Apache PredictionIO - Machine Learning 15minutes! #12 - Speaker Deck

Apache PredictionIO - Machine Learning 15minutes! #12

by takahiro-hagino

Speaker Deck

Tweet

Tweet

Slide 1

Slide 1 text

Open Source Machine Learning Server 15 Machine Learning minutes!

Slide 2

Slide 2 text

Apache PredictionIOͷίϛολ͕ޠΔ Spark&Elasticsearch ػցֶशج൫

Slide 3

Slide 3 text

ػցֶशͷ͓೰Έ • ֶशɾϞσϧσʔλͷετϨʔδ • ػցֶशͷ෼ࢄॲཧϑϨʔϜϫʔΫ • σʔλ༧ଌͷWebαʔϏεʢAPIʣԽ ղܾͷબ୒ࢶ → ఻͍͑ͨ͜ͱ

Slide 4

Slide 4 text

ػցֶशͷ͓೰Έ • ֶशɾϞσϧσʔλͷετϨʔδ • ػցֶशͷ෼ࢄॲཧϑϨʔϜϫʔΫ • σʔλ༧ଌͷWebαʔϏεʢAPIʣԽ ղܾͷબ୒ࢶ → ఻͍͑ͨ͜ͱ

Slide 5

Slide 5 text

Photo by Bernard Spragg. NZ About me

Slide 6

Slide 6 text

Proﬁle Takahiro Hagino Bizreach גࣜձࣾϏζϦʔν • ٻਓݕࡧΤϯδϯʮελϯόΠʯ • AIࣨ

Slide 7

Slide 7 text

Open Source Machine Learning Server

Slide 8

Slide 8 text

ϏζϦʔνͰ͸ػցֶशΞϓϦέʔγϣϯͷ ։ൃɾӡ༻ج൫ͱͯ͠ "QBDIF1SFEJDUJPO*0ʹऔΓ૊ΜͰ͍·͢ɻ ೥݄ฐ͔ࣾΒ໊͕ίϛολͱͯ͠ 1SFEJDUJPO*0ͷ։ൃʹࢀՃ͢Δ͜ͱʹͳΓ·ͨ͠ɻ Shinsuke Sugaya Naoki  Takezoe Takako  Shimamoto Takahiro Hagino

Slide 9

Slide 9 text

ϏζϦʔνͰ͸ػցֶशΞϓϦέʔγϣϯͷ ։ൃɾӡ༻ج൫ͱͯ͠ "QBDIF1SFEJDUJPO*0ʹऔΓ૊ΜͰ͍·͢ɻ ೥݄ฐ͔ࣾΒ໊͕ίϛολͱͯ͠ 1SFEJDUJPO*0ͷ։ൃʹࢀՃ͢Δ͜ͱʹͳΓ·ͨ͠ɻ Shinsuke Sugaya Naoki  Takezoe Takako  Shimamoto Takahiro Hagino

Slide 10

Slide 10 text

No content

Slide 11

Slide 11 text

x ML ٻ৬ऀͱٻਓͷϚονϯά Search Quality and Recommendation ೥ऩਪఆ Salary Prediction ۀछɾ৬छਪఆ Job Category Prediction ٻਓಛ௃ਪఆ Prediction of Job Characteristics

Slide 12

Slide 12 text

ٻ৬ऀͱٻਓͷϚονϯά Search Quality and Recommendation ೥ऩਪఆ Salary Prediction ۀछɾ৬छਪఆ Job Category Prediction ٻਓಛ௃ਪఆ Prediction of Job Characteristics x ML

Slide 13

Slide 13 text

ٻ৬ऀͱٻਓͷϚονϯά Search Quality and Recommendation ೥ऩਪఆ Salary Prediction ۀछɾ৬छਪఆ Job Category Prediction ٻਓಛ௃ਪఆ Prediction of Job Characteristics x ML

