal
data analysis in this talk. With Kepler's research on planetary motion, spatial-temporal data
analysis has been researched as far back as the 17th century. The research in this area has
grown significantly over the past ten years since spatio-temporal data makes up the majority
of big data. The statistical models applied in those investigations are briefly explained.
Applications will be discussed, with examples including spatiotemporal analysis of tweet data
and extreme weather events for urban risk management and spatiotemporal study of heat waves in
urban locations utilizing the Tukey g-and-h random field model.
- 10:40-11:10:
- Prof. Bui Thu Lam (ACT): Artificial Intelligence and Applications in Cyber
Security
Abstract: Cyber security is a fasnating field of research with a wide range of problems
involved with data analysis and prediction. That is why the applications of aritificial
intelligenc in this field has been a hot topic. In this talk, I will cover the recent
development in using AI for cyber security applications, such as deep learning and adversarial
machine learning. Furthermore, the relationship between machine learning and cryptography will
also be covered in this talk and how to protect machine learning models with detailed
examples.
- 11:10-11:40:
- Prof. Yuji Matsumoto (RIKEN Center for Advanced Intelligence Project) (online):
Information Extraction from Scientific Literature
Abstract: This talk will introduce our recent research activities for information
extraction from scientific papers, which some successive results of our
two government-funded research projects. The first is for overall
technologies for scientific document analysis and applications, and the
second is for relation and process information extraction in the
material science domain. After introducing some results on process
extraction in the material science domain, I will point out some
difficulties and issues we experienced that should be conducted for
further development.
- 11:40-12:10:
- Dr. Phan Viet Anh (LQDTU): AI technologies to support blind people in information
access
Abstract: Visually impaired or blind (VIB) people have faced numerous difficulties in
the daily activities. Providing the ability to access text information will enable them to
independently exploring the surrounding environment via reading books, and other printed
documents. This talk will demonstrate the applications of AI in image, language, and speech
processing to assist VIB people in information access via smartphones. I also analyze the main
challenges and our solutions to build high-performance AI models that can process low-quality
input data made by VIB people.
- 12:10-14:00:
- 14:00-14:40:
- Prof. Ken Satoh (NII) (online): Overview of the project "Advanced Reasoning
Support for Judicial Judgment by Artificial Intelligence" and judgement phase support
Abstract: We provide an overview of the achievement of the project of "Advanced
Reasoning Support for Judicial Judgment by Artificial Intelligence" supported by JSPS (Japan
Society for the Promotion of Science) and especially report support for judgement phase in
civil litigation.
- 14:40-15:10:
- Prof. Satoshi Tojo (JAIST): Quantum Knowledge Representation
Abstract: Quantum encryption is known to be very secure since if it is interfered by a
third party, the original information cannot be restored. In this lecture, I begin from the
fundamentals of quantum mechanics, and then expound quantum computation, including quantum
logic. The quantum computation lies in several mysterious notions such as qubit,
superposition, entanglement, and so on. Among which, I especially consider the superposition
and discuss how it can be applied to knowledge representation in AI.
- 15:10-15:40:
- Prof. Mizuhito Ogawa (JAIST): Formal semantics exraction from instruction set
manual
Abstract: For developing binary code tools, understanding the formal semantics of
instruction sets is needed. There are various instruction sets, especially for IoT devices. In
this presentation, we will briefly introduce our recent trial of auttomatic formal sematics
extraction by NLP on manuals, which comes during the development of symbolic execution tools
for analyzing malware, e.g., BE-PUM for x86, SyMIPS for MIPS, and CORANA for ARM.
- 15:40-16:10:
- Dr. Tran Duc Vu (ISM): COVID-19 Scenario Planning & Forecasting With Tweet
Analysis
Abstract: We sought clues to the complicated COVID-19 progression by investigating the
relationship between reactions on social media and the COVID-19 epidemic in Japan. Analysis
using Japanese Twitter data suggested that reactions on social media and the progression of
the COVID-19 epidemic may have a close relationship. We propose using observations of the
reaction trend represented by tweet counts and the trend of COVID-19 epidemic progression in
Japan and a deep neural network model to capture the relationship between social reactions and
COVID-19 progression and to predict the future trend of COVID-19 progression.
