[
https://issues.apache.org/jira/browse/FLINK-1807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14526784#comment-14526784
]
ASF GitHub Bot commented on FLINK-1807:
---------------------------------------
Github user thvasilo commented on a diff in the pull request:
https://github.com/apache/flink/pull/613#discussion_r29597434
--- Diff:
flink-staging/flink-ml/src/main/scala/org/apache/flink/ml/optimization/LossFunction.scala
---
@@ -0,0 +1,101 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.flink.ml.optimization
+
+import org.apache.flink.ml.common.{WeightVector, LabeledVector}
+import org.apache.flink.ml.math.{Vector => FlinkVector, BLAS}
+
+
+abstract class LossFunction extends Serializable{
+
+
+ /** Calculates the loss for a given prediction/truth pair
+ *
+ * @param prediction The predicted value
+ * @param truth The true value
+ */
+ protected def loss(prediction: Double, truth: Double): Double
+
+ /** Calculates the derivative of the loss function with respect to the
prediction
+ *
+ * @param prediction The predicted value
+ * @param truth The true value
+ */
+ protected def lossDerivative(prediction: Double, truth: Double): Double
+
+ /** Compute the gradient and the loss for the given data.
+ * The provided cumGradient is updated in place.
+ *
+ * @param example The features and the label associated with the example
+ * @param weights The current weight vector
+ * @param cumGradient The vector to which the gradient will be added
to, in place.
+ * @return A tuple containing the computed loss as its first element
and a the loss derivative as
+ * its second element.
+ */
+ def lossAndGradient(
+ example: LabeledVector,
+ weights: WeightVector,
+ cumGradient: FlinkVector,
+ regType: RegularizationType,
+ regParameter: Double): (Double, Double) = {
+ val features = example.vector
+ val label = example.label
+ // TODO(tvas): We could also provide for the case where we don't want
an intercept value
+ // i.e. data already centered
+ val prediction = BLAS.dot(features, weights.weights) +
weights.intercept
+ val lossValue: Double = loss(prediction, label)
+ // The loss derivative is used to update the intercept
+ val lossDeriv= lossDerivative(prediction, label)
+ BLAS.axpy(lossDeriv , features, cumGradient)
+ val adjustedLoss = {
+ regType match {
+ case x : DiffRegularizationType => {
+ x.regularizedLossAndGradient(lossValue, weights.weights,
cumGradient, regParameter)
--- End diff --
I think you're right. I'll try to think a bit more about this and how it
fits with the L-BFGS solver as well, since that requires a regularized loss to
be available. But it should be doable, I'll get on this tomorrow.
> Stochastic gradient descent optimizer for ML library
> ----------------------------------------------------
>
> Key: FLINK-1807
> URL: https://issues.apache.org/jira/browse/FLINK-1807
> Project: Flink
> Issue Type: Improvement
> Components: Machine Learning Library
> Reporter: Till Rohrmann
> Assignee: Theodore Vasiloudis
> Labels: ML
>
> Stochastic gradient descent (SGD) is a widely used optimization technique in
> different ML algorithms. Thus, it would be helpful to provide a generalized
> SGD implementation which can be instantiated with the respective gradient
> computation. Such a building block would make the development of future
> algorithms easier.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
