How can SA benefit from Artificial Intelligence and Big Data Introduction Dr Craig Mudge AO FTSE Managing Partner, www.pacific-challenge.com and Research Fellow, CSIRO Science meets parliament SA, Tues March 28 2017, Parliament House 1
Outline 1. Machine learning in industrial processes 2. Automation, including software automation, 3. Jobs that are most threatened retail, driving, radiology, report writing,... 4. A call to action for South Australia. Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 2
Machine learning key part of AI - involves programming computers to learn from example data or past experience -- to allow prediction, recognising images, and discovering anomalies It has given us Siri, excellent translation of languages, self-driving cars, seeing robots, and web search. Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 3
Case study:- modelling mill throughput at Olympic Dam Premise: Statistical Analysis of historical data -- vast amount sensor data would reveal useful patterns and key influencing factors Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 4
Mill fill Results 1. Supervised machine learning gave a model fitting the sensor data 2. Optimisation gave an increase in throughput 3. Sensitivity analysis on key variables improved understanding for the metallurgists Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 5
New project:- Predictive maintenance at Olympic Dam Predictive Maintenance is a method which uses machine learning to model equipment health from historical sensor data and downtime logs. It predicts when maintenance should be performed, so promising cost savings over time-based preventive maintenance, because tasks are performed only when warranted. Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 6
Woodside LNG - Another predictive analytics example Streaming 200,000 sensors into Big Data predicting plant interruptions and maintenance issues Note: Woodside s Senior Vice President and Chief Technology Officer, Shaun Gregory, is a mathematician! Peter J Coleman, CEO, is an engineer and a Fellow of our Academy Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 7
Woodside joint work with IBM Watson IBM Watson is well suited to solving problems that involve sifting through large volumes of unstructured data. Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 Watson for Projects -- staff query 30 years of historical internal data and documents -- 33,000 in total. 320,000 documents for Watson for Maintenance Watson for Drilling -- review 20 or 30 of these well completion reports 8
Changes in jobs due to AI-enabled automation Most threatened jobs:- retail, driving, radiology, report writing, tax preparers accountants, real-estate, para legal,... Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 9
Machine-Learning-as-a-service available from the top 3 cloud services on their massive data centres accessible over the web Amazon Microsoft Google Google, Dalles Oregon Microsoft Azure, Chicago 10
Call to action for South Australia 1. Promote SA as Big Data state machine leaning, AI 2. Strengths Defence (DST Group, Edinburgh) Jackie Craig; University of Adelaide (Aust Centre for Visual Technologies) Anton van den Hengel; D2D CRC Brenton Cooper 3. Opportunities AI and Informatics in agriculture DIY data analytics is here 4. Focus, focus, focus * business Mining Equipment, Technology and Services (METS) strategy 2017 -- an example of what government can do * debate discussions/informed debate on Job Susceptibility to Automation Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 11
Thank you Dr Craig Mudge AO FTSE 0417 679 266 craig.mudge@gmail.com Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 12
Dr. Craig Mudge AO FTSE --- innovator, advocate for machine learning, Big Data practitioner craig.mudge@gmail.com 0417 679 266 Managing Partner, www.pacific-challenge.com and Research Fellow, CSIRO SUGGESTED READING Information Technology and the U.S. Workforce: Where Are We and Where Do We Go from Here?. National Academies Press, National Academy of Science, 2017. http://www.nap.edu/24649 Rise of the Robots- technology and the threat of a jobless future. Martin Ford, Basic Books, New York, 2015 The Second Machine Age work, progress, and prosperity in a time of brilliant technologies. Erik Brynjolfsson and Andrew McAfee, Norton, New York. 2014 The future of employment: how susceptible are jobs to computerisation?, C B Frey and M A Osborne, Machines and Employment Workshop, Oxford University, 2013 http://www.oxfordmartin.ox.ac.uk/downloads/academic/the_future_of_employment.pdf Dr Craig Mudge AO, Science meets parliament SA, 28.3.2017 14