The Golden Age of Artificial Intelligence AI Improving
























- Slides: 24
The Golden Age of Artificial Intelligence (AI) Improving connections with government agencies ATO, Canberra, ACT, Australia; 18 July 2017 Adj/Prof Hanna Suominen, hanna. suominen@anu. edu. au Senior Lecturer in Machine Learning (ML) Team Leader of Natural Language Processing (NLP) 1
Outline 1. What is AI and how does it work? 2. What is the role of AI now and in the future? 3. How will AI impact the facets of our lives: e. g. economy, education, and healthcare? 2
Outline 1. What is AI and how does it work? 2. What is the role of AI now and in the future? 3. How will AI impact the facets of our lives? 3
Knowledge is Power in Information Society 4
From Data Recording to Data Analytics 5
AI is an Enabler Actionable evidence-based decisions Shedroff N (2000). An Overview of Understanding. In RS Wurman (ed. ): Information 6
AI Works By Converting Data Into Insights From your audio or video, reports , measurements, and other data Through a model To your customised insight: • • Filled out forms Visual surveillance Situational awareness Decision support 7
Outline 1. What is AI and how does it work? 2. What is the role of AI now and in the future? 3. How will AI impact the facets of our lives? 8
Trusted Analytics 9
Science and Art of AI AI research and development should be driven by evaluation, covering • • • performance measurement, behavior of the system/method, its limitations, its generalisability, and its future development. Cohen & Howe 1988. How evaluation guides AI research: The message still counts more than the medium. AI Magazine 9(4): 35. Cohen & Howe 1989. Toward AI research methodology: Three case studies in evaluation. IEEE Transactions on Systems, Man and Cybernetics 19(3): 63446. 10
Key Issue: “Trust me, it works like magic!” 11
Sound an Alarm http: //nicta-ifi. s 3 -website-ap-southeast-2. amazonaws. com Cross-sectional status Longitudinal Trends Beds/patients Textual evidence Automatic detection of patients with invasive fungal disease from free-text computed tomography (CT) scans. J Biomed Inform 2015. 12
Fill out a Handover Form http: //nicta-stct. s 3 -website-ap-southeast-2. amazonaws. com Benchmarking clinical speech recognition and information extraction: New data, methods and evaluations. JMIR 13
Protection of Human Subjects in Record Research Personal Data Careful consideration & compliance with the appropriate governance, policy, and legal frameworks Prerequisites for International Exchanges of Health Information for Record Research: Comparison of Australian. . . Medinfo 2017. 14
International Policies 15
Process of Getting Data Plans & policy study Ethics approvals Research permission s Data collection & deidentification Research Data destruction 16
Trustworthiness and Accountability Scientific Evidence Practice-based Evidence Actionable evidence-base for clinical judgement and decision-making Applying language technology to nursing documents: pros and cons with a focus on ethics. Int J Med Inform 2007. 17
Disruptive Technology and Power Structures 18
Outline 1. What is AI and how does it work? 2. What is the role of AI now and in the future? 3. How will AI impact the facets of our lives? 19
Sound an Alarm http: //nicta-ifi. s 3 -website-ap-southeast-2. amazonaws. com Cross-sectional status Longitudinal Trends Beds/patients Textual evidence Automatic detection of patients with invasive fungal disease from free-text computed tomography (CT) scans. J Biomed Inform 2015. 20
Healthcare 21
Education: Jobs Explorer http: //jobs. t 3 as. org • Annual quantification on all AU job ads of – which industries and geolocations absorb Ph. D, – what are their most soughtafter skills, & – how are these posts characterised • Pilot funded by AU Gov Dep of Industry, Innovation, and Science • Invited speech on the capitol hill to Hon Greg Hunt MP and the Group of Eight → Discovery Translation Fund 22
Business: Compliance Auditing 23
Thank You! 24