Slide 14

Slide 14 text

ٻ৬ऀͱٻਓͷϚονϯά Search Quality and Recommendation ೥ऩਪఆ Salary Prediction ۀछɾ৬छਪఆ Job Category Prediction ٻਓಛ௃ਪఆ Prediction of Job Characteristics x ML

Slide 15

Slide 15 text

ٻ৬ऀͱٻਓͷϚονϯά Search Quality and Recommendation ೥ऩਪఆ Salary Prediction ۀछɾ৬छਪఆ Job Category Prediction ٻਓಛ௃ਪఆ Prediction of Job Characteristics x ML

Slide 16

Slide 16 text

ٻ৬ऀͱٻਓͷϚονϯά Search Quality and Recommendation ೥ऩਪఆ Salary Prediction ۀछɾ৬छਪఆ Job Category Prediction ٻਓಛ௃ਪఆ Prediction of Job Characteristics x ML

Slide 17

Slide 17 text

ٻ৬ऀͱٻਓͷϚονϯά Search Quality and Recommendation ೥ऩਪఆ Salary Prediction ۀछɾ৬छਪఆ Job Category Prediction ٻਓಛ௃ਪఆ Prediction of Job Characteristics x ML

Slide 18

Slide 18 text

τϨʔχϯάσʔλ σʔλ͝ͱES΁ͷόονΠϯϙʔτεΫϦϓτΛ४උ͍ͯͨ͠  ֶशͷ؀ڥ΍࣮ߦϑϩʔ͕୲౰ऀ͝ͱʹଐਓԽ͓ͯ͠Γɺ  τϨʔχϯάσʔλͷϑΥʔϚοτ΋ఆ·͍ͬͯ·ͤΜɻ ֶशॲཧͷ࣮ߦ࣌ؒ σʔλྔͷ૿Ճʹͱ΋ͳ͏ֶशͷ௕࣌ؒԽ x MLͷ͓ͳ΍Έ

Slide 19

Slide 19 text

τϨʔχϯάσʔλ σʔλ͝ͱES΁ͷόονΠϯϙʔτεΫϦϓτΛ४උ͍ͯͨ͠  ֶशͷ؀ڥ΍࣮ߦϑϩʔ͕୲౰ऀ͝ͱʹଐਓԽ͓ͯ͠Γɺ  τϨʔχϯάσʔλͷϑΥʔϚοτ΋ఆ·͍ͬͯ·ͤΜɻ ֶशॲཧͷ࣮ߦ࣌ؒ σʔλྔͷ૿Ճʹͱ΋ͳ͏ֶशͷ௕࣌ؒԽ x MLͷ͓ͳ΍Έ

Slide 20

Slide 20 text

ֶशϞσϧͷετϨʔδ ػೳ͝ͱʹֶशϞσϧͷอଘઌ΋ଐਓԽ ༧ଌͷWeb API ػೳ͝ͱʹTornadoͳͲͰ؆қͳAPIαʔόΛ࡞੒͓ͯ͠Γɺ  ͢͹΍͘ΞϓϦʹ൓өͰ͖ͳ͍ ༧ଌ݁ՌΛฦ͢APIͱͯ͠ESΛར༻͢ΔͳͲ͍ͯͨ͠ x MLͷ͓ͳ΍Έ

Slide 21

Slide 21 text

ֶशϞσϧͷετϨʔδ ػೳ͝ͱʹֶशϞσϧͷอଘઌ΋ଐਓԽ ༧ଌͷWeb API ػೳ͝ͱʹTornadoͳͲͰ؆қͳAPIαʔόΛ࡞੒͓ͯ͠Γɺ  ͢͹΍͘ΞϓϦʹ൓өͰ͖ͳ͍ ༧ଌ݁ՌΛฦ͢APIͱͯ͠ESΛར༻͢ΔͳͲ͍ͯͨ͠ x MLͷ͓ͳ΍Έ

