Dimensions overall score 7.0

GitHub Code Pulse

No public code linked for this paper yet.

• LivestockFL is the first federated learning framework specifically designed for livestock growth prediction.method 95%
• The proposed federated learning framework enables collaborative model training across distributed farms without sharing raw data, preserving data privacy.technical 90%
• The framework alleviates data sparsity, particularly for farms with limited historical records.result 85%
• The neural architecture employs a Gated Recurrent Unit (GRU) combined with a multilayer perceptron (MLP) to model temporal growth patterns.method 90%
• A novel personalized federated learning framework (LivestockPFL) extends the base FL framework with a personalized prediction head trained on each farm's local data.method 90%
• Existing biophysical models have difficulty in parameterising at scale due to requirements for precise inputs that farmers often lack.limitation 85%
• The model predicts both the mean and variance of future weights to quantify prediction uncertainty, using a Gaussian Negative Log-Likelihood loss function.technical 90%
• Experiments on a real-world dataset demonstrate the effectiveness and practicality of the proposed approaches.result 80%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card