TEMPO: A Hybrid-Neuromorphic Pipeline for Efficient Multi-Omic Cancer Data Integration

TEMPO is a professional-grade, energy-efficient Python package designed for multi-omic cancer data integraton. Using an innovative hybrid ANN-SNN hybrid model and advanced features, TEMPO has drastically reduces computational overhead (Synaptic Operations and FLOPs) while maintaining state-of-the-art classification performance on highly imbalanced pan-cancer datasets.

Key Features

• Multi-Omic Integration: Seamlessly aligns disparate modalities (e.g., mRNA expression, Proteomics, CNV, Clinical data).
• Neuromorphic Efficiency: Leverages Leaky Integrate-and-Fire (LIF) neurons to minimize computational power.
• Smart Routing & Gating: Implements Gumbel-Softmax modality routing and Fast Fourier Transform (FFT) semantic addressing.
• Scikit-Learn Style API: Clean, intuitive .fit() and .predict() wrapper built over PyTorch execution blocks.
• Imbalance-Aware: Built-in balanced Focal Loss and automated sample weighting to handle rare cancer subtypes.

Installation

To install TEMPO locally in editable development mode, navigate to your root package directory and run:

pip install -e .