The impacts of Data Science, Machine Learning and Artificial Intelligence on Learning and Development
Summary of a talk by Alan Bostakian, Senior Analyst, TD Bank - Future Ready project
The field is moving extremely fast and very soon all professionals will need to have some level of understanding about them. These areas help is in upskilling, instructional design, instructon, coaching, more.
Data science touches on artificial intelligence, machine learning learning, deep learning.
Data science: methodologies, tools (statistical, algebra, programming, etc.), processes and techniques for understanding data to drive insights and value from data sets. In the banking sector, examples include things like credit risk monitoring, etc.
Artificial Intelligence: development of computer systems that have a type or level of intelligence, mostly focused on problem-solving, and well as capability of learning from the environment, the user, or the stakeholder. A popular application of AI is Siri - it's trying to act like a personal assistant for your questions.
(Video on 'responsible AI)
AI is cool, but you have so many challenges - compliance, legal, and ethical challenges - to solve the problems without creating new problems.
Machine Learning is a subset of AI, it is a field of computer science that gives the computer the ability to learn from data and environments. There are three main types of machine learning:
- supervised learning, where you provide lables or tags with the data, and the computer learns to identify unlabled data
- unsupervised learning, where we don't have lables in the data we provide, where the machine tries to find patterns, clusters or similarities in the data, which it then uses to classify everything else
- reinforcement learning - eg. chess game - you provide a way for the computer to think about methods to think about the whole game based on the final result, to come up with a solution to repeat it.
- deep learning - uses several layers where the output of each layer is the input for the next layer, learning from data by processing it again and and again. An application is this: learning from a humn voice, learning to speak like him.
Proposal to create a centre of excellence for AI a DS & AI Academy - to centralize all your decision-mking on AI, combined with a centre of excellence to bring all the research together. "You are the person who needs to come up with some kind of narrative as to why something like this is necessary in your organization."
Another benefit of such a Hub - manage relationships with three entitie:
- external bodies - AI startups, external components
Business Case for DS/AI/ML Projects
Focus on the question you want answered, and wht the audience is for that answer. Do you have a strategy for using that answer? Who is the audience for that answers? What would be the 'Day 2' requirement, after it's implemented, for support - do you have the people, talent, money? What is the project management strategy - agile? waterfall?
"In these types of projects you need to keep momentum and you need to wait a bit to get results - you're not going to show results in 3-4 months."
What constitutes success? What are the KPI? You need to be able to find a mapping between project outcomes and business objectives.
Change Management and Communication
You will need to track progress, celebrate success and help people adapt. Communication is essential to promote the culture change that is necessary. So you need things like adoption strategies, reinforcement tactics, coaching, etc.
Personalization - Customization - Examples - accounting for demographics, learning history, work experience, patterns of behaviour, requests or needs, etc., and this can be done in real time.
Also: better course recommendation engines, learning acceleration, and an adaptive and responsive approach to a personal learning journey.
Content development and Instructional design - t won't replace ID, but will offer a support tool to provide better material for creating a support program. Deep learning could also help with instructional modeling, learning needs, assessment, etc.
There still a challenge here meeting the social and emotional needs of students. Still, AI/ML systems can also provide feedback and coaching (but this is mostly just to free up time for the real coach or mentor). This would also include chatbots, as well as administration/ operations support.
(Another video - video analysis of a person to detect job suitability and training recommendations)
Assessment / Proctoring / Grading
It's proving this now, but it's still improving.
(Activity - moving people into groups - numerous people left the session at this point)
(My thoughts - I was hoping for a higher-level talk here.)