Creating an Effective and Enjoyable Learning Experience

Dr Arron Lacey

A little bit about me…

• 2021: Senior Community Manager at the Alan Turing Institute
  • Work on the EDoN Initiative
• 2017:2021: Lecturer in Machine Learning at Swansea University
  • Now an Honorary Lecturer
• 2011:2014 Epidemiologist at the SAIL Databank
  • Mainly epilepsy and cardiology

About this talk …

• We learn best through doing, not just listening.

• We will focus on just two themes: Achievement and Action.

• Achievement
• Action

About this talk …

• Achievement
  • Expectations
  • Understand your audience
  • Relevance
  • Accessible Content

• Action
  • Instant Gratification
  • Learning Environment
  • Collaboration
  • Future Learning

Expectations

• Onboard the class - we do it in new jobs so why not in a learning environment?

• Provide a roadmap for the session/course/workshop
  • The class knows what to expect and to prepare

• You will be able to do / understand this at the end

Understand your audience

• Course participants will already have their own experience
  • we want to build on that, not teach them what they already know

• Assess their motivation and needs
  • Try a pre-survey to get participant experience and tailor your delivery.

• Encourage constant feedback
  • Listen and pivot if needed
  • Direct the audience to what kind of feedback will be helpful
  • If I learn someone is a cardiologist I will try throw in some relevant examples later in the course

Relevance

• No vague tutorials - embrace abstraction

• Think of real-world challenges that students might face

• Is this a health related course?
  • Start scouring the internet for healthcare datsets to use!
  • Although it’s ok to use a really out there example to keep things fresh

Abstraction

Abstraction is a powerful tool to help remember concepts - but are foo, bar and baz?

# Function definition is here
def foo( arg1, arg2 ):
   # Divide and return"
   bar = arg1 / arg2
   return bar;

# Now you can call sum function
bar = foo( 10, 20 );

Abstraction

Use abstraction in function and variable names

# Function definition is here
def epilepsyIncidence( numEpilepsy, Population ):
   # Divide and return"
   RateRatio = (numEpilepsy / Population)*1000
   return RateRatio;

# Now you can call sum function
IncidenceRate = epilepsyIncidence( 10, 5145 );

print "The Incidence Rate Ratio for Epilepsy is : ", IncidenceRate, "per 10,000"
"The Incidence Rate Ratio for Epilepsy is : 19.4 per 10,000"

Abstraction

Action

• Absorbing knowledge is important, but acting on it re-enforces learned material

• We should aim to minimize the time between learning a concept and practicing it.

• Try to jump into an interactive Minimal Working Example (MWE) as often as possible

• I try to stick to 20 minutes theory with MWEs followed by 30 mins practical

Instant gratification

• Instant gratification is a powerful vehicle for learning
  • Even after 15 years of coding I celebrate when my code works!

• (Easy) Practice examples also allows us to fail (gracefully) and learn from mistakes
  • This is why I can’t do DIY…

Learning Environment - build on MWEs

• A consistent learning environment that is easy to use

• Keeps a tight focus for you to control the environment
  • I recommend open source cloud based solutions where possible
  • Kaggle, Colab, Binder etc

• Can self assess and understand how they are learning

Kaggle for Data Science and Machine Learning

• My favourite cloud solution for teaching Data Science and ML
• Dockerized notebooks
• Free GPU
• Access to many and huge datasets (including those hairy beasts)
• Version control

Collaboration

• Set the tone and expectations for collaborative sessions
  • Code of conduct
  • Set clear goals for sessions

• Try and pair up participants that will learn from eachother

• Some ideas:
  • Discuss the benefits and potential harms of using digital technology to attempt to detect dementia 20 years early
  • Create a COVID Dashboard that tracks new cases by variant over time
  • GitHub Pull Request Practice

Future Learning

• Move the participants towards independence in the topic

• Set them up to succeed
  • Recommend learning materials that will build on the course and challenge them
  • Tailor by discussing their motivations for enrolling on the course/workshop

Introduce relevant GitHub projects

Thank you for your time!

Slides made with Quarto + revealjs