Slide 22

Slide 22 text

AIࣨͷ͓ͳ΍Έ શࣾͷ༷ʑͳࣄۀͱ࿈ܞ ෳ਺ͷࣄۀ͕͋ΓɺػցֶशͷػೳΛఏڙ͍ͯ͠Δ ֤ࣄۀͱσʔλ࿈ܞͷI/FΛڞ௨Խ͍ͨ͠ ։ൃϑϩʔΛڞ௨Խ͍ͨ͠ ઐ೚ͷΠϯϑϥΤϯδχΞ΋͍ͳ͍ ఏڙ͢Δػೳ͝ͱʹɺݸผʹΠϯϑϥΛ੔͑Δͱແବ͕ଟ͍ ग़དྷΔݶΓϑϨʔϜϫʔΫԽ͍ͨ͠

Slide 23

Slide 23 text

AIࣨͷ͓ͳ΍Έ શࣾͷ༷ʑͳࣄۀͱ࿈ܞ ෳ਺ͷࣄۀ͕͋ΓɺػցֶशͷػೳΛఏڙ͍ͯ͠Δ ֤ࣄۀͱσʔλ࿈ܞͷI/FΛڞ௨Խ͍ͨ͠ ։ൃϑϩʔΛڞ௨Խ͍ͨ͠ ઐ೚ͷΠϯϑϥΤϯδχΞ΋͍ͳ͍ ఏڙ͢Δػೳ͝ͱʹɺݸผʹΠϯϑϥΛ੔͑Δͱແବ͕ଟ͍ ग़དྷΔݶΓϑϨʔϜϫʔΫԽ͍ͨ͠

Slide 24

Slide 24 text

Solution ղܾࡦΛ୳͠·ͨ͠

Slide 25

Slide 25 text

No content

Slide 26

Slide 26 text

4 3FETIJGU 3%4ͷσʔλ͔Β ֶशϞσϧΛ࡞੒Ͱ͖Δ ༧ଌͷͨΊʹΫΤϦΛ࣮ߦͰ͖Δ ֶशϞσϧ͔Β༧ଌ஋Λฦ͢ 8FCΞϓϦέʔγϣϯΛࣗಈੜ੒

Slide 27

Slide 27 text

Solution

Slide 28

Slide 28 text

Solution Machine Leaning as a Service ML Tools

Slide 29

Slide 29 text

Solution Machine Leaning as a Service ML Tools Open Source

Slide 30

Slide 30 text

Machine Learning Stacks Apps Algorithm Processing Datastore API Server (Tornado…) Scikitlearn, SparkML … DL: Caffe2, DL4j, Tensorﬂow, Chainer … Hadoop, Spark, Storm … Elasticsearch, HBASE, Redshift …

Slide 31

Slide 31 text

Machine Learning Stacks Apps Algorithm Processing Datastore API Server (Tornado…) Scikitlearn, SparkML … DL: Caffe2, DL4j, Tensorﬂow, Chainer … Hadoop, Spark, Storm … Elasticsearch, HBASE, Redshift … PredictionIO

Slide 32

Slide 32 text

No content

Slide 33

Slide 33 text

PredictionIO?

Slide 34

Slide 34 text

Salesforce Acquires PredictionIO

Slide 35

Slide 35 text

Salesforce Acquires PredictionIO Feb 19, 2016 - TechCrunch 4BMFTGPSDF͕ػցֶशϓϥοτϑΥʔϜͷ 1SFEJDUJPO*0Λങऩ

Slide 36

Slide 36 text

Salesforce Introduces Salesforce Einstein

Slide 37

Slide 37 text

Salesforce Introduces Salesforce Einstein Sep 18, 2016 - TechCrunch "*ΛऔΓࠐΉ4BMFTGPSDFͷ໺๬ ػցֶशϓϥοτϑΥʔϜʮ&JOTUFJOʯΛൃද

Slide 38

Slide 38 text

The most stars repositories on Github? spark apache/spark ˒ 12.8k incubator-predictionio apache/incubator-predictionio ˒ 10.2k playframework playframework/playframework ˒ 9.3k scala scala/scala ˒ 8.2k

Slide 39

Slide 39 text

spark apache/spark ˒ 12.8k incubator-predictionio apache/incubator-predictionio ˒ 10.2k playframework playframework/playframework ˒ 9.3k scala scala/scala ˒ 8.2k The most stars repositories on Github? ˒10.2k

Slide 40

Slide 40 text

What is PredictionIO?

