Multimodal aging clock

Repository Link

https://github.com/danikolbe/opencampus-pytorch-agingclock

Description

Existing biological aging clocks predominantly operate on single data modalities and are trained to predict chronological age rather than health outcomes. While the UK Biobank provides unprecedented multimodal data (clinical, proteomic, metabolomic, blood markers, anthropometry) on the same individuals, no publicly available model comprehensively integrates these modalities. This project develops a multimodal aging clock.

Task Type

Regression

Results Summary

Best Model Performance

• Best Model: [Name and type of the best-performing model"]
• Evaluation Metric: [Primary metric used, e.g., Accuracy, F1-Score, MSE, MAE]
• Final Performance: [Best score achieved, e.g., 95% accuracy, F1-score of 0.87, MSE of 0.12]

Model Comparison

• Baseline Performance: [Baseline model performance for comparison]
• Improvement Over Baseline: [Quantitative improvement, e.g., "+12% accuracy", "25% reduction in MSE"]
• Best Alternative Model: [Second-best model and its performance]

Key Insights

• Most Important Features: [Top 3-5 features that drive model performance]
• Model Strengths: [What the model does well]
• Model Limitations: [Known limitations and failure cases]
• Business Impact: [Practical implications of the model performance]

Documentation

1. Literature Review
2. Dataset Characteristics
3. Baseline Model
4. Model Definition and Evaluation
5. Presentation

Cover Image