September 27, 2022
- 10:00-10:30:
- Dr. Vu Viet Anh (Cambridge - UK): Cybercrime Underground in the
Russia-Ukraine Conflic
Abstract: There has been substantial commentary on the role of cyberattacks,
hacktivists, and the cybercrime underground in the Russia-Ukraine conflict. Drawing on a range
of data sources, we argue that the widely-held narrative of a cyberwar fought by committed
'hacktivists' linked to cybercrime groups is misleading. We collected 281K web defacement
attacks, 1.7M reflected DDoS attacks, and 441 announcements (with 58K replies) of a volunteer
hacking discussion group for two months before and four months after the invasion. To enrich
our quantitative analysis, we conducted interviews with website defacers who were active in
attacking sites in Russia and Ukraine during the period. Our findings indicate that the
conflict briefly but significantly caught the attention of the low-level cybercrime community,
with notable shifts in the geographical distribution of both defacement and DDoS attacks.
However, the role of these players in so-called cyberwarfare is minor, and they do not
resemble the 'hacktivists' imagined in popular criminological accounts. Initial waves of
interest led to more defacers participating in attack campaigns, but rather than targeting
critical infrastructure, there were mass attacks against random websites within '.ru' and
'.ua'. We can find no evidence of high-profile actions of the kind hypothesised by the
prevalent narrative. The much-vaunted role of the 'IT Army of Ukraine' co-ordination group is
mixed; the targets they promoted were seldom defaced although they were often subjected to
DDoS attacks. Our main finding is that there was a clear loss of interest in carrying out
defacements and DDoS attacks after just a few weeks. Contrary to some expert predictions, the
cybercrime underground's involvement in the conflict appears to have been minor and
short-lived; it is unlikely to escalate further.
- 10:30-11:00:
- Assist. Prof. Teeradaj Racharak: Explanation in AI (not only ML): From KRR to
ML and Beyond
Abstract: Explainable AI is not a new topic. The earliest work on Explainable AI could be
found in the literature published 40 years ago, where expert systems explained their results via
the applied rules or by backtracking the reasoning. Since AI research began, scientists have
argued that intelligent systems should explain the AI results, mostly when it comes to
decisions. In this talk, I begin from knowledge representation formalisms in AI, and traditional
machine learning approaches to the latest progress in the context of modern deep learning, and
then describe the major research areas and the state-of-the-art approaches in recent years.
Three main topics will be covered in this lecture: (1) interpretability and explainablity in
logic-based AI, (2) interpretability and explainbility in machine learning, and (3)
interpretability and explainability in hybrid AI that synergises logic with machine learning.
The talk ends with a discussion on challenges and future directions.
- 11:00-11:30:
- Dr. Nguyen Ha Thanh (NII): How incorporating neural networks and symbolic
reasoning can improve the quality of legal decision-making?
Abstract: With the development of pretrained language models, the legal domain has seen
a recent surge of interest in using machine learning to support legal decision-making.
However, while these models can provide some insights, they are often limited by their
reliance on statistical methods and lack of understanding of the underlying legal reasoning.
We propose a framework that combines the strengths of neural networks and symbolic reasoning
to improve the quality of legal decision-making. Our framework first uses a neural network to
learn the relevant concepts from a training set of cases. The concepts are then represented as
symbolic rules, which are used by a symbolic reasoning system to predict the outcomes of the
cases. The reasoning system can also provide justifications for its predictions and improve
the transparency of the system.
- 11:30-11:40:
- Prof. Satoshi Tojo: Closing session
Workshop Chairs
Prof. Nguyen Le Minh, Japan Advanced Institute of Science and Technology
Prof. Satoshi Tojo, Japan Advanced Institute of Science and Technology
Prof. Mizuhito Ogawa, Japan Advanced Institute of Science and Technology
For any inquiry concerning the workshop, please send it to "nguyenml[at]jaist.ac.jp"