Examples with Gradient1Function uk.ac.sussex.gdsc.smlm.function.Gradient1Function used on opensource projects

Example 6 with Gradient1Function

use of uk.ac.sussex.gdsc.smlm.function.Gradient1Function in project GDSC-SMLM by aherbert.

the class LsqVarianceGradientProcedureTest method gradientProcedureUnrolledComputesSameAsGradientProcedure.

private void gradientProcedureUnrolledComputesSameAsGradientProcedure(RandomSeed seed, int nparams, boolean precomputed) {
    final int iter = 10;
    final ArrayList<double[]> paramsList = new ArrayList<>(iter);
    createFakeParams(RngUtils.create(seed.getSeed()), nparams, iter, paramsList);
    Gradient1Function func = new FakeGradientFunction(blockWidth, nparams);
    if (precomputed) {
        func = OffsetGradient1Function.wrapGradient1Function(func, SimpleArrayUtils.newArray(func.size(), 0.1, 1.3));
    }
    final IntArrayFormatSupplier msg = new IntArrayFormatSupplier("[%d] Observations: Not same variance @ %d", 2);
    msg.set(0, nparams);
    for (int i = 0; i < paramsList.size(); i++) {
        final LsqVarianceGradientProcedure p1 = new LsqVarianceGradientProcedure(func);
        p1.variance(paramsList.get(i));
        final LsqVarianceGradientProcedure p2 = LsqVarianceGradientProcedureUtils.create(func);
        p2.variance(paramsList.get(i));
        // Exactly the same ...
        Assertions.assertArrayEquals(p1.variance, p2.variance, msg.set(1, i));
    }
}

Example 7 with Gradient1Function

use of uk.ac.sussex.gdsc.smlm.function.Gradient1Function in project GDSC-SMLM by aherbert.

the class PoissonGradientProcedureTest method gradientProcedureUnrolledComputesSameAsGradientProcedure.

private void gradientProcedureUnrolledComputesSameAsGradientProcedure(RandomSeed seed, int nparams, boolean precomputed) {
    final int iter = 10;
    final ArrayList<double[]> paramsList = new ArrayList<>(iter);
    createFakeParams(RngUtils.create(seed.getSeed()), nparams, iter, paramsList);
    Gradient1Function func = new FakeGradientFunction(blockWidth, nparams);
    if (precomputed) {
        func = OffsetGradient1Function.wrapGradient1Function(func, SimpleArrayUtils.newArray(func.size(), 0.1, 1.3));
    }
    // Create messages
    final IntArrayFormatSupplier msgOal = getMessage(nparams, "[%d] Observations: Not same alpha linear @ %d");
    final IntArrayFormatSupplier msgOam = getMessage(nparams, "[%d] Observations: Not same alpha matrix @ %d");
    for (int i = 0; i < paramsList.size(); i++) {
        final PoissonGradientProcedure p1 = new PoissonGradientProcedure(func);
        p1.computeFisherInformation(paramsList.get(i));
        final PoissonGradientProcedure p2 = PoissonGradientProcedureUtils.create(func);
        p2.computeFisherInformation(paramsList.get(i));
        // Exactly the same ...
        Assertions.assertArrayEquals(p1.getLinear(), p2.getLinear(), msgOal.set(1, i));
        final double[][] am1 = p1.getMatrix();
        final double[][] am2 = p2.getMatrix();
        Assertions.assertArrayEquals(am1, am2, msgOam.set(1, i));
    }
}

Example 8 with Gradient1Function

use of uk.ac.sussex.gdsc.smlm.function.Gradient1Function in project GDSC-SMLM by aherbert.

the class FastMleSteppingFunctionSolver method computeFunctionFisherInformationMatrix.

@Override
protected FisherInformationMatrix computeFunctionFisherInformationMatrix(double[] y, double[] a) {
    // The fisher information is that for a Poisson process.
    // We must wrap the gradient function if weights are present.
    Gradient1Function f1 = (Gradient1Function) function;
    if (obsVariances != null) {
        f1 = OffsetGradient1Function.wrapGradient1Function(f1, obsVariances);
    }
    final PoissonGradientProcedure p = PoissonGradientProcedureUtils.create(f1);
    p.computeFisherInformation(a);
    if (p.isNaNGradients()) {
        throw new FunctionSolverException(FitStatus.INVALID_GRADIENTS);
    }
    return new FisherInformationMatrix(p.getLinear(), p.numberOfGradients);
}

Example 9 with Gradient1Function

use of uk.ac.sussex.gdsc.smlm.function.Gradient1Function in project GDSC-SMLM by aherbert.

the class LsqLvmGradientProcedureTest method gradientProcedureLinearIsFasterThanGradientProcedureMatrix.

