NeuralTSNE

NeuralTSNE implements a parametric version of t-distributed stochastic neighbor embedding (t-SNE) using neural networks for dimensionality reduction of molecular dynamics data.

Key Features:

• Parametric t-SNE Implementation: NeuralTSNE employs neural networks to implement parametric t-SNE, which has demonstrated superior performance in dimensionality reduction compared to standard t-SNE.
• PyTorch-based Architecture: The implementation uses the PyTorch library and the PyTorch Lightning framework.
• Flexible Usage Modes: NeuralTSNE can be imported as a module or used from the command line.

Scientific Applications:

• Molecular Dynamics Data Analysis: NeuralTSNE analyzes large volumes of high-dimensional molecular dynamics data to gain insight into hidden information encoded in molecular dynamics trajectories.
• Dimensionality Reduction: NeuralTSNE reduces the dimensionality of molecular dynamics datasets to facilitate visualization and interpretation of complex simulation data.

Methodology:

NeuralTSNE implements parametric t-SNE using neural networks with the PyTorch library and PyTorch Lightning framework.