Download as PDF, PPTX
![Immutable Facts
case class PreyReading(sensorId: Int,
locationId: Int,
timestamp: Long,
animalsLowerBound: Double,
animalsUpperBound: Double,
percentZebras: Double)
implicit val preyReadingFormatter = Json.format[PreyReading]](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-20-2048.jpg)
, reading)](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-24-2048.jpg)
![Almost Immutable Data Model
{
"sensor_id": 123,
"readings": [
{
"sensor_id": 456,
"timestamp": 1457403598289,
"lower_bound": 12.0,
"upper_bound": 18.0,
"percent_zebras": 0.60
}
]
}](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-26-2048.jpg)
![Original Features
object SquawkLength extends FeatureType[Int]
object Super extends LabelType[Boolean]
val originalFeatures: Set[FeatureType] = Set(SquawkLength)
val label = Super](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-35-2048.jpg)
![Basic Features
object PastSquawks extends FeatureType[Int]
val basicFeatures = originalFeatures + PastSquawks](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-36-2048.jpg)
![More Features
object MobileSquawker extends FeatureType[Boolean]
val moreFeatures = basicFeatures + MobileSquawker](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-37-2048.jpg)
![Feature Collections
case class FeatureCollection(id: Int,
createdAt: DateTime,
features: Set[_ <: FeatureType[_]],
label: LabelType[_])](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-38-2048.jpg)
![Fallback Collections
def validCollection(collections: RDD[FeatureCollection],
invalidFeatures: Set[FeatureType[_]]) = {
val validCollections = collections.filter(
fc => !fc.features
.exists(invalidFeatures.contains))
.sortBy(collection => collection.id)
if (validCollections.count() > 0) {
validCollections.first()
} else
FallbackCollection
}](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-41-2048.jpg)
![Calling the Service
def callServer(remoteServer: PyroProxy, jobConfiguration:
JobConfiguration) = {
Future.firstCompletedOf(
List(
timedOut,
Future {
remoteServer.call("generate",
jobConfiguration.contentPath,
jobConfiguration.stylePath,
jobConfiguration.modelPath,
jobConfiguration.modelType.toString,
jobConfiguration.width,
jobConfiguration.alpha,
jobConfiguration.beta,
jobConfiguration.iterations).asInstanceOf[Boolean]
}))}](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-57-2048.jpg)
![Building Lineages
val rawData: RawData
val featureSet: Set[FeatureType]
val model: ClassificationModel
val modelMetrics: BinaryClassificationMetrics](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-63-2048.jpg)
![Models as Services
def predict(model: Features => Prediction)
(features: Features):
Future[Either[String, Prediction]] = { ??? }](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-67-2048.jpg)
![Models as Services
val routes = {
pathPrefix("model") {
(post & entity(as[PredictionRequest])) { context =>
complete {
predict(model: Features => Prediction)
(extractFeatures(context))
.map[ToResponseMarshallable] {
case Right(prediction) => PredictionSummary(prediction)
case Left(errorMessage) => BadRequest -> errorMessage
}
}
}
}
}](https://image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-68-2048.jpg)
![Immutable Facts
case class PreyReading(sensorId: Int,
locationId: Int,
timestamp: Long,
animalsLowerBound: Double,
animalsUpperBound: Double,
percentZebras: Double)
implicit val preyReadingFormatter = Json.format[PreyReading]](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-20-2048.jpg)
, reading)](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-24-2048.jpg)
![Almost Immutable Data Model
{
"sensor_id": 123,
"readings": [
{
"sensor_id": 456,
"timestamp": 1457403598289,
"lower_bound": 12.0,
"upper_bound": 18.0,
"percent_zebras": 0.60
}
]
}](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-26-2048.jpg)
![Original Features
object SquawkLength extends FeatureType[Int]
object Super extends LabelType[Boolean]
val originalFeatures: Set[FeatureType] = Set(SquawkLength)
val label = Super](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-35-2048.jpg)
![Basic Features
object PastSquawks extends FeatureType[Int]
val basicFeatures = originalFeatures + PastSquawks](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-36-2048.jpg)
![More Features
object MobileSquawker extends FeatureType[Boolean]
val moreFeatures = basicFeatures + MobileSquawker](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-37-2048.jpg)
![Feature Collections
case class FeatureCollection(id: Int,
createdAt: DateTime,
features: Set[_ <: FeatureType[_]],
label: LabelType[_])](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-38-2048.jpg)
![Fallback Collections
def validCollection(collections: RDD[FeatureCollection],
invalidFeatures: Set[FeatureType[_]]) = {
val validCollections = collections.filter(
fc => !fc.features
.exists(invalidFeatures.contains))
.sortBy(collection => collection.id)
