Making informed decisions requires the right information to be delivered to the right people – in the right time to be usable and in the right form to be useful.
But what do we mean by right? How do we make these things right, in all our different circumstances?
Unusual Systems builds and updates information software systems that are used to inform enterprise-level decision makers. To direct these systems we research topics such as knowledge distribution and analytic techniques, and develop new research methods and new training materials to improve the outputs of the intelligence operators that inform decision makers.
Knowledge Distribution
Typically we have too much information and not enough time to assimilate it all – let alone understand the relevant aspects, extract the salient points, and condense them into a form to communicate to the decision maker.
When we work as groups, passing around what we know, the problems escalate.
Our research encompasses several domains. We study – and practice and teach – how to understand the local world from limited reports about it, and how to describe this. We explore methods to recognise, expose and mitigate our biases. We help teams understand what is involved when they work and communicate in groups.
We are assembling a catalogue of knowledge-distribution concepts and relationships: ask us about Iconoclastic Sausages and Trojan Mice.
Research Methods
Analytic Techniques: We are assembling the techniques used by intelligence agencies and scientists to assess the information they have been given. Competitive Hypothesis Testing, for example, can be a rigorous way of reviewing the evidence to distinguish between likely hypothesese, but the costs in time and cognition can be too much for many practical situations. We are preparing exercises to test these techniques for efficacy.
SBOK: Our research focus just now is on Searching the Body of Knowledge: the systems that we can build for various kinds of searches across what humans already know, in order to find existing answers. We have reviewed the methods that are currently used, and developed a systematic approach (which is not what most ‘systematic literature reviews’ describe). A paper has been submitted for publication, and a more detailed critical review of methods is under way.
Training
We create training games, exercises and online courses to help your analysts improve their assessments.
Our ‘Arnhem’ game gives students too much data and not enough relevant information for the time available, and can be used to exercise a teams’ ability to filter, assimilate, discuss and communicate uncertain assessments. Our IN.security game is being developed to help students understand the trade-offs in securing a small company against cyber attack.
We have systems for ‘re-skinning’ exercises so that we can re-use the same underlying situation by making it look and feel different to the students.
Our ‘think smart’ series provides bite-sized introductions to aspects of knowledge distribution: biases, techniques, blind spots and mitigations.
Mining Data
We have experience building distributed grid systems for astronomers, warehouses for the Home Office, data pipelines for HMRC and logistics networks for Ocado. We are currently developing systems for monitoring computer networks to spot signatures of attacking systems hiding on the network.