Slide 41

Slide 41 text

Apache PredictionIO? Apache PredictionIO (incubating) is an open source Machine Learning Server built on top of state-of-the-art open source stack for developers and data scientists create predictive engines for any machine learning task.

Slide 42

Slide 42 text

Apache PredictionIO (incubating) is an open source Machine Learning Server built on top of state-of-the-art open source stack for developers and data scientists create predictive engines for any machine learning task. Apache PredictionIO? ࠷ઌ୺ͷΦʔϓϯιʔεΛ ૊߹Θͤͨػցֶशαʔό

Slide 43

Slide 43 text

Apache PredictionIO (incubating) is an open source Machine Learning Server built on top of state-of-the-art open source stack for developers and data scientists create predictive engines for any machine learning task. Apache PredictionIO? ࠷ઌ୺ͷΦʔϓϯιʔεΛ ૊߹Θͤͨػցֶशαʔό ͲΜͳػցֶशλεΫͰ΋ ༧ଌΤϯδϯ͕ͭ͘ΕΔ

Slide 44

Slide 44 text

Apache PredictionIO let you ର৅໰୊͝ͱʹςϯϓϨʔτΛ࡞Γɺ  ͙͢ʹσϓϩΠͰ͖Δ quickly build and deploy an engine as a web service on production with customizable templates; ΫΤϦ౤͛ͯ݁ՌΛฦ͢API͕͋Δ respond to dynamic queries in real-time once deployed as a web service;

Slide 45

Slide 45 text

Apache PredictionIO let you ޡࠩͷௐ੔΍ɺධՁͷ࢓૊Έ΋͋Δ evaluate and tune multiple engine variants systematically; όον or ϦΞϧλΠϜͰ  ֶशσʔλΛొ࿥͢ΔI/F͕͋Δ unify data from multiple platforms in batch or in real-time for comprehensive predictive analytics;

Slide 46

Slide 46 text

No content

Slide 47

Slide 47 text

No content

Slide 48

Slide 48 text

No content

Slide 49

Slide 49 text

REST: EventAPI SDK: EventClient

Slide 50

Slide 50 text

No content

Slide 51

Slide 51 text

Engine Template

Slide 52

Slide 52 text

No content

Slide 53

Slide 53 text

Photo by Bernard Spragg. NZ Quick Start

Slide 54

Slide 54 text

Versions Latest Release Version v0.11.0

Slide 55

Slide 55 text

Quick Startʢ೔ຊޠʣ takezoe.hatenablog.com/entry/2017/05/11/132410

Slide 56

Slide 56 text

Photo by Bernard Spragg. NZ Algorithm

Slide 57

Slide 57 text

No content

Slide 58

Slide 58 text

Photo by Bernard Spragg. NZ System Architecture

Slide 59

Slide 59 text

System Architecture Apache Hadoop up to 2.7.2 required only if YARN and HDFS are needed  Apache HBase up to 1.2.4 Apache Spark up to 1.6.3  for Hadoop 2.6 not Spark 2.x version Elasticsearch up to 1.7.5 not the Elasticsearch 2.x version

