ML Predict: Regression using Feature Input

This floe predicts properties of small, drug-like molecules using a pretrained ML model.

It runs a TensorFlow-based fully connected neural network regression model for prediction. This user-provided model can be generated using the ML Build: Regression Model with Tuner using Feature Input Floe. Every molecule needs user-provided features as float vectors for inputs.

This floe uses a convex box approach for domain of application predictions. The input TensorFlow dataset also contains a model agnostic system to explain the predictions on the molecule.

It is very cheap and quick, costing a few cents for property predictions of 50 molecules.

Outputs:

Failure Data: The molecule (a) is too large or too small or (b) has an unknown atom.

No confidence Data: The molecule’s property falls out of scope of the training set. In this case, the model predicts with no guarantees. The explainer image has a red background.

Success Data: The molecule falls (a) within scope and the explainer has a green background or (b) at the edge of scope and the explainer has a yellow background.

Molecules outside the scope of the training set will be sent to the “No Confidence” port, as a prediction cannot be considered reliable. Specifically, the scope is defined as a range in molecular weight, atom count, polar surface area, and calculated logP from the training set molecules. These ranges are reported in the Floe Report.

Inputs

Name	Description	Type
Input Small Molecule(s) Dataset to Predict Property of	The dataset(s) to read records from.	Molecule Dataset
Input TensorFlow Model	Machine learning model to predict property.	Machine Learning TensorFlow Model Dataset

Machine Learning Model Options

Name	Description	Type
Model ID of Which TensorFlow Model to Use to Predict	Which model to select. Make sure this matches the input model ID.	Int
Preprocess Molecule	For every molecule, stores only largest component, adjusts ionization to neutral pH.	Bool
Apply Blockbuster Filter	Apply Blockbuster filter.	Bool
Number of Features to Explain	Number of top features to provide LIME explanations for.	Int

Explanation and Validation

Name	Description	Type
Property Validation Field	If the dataset has a baseline, the floe reports a comparison between predictions in Floe Report.	Float
Custom Feature	Field containing feature vector to train model on.	FloatVec

Outputs

Name	Description	Type
Output Property	Output dataset to write to.	Dataset
Failure Property	Output dataset to write to.	Dataset