Examples with Saliency org.kie.kogito.explainability.model.Saliency used on opensource projects

Example 26 with Saliency

use of org.kie.kogito.explainability.model.Saliency in project kogito-apps by kiegroup.

the class ShapResultsTest method buildShapResults.

ShapResults buildShapResults(int nOutputs, int nFeatures, int scalar1, int scalar2) {
    Saliency[] saliencies = new Saliency[nOutputs];
    for (int i = 0; i < nOutputs; i++) {
        List<FeatureImportance> fis = new ArrayList<>();
        for (int j = 0; j < nFeatures; j++) {
            fis.add(new FeatureImportance(new Feature("f" + String.valueOf(j), Type.NUMBER, new Value(j)), i * j * scalar1));
        }
        saliencies[i] = new Saliency(new Output("o" + String.valueOf(i), Type.NUMBER, new Value(i), 1.0), fis);
    }
    RealVector fnull = MatrixUtils.createRealVector(new double[nOutputs]);
    fnull.mapAddToSelf(scalar2);
    return new ShapResults(saliencies, fnull);
}

Example 27 with Saliency

use of org.kie.kogito.explainability.model.Saliency in project kogito-apps by kiegroup.

the class ExplainabilityMetricsTest method testFidelityWithTextClassifier.

@Test
void testFidelityWithTextClassifier() throws ExecutionException, InterruptedException, TimeoutException {
    List<Pair<Saliency, Prediction>> pairs = new LinkedList<>();
    LimeConfig limeConfig = new LimeConfig().withSamples(10);
    LimeExplainer limeExplainer = new LimeExplainer(limeConfig);
    PredictionProvider model = TestUtils.getDummyTextClassifier();
    List<Feature> features = new LinkedList<>();
    features.add(FeatureFactory.newFulltextFeature("f-0", "brown fox", s -> Arrays.asList(s.split(" "))));
    features.add(FeatureFactory.newTextFeature("f-1", "money"));
    PredictionInput input = new PredictionInput(features);
    Prediction prediction = new SimplePrediction(input, model.predictAsync(List.of(input)).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit()).get(0));
    Map<String, Saliency> saliencyMap = limeExplainer.explainAsync(prediction, model).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit());
    for (Saliency saliency : saliencyMap.values()) {
        pairs.add(Pair.of(saliency, prediction));
    }
    Assertions.assertDoesNotThrow(() -> {
        ExplainabilityMetrics.classificationFidelity(pairs);
    });
}

Example 28 with Saliency

use of org.kie.kogito.explainability.model.Saliency in project kogito-apps by kiegroup.

the class PmmlScorecardCategoricalLimeExplainerTest method testPMMLScorecardCategorical.

@Test
void testPMMLScorecardCategorical() throws Exception {
    PredictionInput input = getTestInput();
    Random random = new Random();
    LimeConfig limeConfig = new LimeConfig().withSamples(10).withPerturbationContext(new PerturbationContext(0L, random, 1));
    LimeExplainer limeExplainer = new LimeExplainer(limeConfig);
    PredictionProvider model = getModel();
    List<PredictionOutput> predictionOutputs = model.predictAsync(List.of(input)).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit());
    assertThat(predictionOutputs).isNotNull().isNotEmpty();
    PredictionOutput output = predictionOutputs.get(0);
    assertThat(output).isNotNull();
    Prediction prediction = new SimplePrediction(input, output);
    Map<String, Saliency> saliencyMap = limeExplainer.explainAsync(prediction, model).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit());
    for (Saliency saliency : saliencyMap.values()) {
        assertThat(saliency).isNotNull();
        double v = ExplainabilityMetrics.impactScore(model, prediction, saliency.getTopFeatures(2));
        assertThat(v).isGreaterThan(0d);
    }
    assertDoesNotThrow(() -> ValidationUtils.validateLocalSaliencyStability(model, prediction, limeExplainer, 1, 0.4, 0.4));
    List<PredictionInput> inputs = getSamples();
    DataDistribution distribution = new PredictionInputsDataDistribution(inputs);
    String decision = "score";
    int k = 1;
    int chunkSize = 2;
    double f1 = ExplainabilityMetrics.getLocalSaliencyF1(decision, model, limeExplainer, distribution, k, chunkSize);
    AssertionsForClassTypes.assertThat(f1).isBetween(0d, 1d);
}

Example 29 with Saliency

use of org.kie.kogito.explainability.model.Saliency in project kogito-apps by kiegroup.

the class LimeExplainerTest method testWithDataDistribution.

