A library for data valuation.

pyDVL collects algorithms for Data Valuation and Influence Function computation.

Refer to the Methods page of our documentation for a list of all implemented methods.

Data Valuation for machine learning is the task of assigning a scalar to each element of a training set which reflects its contribution to the final performance or outcome of some model trained on it. Some concepts of value depend on a specific model of interest, while others are model-agnostic. pyDVL focuses on model-dependent methods.

Comparison of different data valuation methods on best sample removal.

The Influence Function is an infinitesimal measure of the effect that single training points have over the parameters of a model, or any function thereof. In particular, in machine learning they are also used to compute the effect of training samples over individual test points.

Influences of input points with corrupted data. Highlighted points have flipped labels.

To install the latest release use:

$ pip install pyDVL

You can also install the latest development version from TestPyPI:

pip install pyDVL --index-url https://test.pypi.org/simple/

pyDVL has also extra dependencies for certain functionalities (e.g. influence functions).

For more instructions and information refer to Installing pyDVL in the documentation.

In the following subsections, we will showcase the usage of pyDVL for Data Valuation and Influence Functions using simple examples.

For more instructions and information refer to Getting Started in the documentation. We provide several examples for data valuation (e.g. Shapley Data Valuation) and for influence functions (e.g. Influence Functions for Neural Networks) with details on the algorithms and their applications.

For influence computation, follow these steps:

1. Import the necessary packages (The exact packages depend on your specific use case).

   import torch
   from torch import nn
   from torch.utils.data import DataLoader, TensorDataset
   
   from pydvl.influence.torch import DirectInfluence
   from pydvl.influence.torch.util import NestedTorchCatAggregator, TorchNumpyConverter
   from pydvl.influence import SequentialInfluenceCalculator

2. Create PyTorch data loaders for your train and test splits.

   input_dim = (5, 5, 5)
   output_dim = 3
   train_x = torch.rand((10, *input_dim))
   train_y = torch.rand((10, output_dim))
   test_x = torch.rand((5, *input_dim))
   test_y = torch.rand((5, output_dim))
   
   train_data_loader = DataLoader(TensorDataset(train_x, train_y), batch_size=2)
   test_data_loader = DataLoader(TensorDataset(test_x, test_y), batch_size=1)

3. Instantiate your neural network model.

   nn_architecture = nn.Sequential(
     nn.Conv2d(in_channels=5, out_channels=3, kernel_size=3),
     nn.Flatten(),
     nn.Linear(27, 3),
   )

4. Define your loss:

   loss = nn.MSELoss()

5. Instantiate an InfluenceFunctionModel and fit it to the training data

   infl_model = DirectInfluence(nn_architecture, loss, hessian_regularization=0.01)
   infl_model = infl_model.fit(train_data_loader)

6. For small input data call influence method on the fitted instance.

   influences = infl_model.influences(test_x, test_y, train_x, train_y)

   The result is a tensor of shape (training samples x test samples) that contains at index (i, j) the influence of training sample i on test sample j.

7. For larger data, wrap the model into a calculator and call methods on the calculator.

   infl_calc = SequentialInfluenceCalculator(infl_model)
   
    # Lazy object providing arrays batch-wise in a sequential manner
   lazy_influences = infl_calc.influences(test_data_loader, train_data_loader)
   
   # Trigger computation and pull results to memory
   influences = lazy_influences.compute(aggregator=NestedTorchCatAggregator())
   
   # Trigger computation and write results batch-wise to disk
   lazy_influences.to_zarr("influences_result", TorchNumpyConverter())

   The higher the absolute value of the influence of a training sample on a test sample, the more influential it is for the chosen test sample, model and data loaders. The sign of the influence determines whether it is useful (positive) or harmful (negative).

Note pyDVL currently only support PyTorch for Influence Functions. We are planning to add support for Jax and perhaps TensorFlow or even Keras.

The steps required to compute data values for your samples are:

1. Import the necessary packages (The exact packages depend on your specific use case).

   import matplotlib.pyplot as plt
   from sklearn.datasets import load_breast_cancer
   from sklearn.linear_model import LogisticRegression
   from pydvl.utils import Dataset, Scorer, Utility
   from pydvl.value import (
      compute_shapley_values,
      ShapleyMode,
      MaxUpdates,
   )

2. Create a Dataset object with your train and test splits.

   data = Dataset.from_sklearn(
       load_breast_cancer(),
       train_size=10,
       stratify_by_target=True,
       random_state=16,
   )

3. Create an instance of a SupervisedModel (basically any sklearn compatible predictor).

   model = LogisticRegression()

4. Create a Utility object to wrap the Dataset, the model and a scoring function.

   u = Utility(
      model,
      data,
      Scorer("accuracy", default=0.0)
   )

5. Use one of the methods defined in the library to compute the values. In our example, we will use Permutation Montecarlo Shapley, an approximate method for computing Data Shapley values.

   values = compute_shapley_values(
      u,
      mode=ShapleyMode.PermutationMontecarlo,
      done=MaxUpdates(100),
      seed=16,  
      progress=True
   )

   The result is a variable of type ValuationResult that contains the indices and their values as well as other attributes.

   The higher the value for an index, the more important it is for the chosen model, dataset and scorer.

6. (Optional) Convert the valuation result to a dataframe and analyze and visualize the values.

   df = values.to_dataframe(column="data_value")

Please open new issues for bugs, feature requests and extensions. You can read about the structure of the project, the toolchain and workflow in the guide for contributions.

pyDVL is distributed under LGPL-3.0. A complete version can be found in two files: here and here.

All contributions will be distributed under this license.