Optimising financial processes

Posted on:

Predictions & Contradictions: Artificial Intelligence, Machine Learning and Data Science

Artificial Intelligence, Machine Learning and Data Science

There is much discussion on these topics. Simply put, Machine Learning is the link that connects Data Science and AI.

AI is often confused and conflated with RPA and other technologies, but a different kettle of fish (or discipline’ if you prefer), AI has been with us for a long time now and we experience its value in our daily lives. Exploiting AI does not require us all to hire armies of PhD data scientists, rather we can exploit AI enabled services that already exist. Machine Learning, specifically, is well tuned to a service-oriented SaaS model.

We will see more AI and specifically, Machine Learning, entering enterprise business processes this year. Low hanging fruit includes improving the quality of master data, identifying errors and exceptions and streamlining processes. We are learning the nuances of the AI world and some of the dangers, such as confusing causation with correlation. The availability of large bodies of data has made Machine Learning techniques much more practical than before. With so much data to evaluate, the chances increase that random correlations are discovered by learning models and misunderstood. One such study correlates the age of Miss America and the total number of murders by steam, hot vapours and hot objects and shows a relationship. The lesson? Don’t Confuse Causation with Correlation. On a more optimistic note, there have been some great ideas developing rapidly. The idea of the ‘Generative Adversarial Network’ has been hailed as the ‘coolest thing in Deep Learning in the last 20 years’. In this scheme, two deep neural networks challenge each other to create and improve content, teaching themselves as they go along! The technique has been used to create images and music and is touted as having huge future potential. Maybe AI could develop imagination! Even better, a sense of humour?

Also, often conflated, Blockchain will get productive in business applications such as ‘ethical sourcing’ and will actually become boring this year! But getting past the frothy days of inflated expectations is always a good thing. In response to the somewhat rhetorical question “Blockchain? Or are your pulling my chain?” 31% of CFOs surveyed now feel broadly positive about its potential applications, with 61% taking the diametrically opposite view. As with Machine Learning, the trend seems to be that Blockchain enabled solutions delivered as a service will win out over single entity attempts to deliver value.