opendataval.dataval.InfluenceFunction

class opendataval.dataval.InfluenceFunction(*args, **kwargs)

Influence Function Data evaluation implementation.

TODO it may be useful to compute gradients of the validation dataset in batches to save time/space. TODO H^{-1} implementation, Current implementation is for first-order gradients

References

[1]

P. W. Koh and P. Liang, Understanding Black-box Predictions via Influence Functions, arXiv.org, 2017. https://arxiv.org/abs/1703.04730.

Parameters

grad_argstuple, optional

Positional arguments passed to the model.grad function

grad_kwargsdict[str, Any], optional

Key word arguments passed to the model.grad function

__init__(*grad_args, **grad_kwargs)

Methods

__init__(*grad_args, **grad_kwargs)
evaluate(y, y_hat)	Evaluate performance of the specified metric between label and predictions.
evaluate_data_values()	Return influence (data values) for each training data point.
input_data(x_train, y_train, x_valid, y_valid)	Store and transform input data for Influence Function Data Valuation.
input_fetcher(fetcher)	Input data from a DataFetcher object.
input_metric(metric)	Input the evaluation metric.
input_model(pred_model)	Input the prediction model with gradient.
input_model_metric(pred_model, metric)	Input the prediction model and the evaluation metric.
setup(fetcher[, pred_model, metric])	Inputs model, metric and data into Data Evaluator.
train(fetcher[, pred_model, metric])	Store and transform data, then train model to predict data values.
train_data_values(*args, **kwargs)	Trains model to compute influence of each data point (data values).

Attributes

Evaluators
data_values	Cached data values.