Instructions for Designing Your Experiments and Creating a Motivation Example

[pooyanjamshidi]

Specific Task:

You are building a federated learning system for processing patient journals and emotion-related data. Your key challenge is maintaining privacy while correlating text data to emotional ratings.

Experiment Design:

• Federated Setup: Start by simulating a small-scale federated learning system. Use journals and surveys from a couple of test cases, and evaluate how well the system learns emotional correlations over time.
• Baseline Comparison: Compare results from the federated model vs. centralized training to see if privacy-preserving methods maintain accuracy while protecting data.

Motivation Example:

Generate a plot that shows how your federated learning model’s predictions evolve over time. For instance, show how the emotion "sad" is consistently identified over journal entries and how this builds up into useful insights for psychiatric evaluation.

Evaluation Focus:

• Metrics like emotion classification accuracy, privacy scores, and federated model convergence over time compared to a centralized baseline.
• Highlight how this system helps psychiatric evaluation and speeds up diagnoses based on patient data.

[jdubarry]

Deliverable 1.pdf

General rundown of what we discussed today