if (validCollections.count() > 0) {
validCollections.first()
} else
FallbackCollection
}](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-41-2048.jpg)
![Calling the Service
def callServer(remoteServer: PyroProxy, jobConfiguration:
JobConfiguration) = {
Future.firstCompletedOf(
List(
timedOut,
Future {
remoteServer.call("generate",
jobConfiguration.contentPath,
jobConfiguration.stylePath,
jobConfiguration.modelPath,
jobConfiguration.modelType.toString,
jobConfiguration.width,
jobConfiguration.alpha,
jobConfiguration.beta,
jobConfiguration.iterations).asInstanceOf[Boolean]
}))}](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-57-2048.jpg)
![Building Lineages
val rawData: RawData
val featureSet: Set[FeatureType]
val model: ClassificationModel
val modelMetrics: BinaryClassificationMetrics](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-63-2048.jpg)
![Models as Services
def predict(model: Features => Prediction)
(features: Features):
Future[Either[String, Prediction]] = { ??? }](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-67-2048.jpg)
![Models as Services
val routes = {
pathPrefix("model") {
(post & entity(as[PredictionRequest])) { context =>
complete {
predict(model: Features => Prediction)
(extractFeatures(context))
.map[ToResponseMarshallable] {
case Right(prediction) => PredictionSummary(prediction)
case Left(errorMessage) => BadRequest -> errorMessage
}
}
}
}
}](https://crownmelresort.com/image.slidesharecdn.com/introducingreactivemachinelearning-160309200307/75/Introducing-Reactive-Machine-Learning-68-2048.jpg)
Uploaded byJeff Smith
Introducing Reactive Machine Learning
This document introduces reactive machine learning and discusses how reactive strategies can be applied to machine learning systems. It describes how reactive systems are responsive, resilient, elastic, and message-driven. It then discusses how to build reactive machine learning systems that can collect and process data in distributed databases, generate features, learn models, publish models as services, and make predictions in a responsive and resilient manner.
More Related Content
Similar to Introducing Reactive Machine Learning
Introducing Reactive Machine Learning
- 1. Introducing Reactive Machine Learning JeffSmith @jeffksmithjr
- 2. x.ai is apersonal assistant who schedules meetings for you
- 3.
- 4. nom nom, thedata dog Scala & Python Spark & Akka Couchbase Machine Learning
- 5. Reactive + MachineLearning
- 6.
- 7. Traits of ReactiveSystems
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 19.
- 20. Immutable Facts case classPreyReading(sensorId: Int, locationId: Int, timestamp: Long, animalsLowerBound: Double, animalsUpperBound: Double, percentZebras: Double) implicit val preyReadingFormatter = Json.format[PreyReading]
- 21.
- 22.
- 23.
- 24. Immutable Facts val reading= PreyReading(36, 12, currentTimeMillis(), 12.0, 18.0, 0.60) val setDoc = bucket.set[PreyReading](readingId(reading), reading)
- 25. Scaling with DistributedDatabases
- 26. Almost Immutable DataModel { "sensor_id": 123, "readings": [ { "sensor_id": 456, "timestamp": 1457403598289, "lower_bound": 12.0, "upper_bound": 18.0, "percent_zebras": 0.60 } ] }
- 27.
- 28. Truly Immutable DataModel { "sensor_id_reading_id": 123456, "timestamp": 1457403598289, "lower_bound": 12.0, "upper_bound": 18.0, "percent_zebras": 0.60 }
- 29.
- 30.
- 31. Feature Generation Raw DataFeaturesFeature Generation Pipeline
- 32.
- 33. Pipeline Failure Raw DataFeaturesFeature Generation Pipeline Raw Data FeaturesFeature Generation Pipeline
- 34. Supervising Feature Generation RawData FeaturesFeature Generation Pipeline Supervision
- 35. Original Features object SquawkLengthextends FeatureType[Int] object Super extends LabelType[Boolean] val originalFeatures: Set[FeatureType] = Set(SquawkLength) val label = Super
- 36. Basic Features object PastSquawksextends FeatureType[Int] val basicFeatures = originalFeatures + PastSquawks
- 37. More Features object MobileSquawkerextends FeatureType[Boolean] val moreFeatures = basicFeatures + MobileSquawker
- 38. Feature Collections case classFeatureCollection(id: Int, createdAt: DateTime, features: Set[_ <: FeatureType[_]], label: LabelType[_])
- 39. Feature Collections val earlierCollection= FeatureCollection(101, earlier, basicFeatures, label) val latestCollection = FeatureCollection(202, now, moreFeatures, label) val featureCollections = sc.parallelize( Seq(earlierCollection, latestCollection))
- 40. Fallback Collections val FallbackCollection= FeatureCollection(404, beginningOfTime, originalFeatures, label)
- 41. Fallback Collections def validCollection(collections:RDD[FeatureCollection], invalidFeatures: Set[FeatureType[_]]) = { val validCollections = collections.filter( fc => !fc.features .exists(invalidFeatures.contains)) .sortBy(collection => collection.id) if (validCollections.count() > 0) { validCollections.first() } else FallbackCollection }
- 42.