Slide 60

Slide 60 text

Storage roles Meta Data Event Data Model Data ✓ ✓ ✓ ✓ ✓* ✓ ✓ LOCALFS ✓

Slide 61

Slide 61 text

Photo by Bernard Spragg. NZ Implementation

Slide 62

Slide 62 text

Scala ੡ ɾ ػցֶशج൫ PredictionIOͱ SparkʹΑΔϨίϝϯυγ ες Ϝ

Slide 63

Slide 63 text

x ML

Slide 64

Slide 64 text

System Requirements ελϯόΠͷϨίϝϯυཁ݅ ϢʔβͷΫϦοΫϩά ͓ؾʹೖΓ௥Ճϩά S3ʹϩάσʔλ্͕͕͍ͬͯΔ ֶश͸೔࣍Ͱ

Slide 65

Slide 65 text

ElasticsearchͱTasteϓϥάΠϯͰ ࡞ΔϨίϝϯυγεςϜ

Slide 66

Slide 66 text

PIOಋೖલ σʔλ༻ͷESΠϯσοΫΛຖճ࡞੒ σʔλΠϯϙʔτ༻ͷઐ༻εΫϦϓτ Elasticsearch TasteϓϥάΠϯͰ࣮ߦ ศར͕ͩ൚༻ੑɺσʔλ૿ʹΑΔ࣮ߦ͕࣌ؒ՝୊ʹ ֶश݁ՌΛόϧΫϑΝΠϧͱͯ͠ग़ྗ Elasticsearch ༧ଌͷAPIͱͯ͠ར༻ ࣮ߦϑϩʔΛγΣϧεΫϦϓτͰ؅ཧ

Slide 67

Slide 67 text

PIOಋೖલ σʔλ༻ͷESΠϯσοΫΛຖճ࡞੒ σʔλΠϯϙʔτ༻ͷઐ༻εΫϦϓτ Elasticsearch TasteϓϥάΠϯͰ࣮ߦ ศར͕ͩ൚༻ੑɺσʔλ૿ʹΑΔ࣮ߦ͕࣌ؒ՝୊ʹ ֶश݁ՌΛόϧΫϑΝΠϧͱͯ͠ग़ྗ Elasticsearch ༧ଌͷAPIͱͯ͠ར༻ ࣮ߦϑϩʔΛγΣϧεΫϦϓτͰ؅ཧ

Slide 68

Slide 68 text

Click Log Favorite Log Event Server ALS Template

Slide 69

Slide 69 text

Click Log Favorite Log Event Server ALS Template pio import

Slide 70

Slide 70 text

Click Log Favorite Log Elasticsearch v5.3 cluster Event Server ALS Template pio import Data

Slide 71

Slide 71 text

Click Log Favorite Log Elasticsearch v5.3 cluster Event Server ALS Template pio import Data Spark 2 node cluster RDD

Slide 72

Slide 72 text

Click Log Favorite Log Elasticsearch v5.3 cluster Event Server ALS Template pio import Data LOCALFS Spark 2 node cluster RDD Model

Slide 73

Slide 73 text

Click Log Favorite Log Elasticsearch v5.3 cluster Event Server ALS Template pio import Data LOCALFS Spark 2 node cluster RDD Model Query Predicted Result

Slide 74

Slide 74 text

Engine Template?

Slide 75

Slide 75 text

Engine Template

Slide 76

Slide 76 text

D A S E D-A-S-E Data Source and Data Preparator Algorithm Serving Evaluation Metrics

Slide 77

Slide 77 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌॲཧ Prediction Server ༧ଌ݁Ռ Predicted Result

Slide 78

Slide 78 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌॲཧ Prediction Server ༧ଌ݁Ռ Predicted Result

Slide 79

Slide 79 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌॲཧ Prediction Server ༧ଌ݁Ռ Predicted Result

Slide 80

Slide 80 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌॲཧ Prediction Server ༧ଌ݁Ռ Predicted Result

Slide 81

Slide 81 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌॲཧ Prediction Server ༧ଌ݁Ռ Predicted Result

Slide 82

Slide 82 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌॲཧ Prediction Server ༧ଌ݁Ռ Predicted Result

