Workshop Program

October 28th, 2019
10:30 - 10:35 Opening
10:35 - 11:35 Quantitative Information Flow – An Introduction
Hiroyuki Seki (Nagoya University)
11:35 - 12:00 Trading Location Data with Bounded Personalized Privacy Loss
Shuyuan Zheng (Kyoto University), Yang Cao (Kyoto University), and Masatoshi Yoshikawa (Kyoto University)
12:00 - 13:30 Lunch
13:30 - 14:30 Update Propagation Over a Network: Multi-ary Delta Lenses, Tiles, and Categories
Zinovy Diskin (McMaster University)
14:30 - 15:00 Towards a Complete Picture of Lens Laws
Keisuke Nakano (Tohoku University)
15:00 - 15:30 Break
15:30 - 15:55 Complexity Results on Register Pushdown Automata
Ryoma Senda (Nagoya University), Yoshiaki Takata (Kochi University of Technology), and Hiroyuki Seki (Nagoya University)
15:55 - 16:20 Analyzing Trade-offs in Reversible Linear and Binary Search Algorithms
Hiroki Masuda (Nanzan University) and Tetsuo Yokoyama (Nanzan University)
16:20 - 16:45 Toward a view-based data cleaning architecture
Toshiyuki Shimizu (Kyoto University), Hiroki Omori (Kyoto University), and Masatoshi Yoshikawa (Kyoto University)
16:45 - 17:10 Toward Co-existing Database Schemas based on Bidirectional Transformation
Jumpei Tanaka (SOKENDAI), Van-Dang Tran (SOKENDAI / National Institute of Informatics), Hiroyuki Kato (National Institute of Informatics), and Zhenjiang Hu (Peking University / National Institute of Informatics)
17:10 - 17:35 Distributed transaction management for P2P-based update propagation
Makoto Onizuka (Osaka University), Yusuke Wakuta (Osaka University), Yuya Sasaki (Osaka University),and Chuan Xiao (Osaka University)
17:35 - 17:40 Closing
(Reception of Internetware 2019)


[Invited Talk] Quantitative Information Flow – An Introduction

Quantitative information flow (QIF) is the amount of information from secret inputs to observable outputs. QIF is usually defined as the mutual information between the secret input and observable output of a program by regarding the program as a communication channel. To compute QIF, we need to count the possible input and output pairs consistent with the observed output. In fact, QIF computation is PSPACE-complete even for non-recursive Boolean programs. In some cases, however, we can compute QIF efficiently and I would like to share such clever computation methods with the audience.
Hiroyuki Seki (Nagoya University, Japan)
Hiroyuki Seki is a Professor at Graduate School of Informatics, Nagoya University. He received his Ph.D. degree from Osaka University in 1987. He joined Nara Institute of Science and Technology in 1993, where he was a Professor during 1996 to 2013. He co-founded TTATT (International Workshop on Trends in Tree Automata and Tree Transducers) with Prof. Keisuke Nakano of Tohoku University in 2012. Hiroyuki Seki’s research interests include formal language theory and formal approach to software development.

[Invited Talk] Update Propagation Over a Network: Multi-ary Delta Lenses, Tiles, and Categories(updated: Oct. 15)

Synchronization of inter-related models (files, documents, datasets, software and domain models, or your favorite artifacts) distributed over a network is ubiquitous in data management, software development, system engineering, and our lives. Delta lenses is an established mathematical framework for model synchronization via update propagation initially proposed for the binary case (bidirectional update propagation or bx). In the talk, we will first discuss why having binary delta lenses is not enough to manage multi-ary (more than 2 models) synchronization (further referred to as mx), and we thus need multi-ary delta lenses (mx-lenses). Then we discuss mx-lens composition to cover the entire network of models to be synchronized. Finally, we will also discuss basic ideas of building more usable delta-based framework for update propagation, and why “pure” mx-lenses need a whole set of enrichments. Although the notions of mx-lenses and their composition are essentially categorical, no knowledge of category theory is required for the talk: all formalities will be hidden in diagrams and their tiling. (updated: Oct. 15)
Zinovy Diskin (McMaster University, Canada)
Zinovy Diskin is Senior Research Scientist at the McMaster Center for Software Certification and Adjunct Professor at McMaster University. He has 25+ years of experience in building mathematical models for artifacts, languages and processes in software engineering, databases and businesses including automotive, e-commerce and banking in Canada, US, and Latvia. Amongst his and coauthors most known and used models are Delta Lenses for specifying model synchronization via change propagation, DP-Graphs for diagrammatic specification of constraints and queries in data and behaviour modelling, a span-based semantics for relational domain modelling and its implementation in the language Clafer invented in Waterloo, and WorkflowPlus modelling for safety assurance. He has published extensively on category theory foundations for MDE and model management, and mathematics of model consistency, synchronization and transformation. He is a corecipient of the 10-year most influential paper award from J. of Software and System Engineering (2018). He holds Masters in Mechanical Engineering and Research from Bryansk State Technical University (Russia), PhD in Mathematics from Omsk State University (Russia), and Dr. Math from the University of Latvia.