- 43.
- 44. Learning Models Features ModelModelLearning Pipeline
- 45.
- 46.
- 47.
- 48. Data Preparation val labelIndexer= new StringIndexer() .setInputCol("label") .setOutputCol("indexedLabel") .fit(instances) val featureIndexer = new VectorIndexer() .setInputCol("features") .setOutputCol("indexedFeatures") .fit(instances) val Array(trainingData, testingData) = instances.randomSplit( Array(0.8, 0.2))
- 49. Learning a Model valdecisionTree = new DecisionTreeClassifier() .setLabelCol("indexedLabel") .setFeaturesCol("indexedFeatures") val labelConverter = new IndexToString() .setInputCol("prediction") .setOutputCol("predictedLabel") .setLabels(labelIndexer.labels) val pipeline = new Pipeline() .setStages(Array(labelIndexer, featureIndexer, decisionTree, labelConverter))
- 50. Evolving Modeling Strategies valrandomForest = new RandomForestClassifier() .setLabelCol("indexedLabel") .setFeaturesCol("indexedFeatures") val revisedPipeline = new Pipeline() .setStages(Array(labelIndexer, featureIndexer, randomForest, labelConverter))
- 51. Deep Models ofArtistic Style
- 52. Refactoring Command LineTools > python neural-art-tf.py -m vgg -mp ./vgg -c ./images/ bear.jpg -s ./images/style.jpg -w 800 def produce_art(content_image_path, style_image_path, model_path, model_type, width, alpha, beta, num_iters):
- 53. Exposing a Service classNeuralServer(object): def generate(self, content_image_path, style_image_path, model_path, model_type, width, alpha, beta, num_iters): produce_art(content_image_path, style_image_path, model_path, model_type, width, alpha, beta, num_iters) return True daemon = Pyro4.Daemon() ns = Pyro4.locateNS() uri = daemon.register(NeuralServer) ns.register("neuralserver", uri) daemon.requestLoop()
- 54. Encoding Model Types objectModelType extends Enumeration { type ModelType = Value val VGG = Value("VGG") val I2V = Value("I2V") }
- 55. Encoding Valid Configuration caseclass JobConfiguration(contentPath: String, stylePath: String, modelPath: String, modelType: ModelType, width: Integer = 800, alpha: java.lang.Double = 1.0, beta: java.lang.Double = 200.0, iterations: Integer = 5000)
- 56. Finding the Service valns = NameServerProxy.locateNS(null) val remoteServer = new PyroProxy(ns.lookup("neuralserver"))
- 57. Calling the Service defcallServer(remoteServer: PyroProxy, jobConfiguration: JobConfiguration) = { Future.firstCompletedOf( List( timedOut, Future { remoteServer.call("generate", jobConfiguration.contentPath, jobConfiguration.stylePath, jobConfiguration.modelPath, jobConfiguration.modelType.toString, jobConfiguration.width, jobConfiguration.alpha, jobConfiguration.beta, jobConfiguration.iterations).asInstanceOf[Boolean] }))}
- 58.
- 59. Hybrid Model learning FeaturesModelModel Learning Pipeline
- 60.
- 61.
- 62.
- 63. Building Lineages val rawData:RawData val featureSet: Set[FeatureType] val model: ClassificationModel val modelMetrics: BinaryClassificationMetrics
- 64.
- 65.
- 66.
- 67. Models as Services defpredict(model: Features => Prediction) (features: Features): Future[Either[String, Prediction]] = { ??? }
- 68. Models as Services valroutes = { pathPrefix("model") { (post & entity(as[PredictionRequest])) { context => complete { predict(model: Features => Prediction) (extractFeatures(context)) .map[ToResponseMarshallable] { case Right(prediction) => PredictionSummary(prediction) case Left(errorMessage) => BadRequest -> errorMessage } } } } }
- 69.
- 70.
- 71.
- 72.
- 73.
- 74. Traits of ReactiveSystems
- 75.
- 76.
- 77.
- 78. manning.com reactivemachinelearning.com medium.com/data-engineering M A NN I N G Jeff SmithUse the code reactnymu for 39% off!
- 79. x.ai @xdotai hello@human.x.ai New York, NewYork We’re hiring!
- 80.
- 81. Code: reactnymu Jeff Smith @jeffksmithjr reactivemachinelearning.com MA N N I N G Jeff Smith