Guest blog: Informed Decision Making: On the steady growth of Data Driven Prediction Machines in Society
The recent availability and continuous collection of data via various mediums such as sensors, qualitative information and general records has tied in nicely with the rise of computation power and the availability of Data Driven models. The said models of which stem from the area of Machine Learning which is a fast-growing field where models of varied configurations are fed data, iteratively tuned as part of a learning process with the core aim of being able to predict an outcome given a specific kind of input data and are commonly referred to as Prediction Machines.
These prediction machines take a variety of forms depending on their complexity and underlying Machine Learning algorithms, where a few examples of the underlying models used include classical statistical regressions, of which carry a high degree of interpretability, Discriminant Analysis and Decision trees which have greater predictive power at the cost of model transparency and interpretability, all the way towards Support Vector Machines and Artificial Neural Networks of which have greater predictive power at the trade of a much higher model complexity where at times model transparency begins to lack and therein causing for these class of models to be referred to as Black Box models. The choice of which of these models to use is largely dependent upon the application area and requirement, i.e is accuracy favoured over interpretability? Can the model handle high dimensional data by example? Can the internals of the models be visualised and decisions explained?
In various capacities, these said prediction machines are beginning to find application in various aspects of society-at least first in an R&D capacity, where they have been trialled for use in what can be regarded as a Decision Support framework. This tends to require articulation and clarification to the non-expert as the common census is to project scenarios of a prediction machine replacing a human expert and resulting in a human based redundancy. This is not the intention by the AI and Data Scientist of whom are leveraging and introducing these machines into society, instead the underpinning goal is to be able to utilise further data and information collected from sources related to the question at hand, in order for the machine to serve as an auxiliary source of information which can help make a well-rounded and informed decision making process. This combination of Human and Machine based intelligence has been termed by some as a form of “Super-Intelligence” which sees the creation of a platform which would allow for enhanced and proactive decision making which compensate for the apparent shortcomings of each of its individual constituent parts.
With the growth in computation power, mass awareness, data availability and algorithmic power which we can now harness as a society, it can be finely anticipated that these machines would become a staple in the years to come within various walks of life, where they would be augmented with human expert knowledge to reach decisions, examples of which could include the harnessing of societal data to train prediction machines to aid in resource allocation capacities to aid policy makers in preparedness for forms of societal turbulence which include wars, pandemics, recessions. The use of prediction machines to offer an additional opinion and where possible null out biases in the criminal justice system within the law, and of course in the area of clinical medicine for early diagnosis of cancers, proactive care within pregnancy medicine and also strides towards personalised medicine to name a few.
Ejay Nsugbe, Data Scientist and Julia Muraszkiewicz, Head of Sociotech Insights Group at Trilateral Research
Local Public Services Innovation: Creating a catalyst for change
techUK, in collaboration with its Local Public Services Committee, has published a new report making the case for enhanced digital innovation adoption across the UK’s local public services to improve citizens’ lives. The report, ‘Local Public Services Innovation: Creating a catalyst for change’
