Software Technology Exchange Workshop, STEW, 2019

Digitalisation transforms almost all aspects of our society and cities as local authorities are no exception. Done right, we face possibilities to take on societal challenges, strengthen democracy and deliver better welfare and higher quality of life for our citizens. However, it requires us to work in completely new ways internally and to partner up with external actors from business, academia and the civil sector. The City of Malmö has a long history of working in the field of democracy and sustainability that we now draw upon in our digital transformation.

In this keynote Maria will give an overview of the ground laying and ongoing work for digitally transforming the city of Malmö – from recruitment, investments and financial models to principles on end-user focus and a new way of seeing IT as an integrated business development asset in all corners of the administration. Emphasis will be put on how we build on sustainability and cooperation.

EU’s research and development focus around Cooperative Intelligent Transport Systems (C-ITS) is aiming at improving road safety, traffic efficiency and comfort of driving, by helping the driver to take the right decisions and adapt to the traffic situation intend to which will allow road users and traffic managers to share information and use it to coordinate their actions. This calls for the need of new open frameworks for data exchange between business partners and national and regional road and traffic authorities.

In this presentation, Volvo Cars will present their work in the C-ITS area with other business partners and authorities in projects such as Nordic Way 2 and GeoSum. Software solutions in the vehicle and in the cloud for use cases such as geofencing of environmental zones, Emergency Vehicle Awareness and Slippery Road Alert will be presented. The need of and how to successfully establish open frameworks and architectural design for data exchange between new ecosystems will be discussed.

We will present a newly established project, along with its initial results, that aims at developing technical advances enabling automated driving in challenging urban scenarios. The key observation is that cooperation between vehicles is a crucial component to approach the problem of the safety for highly automated traffic in smart cities.

The project is a cooperation between Halmstad University and several industrial partners (including Scania, AstaZero, H&E Solutions, TerraNet, QuviQ, KPIT, and Gutec) and is divided into three sub-projects that address, respectively, Sensing & Localization, Communication & Collaboration, and Decision & Control issues.

The overarching part of the project is that of software testing and software integration of the sub-project developments, we will concentrate on this overarching part in this presentation. Model-Based Testing (MBT) is to be applied to enable efficient testing of components in such a system at different stages of the life-cycle: development, integration, deployment, and also run-time. One of the main challenges is the unification of modelling and testing methods towards this scenario that is inherently hybrid and also needs interfacing with simulation environments.

In this talk Jan Erik will present learnings from using computer vision to build a global world representation for cities, mapping companies, and automotive. The talk will cover some core technical innovations, crowdsourcing of data, large scale compute, open source and open data, and how to get businesses and community to work together. Example use cases from smart connected cities and autonomous vehicles will be shown together with an outlook towards what to come in this field.

Fueled by digitalization and a global smart city megatrend, cities around the world compete to become the smartest. But what is a smart city? And how do smart city initiatives make a difference for citizens like you and me? A smart city crash course combined with real-life examples of how connected devices, data and people create real community and citizen values.

Internet of Things and Smart ”this-and-that” have been buzzwords for some time, but a major challenge slows the adoption and prevents solutions, very much as in the early days of computers and mobility; the incompatibility between standards and different vendor offerings.

We present a game-changing solution that turns the mess of these vertical lock-ins into a horizontal layered open market.

You will learn how the integrated platforms from the Dutch Civity and the Swedish Sensative empowers real estate companies and municipals to take control of their digitalization and open up a volume market for smart service providers.

While Deep Neural Networks (DNN) have revolutionized applications that rely on computer vision, their characteristics introduce substantial challenges to automotive safety engineering. The behavior of a DNN is not explicitly expressed by an engineer in source code, instead enormous amounts of annotated data are used to learn a mapping between input and output. Functional safety as defined by ISO 26262 is not sufficient to match the needs for the new generation of data-driven software.

Earlier this year, ISO/PAS 21148 Safety of the Intended Functionality (SOTIF) was published by ISO. SOTIF is a Publicly Available Specification (PAS), a response to a pressing need of an automotive safety standard appropriate for machine learning. A PAS is a stepping stone toward a new ISO standard, and SOTIF is intended to complement conventional functional safety as defined in ISO 26262.

Based on SOTIF, we highlight that input camera data from environments that differ from what a DNN was trained for constitutes a major hazard. Furthermore, we argue that the hazard mitigation should result in a safety requirement mandating the inclusion of a DNN supervisor, i.e., a reject option for when input does not resemble the training data. Finally, we present results from a comparison of different DNN supervisors.

In a connected society, many software intensive systems are exposed to attacks, e.g., through exposed APIs. The use of third-party components and libraries in software development allows for faster deployment and less maintenance of the code. Reducing the amount of in-house developed code also reduces the possibilities of vulnerabilities in such code. Instead, analysis of vulnerabilities is shifted towards understanding and responding to security problems in the third-party code.

We have developed algorithms and tools, taking advantage of machine learning, for improving the analysis and understanding of security problems in third party code. Applying such algorithms can help software developers, device manufacturers, and integrators to improve the overall security in their offerings. In the talk we show how information from different sources can be combined and processed in order to reduce the time of vulnerability exposure, improve the coverage of the information, and prioritize among vulnerabilities to analyze.

Data is the basis for more and domains, particularly for machine learning applications. As a consequence, the costs for data collection and maintenance increase, leading to the risk of development budgets eaten up by data for commodity features, leaving little resources for differentiation and innovation. We therefore explore if data can be shared in a fashion, similar to open source software. We have surveyed companies for bridges and barriers to opening their data and using data from the open. We identify

1) lack of data providers and brokerage,

2) trust is important for relying on data sources,

3) needs for standardization, and

4) the role of public agencies to provide or broker data is important.

This talk will elaborate our findings and discuss what is needed to make open data commons to function in effective ways in the community.

Bishop Johan Tyrberg’s presentation will deal with the following:
1. Liberate from fear. In times of great change fear is a natural reaction. Theology has always dealt with the concept of hope and trust. This is a good base for deliverance from fear.
2. Quid est homo? What is mankind / a human being? That was a question raised in a poem about 1000 BC and is still valid. Do we have a mechanistic view of humanity and a humanistic view of machines? I want to reflect on the responsibility of being a human.
3. The problem of Evil. Can a machine be evil? What about human mistakes? Mistakes made by a machine? When scholars of law reflect on judicial issues, theologians reflect on forgiveness and atonement. Can we forgive a machine? Can a machine forgive a person?
4. We need to reflect on the difference between what we could do and what we should do, the driving force behind technological development.
5. What is “intelligence”? Why do we use that as a metaphor for machine learning?