Roberta Rizzo (PSG Chairwoman Trenitalia), Luca Mariorenzi (PES Chairman Trenitalia) presented the context of the UIC innovation workshops within the Passenger Service Group activities. Previous innovation workshops show the current research and implementation state-of-the-art in blockchain technology in transport. At the invitation of the UIC Passenger Department, 20 participants representing DB, CFF-SBB, Hitrail, OBB, SNCB, SNCF, SZ, Trenitalia, and UIC gathered in Paris for a day’s meeting.
Firstly a historical view on Artificial Intelligence was presented, from mimicking the human intelligence in the 1950s (LISP and Eliza languages) to machine learning in the 1990s. Today it is much more efficient for machines to learn what humans teach them. AI is about problem solving and improving by learning. Today, Automated Machine learning (AutoML), where the machine chooses the best model to learn from, is the latest technology offered as SaaS by GAFAs.
Stefano Scarsi (EY) then presented the different transport fields of application such as customer analytics, revenue management, predictive models of delays, passenger flow analysis, mobility demand, chatbots, voice and text assistants.
Marie Zeiss (DB) presented Deutsche Bahn’s latest development in three different fields: natural language with DB robot SEMMI which speaks multiple languages and answers easy questions; image recognition for snow clearing on platforms or CCTV passenger face anonymisation for data privacy, reinforcement learning (an area of machine learning) to help prioritise trains on a jammed mass transit network, taking into account the number of passengers.
Nathalie Kuster (CFF-SBB) presented Swiss Railways’ challenges with a network involving a train every two minutes, where a delay of three minutes brings chaos and a delay of five minutes is unacceptable. An initial IA project involves a train, staff, track schedules and timetables for the whole country for the whole year. Another IA study presented by Mrs Kuster was intelligent cameras deployed on a locomotive to monitor tracks for a very accurate location and to check that no humans are on the train tracks.
In the afternoon, the participants worked in small groups with lively sessions and proposals to improve future collaboration on AI and other innovation topics.
The workshops will continue in 2020 on 5G, Internet of Things. All railways experts are very welcome to participate.