Abstract

Abstract: Manufacturing companies often only have very limited visibility into their highly complex and global supply networks. While first tier suppliers are known, suppliers on sub-tiers (and their inter-dependencies) remain unknown. This obscures risks and might lead to a false sense of security. Furthermore, more complete models of the supply network would help predict and possibly mitigate supply chain disruptions. The aim of this workstream (within the research project VIPr) is to develop a system that can reconstruct any company’s supply network from public data. The DIAL seminar will cover the motivation for this project as well as show our current approach and first results. There will also be time to discuss limitations and ideas for future development. Methods we use are Natural Language Processing, Machine Learning and Human Computation (Amazon Mechanical Turk). This seminar is an interim presentation with the objective of scrutinising our approach and collecting feedback.

Speaker: Kshitij Kumar is an undergraduate Computer Science Engineering student from BITS Pilani in India. He joined DIAL as a visitor for two months to work on the VIPr research project under the supervision of Pascal Wichmann and Dr Philip Woodall. He has a background in Data Mining and Machine Learning.