Slide 83

Slide 83 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D

Slide 84

Slide 84 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D Algorithm A

Slide 85

Slide 85 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D Algorithm A Serving S

Slide 86

Slide 86 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D Algorithm A Serving S E Evaluation Metrics

Slide 87

Slide 87 text

No content

Slide 88

Slide 88 text

D

Slide 89

Slide 89 text

D A

Slide 90

Slide 90 text

No content

Slide 91

Slide 91 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D Algorithm A Serving S E Evaluation Metrics

Slide 92

Slide 92 text

D

Slide 93

Slide 93 text

DataSource •Event Store (Event Server) ͔ΒσʔλΛಡࠐ •TrainingDataΛฦ͢

Slide 94

Slide 94 text

No content

Slide 95

Slide 95 text

No content

Slide 96

Slide 96 text

D

Slide 97

Slide 97 text

Preparator • TrainingDataʹର͢Δલॲཧ • ಛ௃நग़ • ෳ਺AlgorithmΛར༻͢Δ৔߹ͷڞ௨ॲཧ • PreparedDataʹม׵ͯ͠Algoritmʹ౉͢

Slide 98

Slide 98 text

No content

Slide 99

Slide 99 text

No content

Slide 100

Slide 100 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D Algorithm A Serving S E Evaluation Metrics

Slide 101

Slide 101 text

A

Slide 102

Slide 102 text

Algorithm • train() Λ࣮૷ • ༧ଌϞσϧͷֶशΛ୲౰͢Δ • pio train ίϚϯυͰݺͼग़͞ΕΔ • HDFSʢLocalFSʣʹετΞ͞ΕΔ • predict() Λ࣮૷ • σϓϩΠޙͷΫΤϦʹରͯ͠ϦΞϧλΠϜʹݺ͹ΕΔ

Slide 103

Slide 103 text

No content

Slide 104

Slide 104 text

No content

Slide 105

Slide 105 text

No content

Slide 106

Slide 106 text

No content

Slide 107

Slide 107 text

No content

Slide 108

Slide 108 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D Algorithm A Serving S E Evaluation Metrics

Slide 109

Slide 109 text

No content

Slide 110

Slide 110 text

Serving • LServeΛܧঝ • serve() Λ࣮૷

Slide 111

Slide 111 text

No content

Slide 112

Slide 112 text

Machine Learning Flow τϨʔχϯάσʔλ Training Data ػցֶशΞϧΰϦζϜ Machine Learning Algorithm ༧ଌϞσϧ Predictive Model લॲཧ Preprocessing Πϯϓοτσʔλ Input Data ༧ଌϞσϧ Predictive Model ༧ଌ݁Ռ Predicted Result Data Source & Preparator D Algorithm A Serving S E Evaluation Metrics

Slide 113

Slide 113 text

Precision@k Precision@5 / Threshold = 2.0 Predicted A ˒ˑˑ Validation B ˒˒˒ C ˒˒ˑ D ˑˑˑ E ˒˒ˑ A ˒ˑˑ B ˒˒˒ X ˒˒ˑ D ˑˑˑ E ˒˒ˑ

Slide 114

Slide 114 text

Precision@k Precision@5 / Threshold = 2.0 Predicted A ˒ˑˑ Validation B ˒˒˒ C ˒˒ˑ D ˑˑˑ E ˒˒ˑ A ˒ˑˑ B ˒˒˒ X ˒˒ˑ D ˑˑˑ E ˒˒ˑ

Slide 115

Slide 115 text

Precision@k Precision@5 / Threshold = 2.0 Predicted A ˒ˑˑ Validation B ˒˒˒ C ˒˒ˑ D ˑˑˑ E ˒˒ˑ A ˒ˑˑ B ˒˒˒ X ˒˒ˑ D ˑˑˑ E ˒˒ˑ PositiveCount: 2.0