@Test
void testWithDataDistribution() throws InterruptedException, ExecutionException, TimeoutException {
    Random random = new Random();
    PerturbationContext perturbationContext = new PerturbationContext(4L, random, 1);
    List<FeatureDistribution> featureDistributions = new ArrayList<>();
    int nf = 4;
    List<Feature> features = new ArrayList<>();
    for (int i = 0; i < nf; i++) {
        Feature numericalFeature = FeatureFactory.newNumericalFeature("f-" + i, Double.NaN);
        features.add(numericalFeature);
        List<Value> values = new ArrayList<>();
        for (int r = 0; r < 4; r++) {
            values.add(Type.NUMBER.randomValue(perturbationContext));
        }
        featureDistributions.add(new GenericFeatureDistribution(numericalFeature, values));
    }
    DataDistribution dataDistribution = new IndependentFeaturesDataDistribution(featureDistributions);
    LimeConfig limeConfig = new LimeConfig().withDataDistribution(dataDistribution).withPerturbationContext(perturbationContext).withSamples(10);
    LimeExplainer limeExplainer = new LimeExplainer(limeConfig);
    PredictionInput input = new PredictionInput(features);
    PredictionProvider model = TestUtils.getSumThresholdModel(random.nextDouble(), random.nextDouble());
    PredictionOutput output = model.predictAsync(List.of(input)).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit()).get(0);
    Prediction prediction = new SimplePrediction(input, output);
    Map<String, Saliency> saliencyMap = limeExplainer.explainAsync(prediction, model).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit());
    assertThat(saliencyMap).isNotNull();
    String decisionName = "inside";
    Saliency saliency = saliencyMap.get(decisionName);
    assertThat(saliency).isNotNull();
}

Example 30 with Saliency

use of org.kie.kogito.explainability.model.Saliency in project kogito-apps by kiegroup.

the class LimeExplainerTest method testSparseBalance.

@ParameterizedTest
@ValueSource(longs = { 0, 1, 2, 3, 4 })
void testSparseBalance(long seed) throws InterruptedException, ExecutionException, TimeoutException {
    for (int nf = 1; nf < 4; nf++) {
        Random random = new Random();
        int noOfSamples = 100;
        LimeConfig limeConfigNoPenalty = new LimeConfig().withPerturbationContext(new PerturbationContext(seed, random, DEFAULT_NO_OF_PERTURBATIONS)).withSamples(noOfSamples).withPenalizeBalanceSparse(false);
        LimeExplainer limeExplainerNoPenalty = new LimeExplainer(limeConfigNoPenalty);
        List<Feature> features = new ArrayList<>();
        for (int i = 0; i < nf; i++) {
            features.add(TestUtils.getMockedNumericFeature(i));
        }
        PredictionInput input = new PredictionInput(features);
        PredictionProvider model = TestUtils.getSumSkipModel(0);
        PredictionOutput output = model.predictAsync(List.of(input)).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit()).get(0);
        Prediction prediction = new SimplePrediction(input, output);
        Map<String, Saliency> saliencyMapNoPenalty = limeExplainerNoPenalty.explainAsync(prediction, model).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit());
        assertThat(saliencyMapNoPenalty).isNotNull();
        String decisionName = "sum-but0";
        Saliency saliencyNoPenalty = saliencyMapNoPenalty.get(decisionName);
        LimeConfig limeConfig = new LimeConfig().withPerturbationContext(new PerturbationContext(seed, random, DEFAULT_NO_OF_PERTURBATIONS)).withSamples(noOfSamples).withPenalizeBalanceSparse(true);
        LimeExplainer limeExplainer = new LimeExplainer(limeConfig);
        Map<String, Saliency> saliencyMap = limeExplainer.explainAsync(prediction, model).get(Config.INSTANCE.getAsyncTimeout(), Config.INSTANCE.getAsyncTimeUnit());
        assertThat(saliencyMap).isNotNull();
        Saliency saliency = saliencyMap.get(decisionName);
        for (int i = 0; i < features.size(); i++) {
            double score = saliency.getPerFeatureImportance().get(i).getScore();
            double scoreNoPenalty = saliencyNoPenalty.getPerFeatureImportance().get(i).getScore();
            assertThat(Math.abs(score)).isLessThanOrEqualTo(Math.abs(scoreNoPenalty));
        }
    }
}