Public Funding for Diagonal Works Ltd

To tackle the climate crisis, build resilient cities, and inclusive communities - we need our cities to change in bold ways. We have never had more information about the built environment than today: transport systems, buildings, populations, air quality, and other city systems can all be modelled digitally. Yet, all this data is not enabling the scale or pace of change needed. While technology is not the whole solution, we currently lack the right tools to make use of this information effectively. Existing tools struggle to handle analysis at the intersections of city systems. Diagonal has reimagined the foundations of a geographic information system (GIS). We are not building one tool to replace all GIS tools. Rather, our technology is an important advancement in data scientists' toolbox, to enable analysis of interactions between city systems - at scale. We have built technology that prepares, processes, and manages big data sets so data scientists can get to work on geospatial analysis. We have built a custom data processing framework that combines in-memory (RAM) geospatial data storage with geospatial-specific analysis functions: including network traversal and geographic search. This technology is new data infrastructure. We have built our technology for scale, so that data size doesn't slow down data scientists. Because our data model is in-memory, data scientists can run analysis that isn't feasible with file-based, or database storage. We want to bring the same tooling to geospatial analysis, which exists in non-spatial fields. Our combined data management and processing tooling enables machine learning, advanced analytics, and large system-of-systems analysis in city planning domains. With the funding from the Fast Start Innovation grant, we will be able to accelerate the publishing of our technology core as open source software. We believe there are many benefits to building our business around an open-core model. At the forefront, we believe that evidence which leads to change in the public realm should be scrutinizable and reproducible.

This challenge seeks to develop Privacy Enhancing Technologies (PETs) that may unlock the sharing of sensitive data between health care providers, and use a federated learning approach to model infection risk. Traditional machine learning techniques can result in a model that embeds sensitive data. Federated learning approaches struggle to promote learning from weak signals that occur in a single node to the aggregate model. Pandemic forecasting relies on inferring from weak signals within sensitive data, making the problem challenging. Our approach will focus on prediction based on individuals' activity and location patterns. We propose shifting the privacy burden to the feature engineering stage. This allows the use of simpler approaches to federated learning that need only handle privacy preserving features, rather than provide privacy preservation themselves. We will do this by applying two key techniques: first through Homomorphic encryption, a cryptographic technique that allows computations to be performed on encrypted data. So that data passed between nodes and a central server are encrypted. Second, we will introduce noise into the system to mask the impact of an individual. The purpose of adding noise is to thwart attempts to discover data about an individual, a concept mathematically defined as _differential privacy_. We will take a place-centric approach to this challenge. The features that we will focus on will be the locations of infected individuals. While our approach is implemented in terms of geospatial data, it could trivially be adapted to produce features based on other graph data, for example, the population contact graph rather than the place visitation graph.