private void gradientProcedureLinearIsFasterThanGradientProcedureMatrix(RandomSeed seed, final int nparams, final BaseLsqLvmGradientProcedureFactory factory1, final BaseLsqLvmGradientProcedureFactory factory2, boolean doAssert) {
    Assumptions.assumeTrue(TestSettings.allow(TestComplexity.MEDIUM));
    final int iter = 100;
    final ArrayList<double[]> paramsList = new ArrayList<>(iter);
    final ArrayList<double[]> yList = new ArrayList<>(iter);
    createData(RngUtils.create(seed.getSeed()), 1, iter, paramsList, yList);
    // Remove the timing of the function call by creating a dummy function
    final Gradient1Function func = new FakeGradientFunction(blockWidth, nparams);
    // Create messages
    final IndexSupplier msgA = new IndexSupplier(1, "A ", null);
    final IndexSupplier msgB = new IndexSupplier(1, "B ", null);
    for (int i = 0; i < paramsList.size(); i++) {
        final BaseLsqLvmGradientProcedure p1 = factory1.createProcedure(yList.get(i), func);
        p1.gradient(paramsList.get(i));
        p1.gradient(paramsList.get(i));
        final BaseLsqLvmGradientProcedure p2 = factory2.createProcedure(yList.get(i), func);
        p2.gradient(paramsList.get(i));
        p2.gradient(paramsList.get(i));
        // Check they are the same
        Assertions.assertArrayEquals(p1.getAlphaLinear(), p2.getAlphaLinear(), msgA.set(0, i));
        Assertions.assertArrayEquals(p1.beta, p2.beta, msgB.set(0, i));
    }
    // Realistic loops for an optimisation
    final int loops = 15;
    // Run till stable timing
    final Timer t1 = new Timer() {

        @Override
        void run() {
            for (int i = 0, k = 0; i < paramsList.size(); i++) {
                final BaseLsqLvmGradientProcedure p = factory1.createProcedure(yList.get(i), func);
                for (int j = loops; j-- > 0; ) {
                    p.gradient(paramsList.get(k++ % iter));
                }
            }
        }
    };
    final long time1 = t1.getTime();
    final Timer t2 = new Timer(t1.loops) {

        @Override
        void run() {
            for (int i = 0, k = 0; i < paramsList.size(); i++) {
                final BaseLsqLvmGradientProcedure p2 = factory2.createProcedure(yList.get(i), func);
                for (int j = loops; j-- > 0; ) {
                    p2.gradient(paramsList.get(k++ % iter));
                }
            }
        }
    };
    final long time2 = t2.getTime();
    final LogRecord record = TestLogUtils.getTimingRecord(factory1.getClass().getSimpleName() + nparams, time1, factory2.getClass().getSimpleName(), time2);
    if (!doAssert && record.getLevel() == TestLevel.TEST_FAILURE) {
        record.setLevel(TestLevel.TEST_WARNING);
    }
    logger.log(record);
}

Example 10 with Gradient1Function

use of uk.ac.sussex.gdsc.smlm.function.Gradient1Function in project GDSC-SMLM by aherbert.

the class LvmGradientProcedureTest method gradientProcedureIsFasterUnrolledThanGradientProcedure.

private void gradientProcedureIsFasterUnrolledThanGradientProcedure(RandomSeed seed, final int nparams, final Type type, final boolean precomputed) {
    Assumptions.assumeTrue(TestSettings.allow(TestComplexity.MEDIUM));
    final int iter = 100;
    final ArrayList<double[]> paramsList = new ArrayList<>(iter);
    final ArrayList<double[]> yList = new ArrayList<>(iter);
    createData(RngUtils.create(seed.getSeed()), 1, iter, paramsList, yList);
    // Remove the timing of the function call by creating a dummy function
    final FakeGradientFunction fgf = new FakeGradientFunction(blockWidth, nparams);
    final Gradient1Function func;
    if (precomputed) {
        final double[] b = SimpleArrayUtils.newArray(fgf.size(), 0.1, 1.3);
        func = OffsetGradient1Function.wrapGradient1Function(fgf, b);
    } else {
        func = fgf;
    }
    final FastLog fastLog = type == Type.FAST_LOG_MLE ? getFastLog() : null;
    final IntArrayFormatSupplier msgA = new IntArrayFormatSupplier("A [%d]", 1);
    final IntArrayFormatSupplier msgB = new IntArrayFormatSupplier("B [%d]", 1);
    for (int i = 0; i < paramsList.size(); i++) {
        final LvmGradientProcedure p1 = createProcedure(type, yList.get(i), func, fastLog);
        p1.gradient(paramsList.get(i));
        p1.gradient(paramsList.get(i));
        final LvmGradientProcedure p2 = LvmGradientProcedureUtils.create(yList.get(i), func, type, fastLog);
        p2.gradient(paramsList.get(i));
        p2.gradient(paramsList.get(i));
        // Check they are the same
        Assertions.assertArrayEquals(p1.getAlphaLinear(), p2.getAlphaLinear(), msgA.set(0, i));
        Assertions.assertArrayEquals(p1.beta, p2.beta, msgB.set(0, i));
    }
    // Realistic loops for an optimisation
    final int loops = 15;
    // Run till stable timing
    final Timer t1 = new Timer() {

        @Override
        void run() {
            for (int i = 0, k = 0; i < paramsList.size(); i++) {
                final LvmGradientProcedure p1 = createProcedure(type, yList.get(i), func, fastLog);
                for (int j = loops; j-- > 0; ) {
                    p1.gradient(paramsList.get(k++ % iter));
                }
            }
        }
    };
    final long time1 = t1.getTime();
    final Timer t2 = new Timer(t1.loops) {

        @Override
        void run() {
            for (int i = 0, k = 0; i < paramsList.size(); i++) {
                final LvmGradientProcedure p2 = LvmGradientProcedureUtils.create(yList.get(i), func, type, fastLog);
                for (int j = loops; j-- > 0; ) {
                    p2.gradient(paramsList.get(k++ % iter));
                }
            }
        }
    };
    final long time2 = t2.getTime();
    logger.log(TestLogUtils.getTimingRecord(new TimingResult(() -> String.format("%s, Precomputed=%b : Standard", type, precomputed), time1), new TimingResult(() -> String.format("Unrolled %d", nparams), time2)));
}