Skip to main content

Challenge instructions

Instruction and application
In Progress

To strengthen your ML solution, you'll now focus on refining your approach through targeted experimentation, comparing results, interpreting model behaviour, and making evidence-based decisions about the most suitable solution.

Hackathon Resource Pack​

If you haven't already, or if you need to re-download it, download the Hackathon Resource Pack now. It contains:

  • Descriptions of the two scenarios.
  • Datasets for each scenario.
  • Hackathon 1 sample solution notebook for each scenario.
  • Working notebook for each scenario that you will use during the hackathon.
  • The presentation template you will use to capture your work throughout the hackathon. AIMLF Hackathon 2 Resource Pack.zip

Once downloaded, open the files and keep them accessible for the rest of the hackathon.

### Instructions

Targeted model improvement​

Make more deliberate improvements to your solution, such as:

  • Trying a stronger alternative model.
  • Refining features.
  • Improving preprocessing decisions.

Focus on meaningful improvements rather than large amounts of experimentation.

Compare approaches​

Compare results across:

  • Different models.
  • Different feature combinations.
  • Different evaluation results.

Identify:

  • Which approach performed best.
  • Which approach was easiest to interpret.
  • Key trade-offs between approaches.

Select and justify final approach​

Select one preferred model approach. Your justification should consider:

  • Performance metrics.
  • Trade-offs (e.g. simplicity vs performance).
  • Suitability for the business problem.
  • Practical implementation considerations.

You may use LLMs to:

  • Critique your reasoning.
  • Challenge assumptions.
  • Help refine explanations and recommendations.

Interpret model behaviour​

Analyse:

  • Where the model performs well.
  • Where the model struggles.
  • Potential causes of weaker performance.

Link findings to:

  • Business impact.
  • Operational or editorial decision-making.

Practical risk and limitation reflection​

Consider:

  • What happens when the model is wrong.
  • Where human oversight may still be needed.
  • Key limitations in the data or modelling approach.
  • What you would improve with more time or data.

Looking ahead to the wrap up workshop​

In the wrap-up workshop, you'll bring this work together by preparing and delivering your presentation, clearly communicating your solution, key decisions, and recommended next steps.

In progress