Slide 116

Slide 116 text

x ML 5 Jobs

Slide 117

Slide 117 text

No content

Slide 118

Slide 118 text

No content

Slide 119

Slide 119 text

Photo by Bernard Spragg. NZ Conclusion

Slide 120

Slide 120 text

τϨʔχϯάσʔλ ϦΞϧλΠϜͰ΋ɺόονͰ΋σʔλΛऔΓࠐΉ͜ͱ͕Ͱ͖Δ ΞΫηετʔΫϯΛൃߦͰ͖ΔͷͰɺ֤αʔϏεͱͷ࿈ܞ͕ศར Elasticsearchͷ෼ࢄετϨʔδͷػೳΛڗडͰ͖Δ ֶशॲཧͷ࣮ߦ࣌ؒ SparkͷΫϥελΛ࢖͏ͨΊɺॲཧΛ෼ࢄֶ͠शʹ͔͔Δ࣌ؒΛ୹ॖ Open Source Machine Learning Server

Slide 121

Slide 121 text

ֶशϞσϧͷετϨʔδ ελϯόΠͰ͸LOCALFSΛར༻͍ͯ͠Δ  ϞσϧͷಛੑʹԠͯ͡HDFSΛબ୒Մೳ ༧ଌͷWeb API “pio deploy” ίϚϯυ͚ͩͰ༧ଌͷAPIΛ࡞੒Ͱ͖Δ APIαʔό͸Akka-Httpϕʔε Open Source Machine Learning Server

Slide 122

Slide 122 text

୤ɾଐਓԽ

Slide 123

Slide 123 text

Photo by Bernard Spragg. NZ Appendix

Slide 124

Slide 124 text

ϏζϦʔνͰ͸ػցֶशΞϓϦέʔγϣϯΛ ૊৫తʹ։ൃ͍ͯ͘͠ʹ͋ͨΓɺ։ൃɾӡ༻ج൫ͱ ͯ͠"QBDIF1SFEJDUJPO*0ʹऔΓ૊ΜͰ͍·͢ɻ ೥݄ฐ͔ࣾΒ໊͕ίϛολͱͯ͠ 1SFEJDUJPO*0ͷ։ൃʹࢀՃ͢Δ͜ͱʹͳΓ·ͨ͠ɻ Shinsuke Sugaya Naoki  Takezoe Takako  Shimamoto Takahiro Hagino

Slide 125

Slide 125 text

ϏζϦʔνͰ͸ػցֶशΞϓϦέʔγϣϯΛ ૊৫తʹ։ൃ͍ͯ͘͠ʹ͋ͨΓɺ։ൃɾӡ༻ج൫ͱ ͯ͠"QBDIF1SFEJDUJPO*0ʹऔΓ૊ΜͰ͍·͢ɻ ೥݄ฐ͔ࣾΒ໊͕ίϛολͱͯ͠ 1SFEJDUJPO*0ͷ։ൃʹࢀՃ͢Δ͜ͱʹͳΓ·ͨ͠ɻ Shinsuke Sugaya Naoki  Takezoe Takako  Shimamoto Takahiro Hagino

Slide 126

Slide 126 text

How to Contribute to PIO

Slide 127

Slide 127 text

Add support for Elasticsearch 5.x

Slide 128

Slide 128 text

No content

Slide 129

Slide 129 text

jpioug.org

Slide 130

Slide 130 text

● ೔ຊ Apache PredictionIO Ϣʔβձ JPIOUG Join Us!

Slide 131

Slide 131 text

No content

Slide 132

Slide 132 text

30 6݄ FRI

Slide 133

Slide 133 text

30 6݄ FRI Open Source Machine Learning Server 02 JPIOUG Meetup 19:30 @Shibuya

Slide 134

Slide 134 text

30 6݄ FRI Open Source Machine Learning Server 02 JPIOUG Meetup 19:30 @Shibuya ʲٸืʳLT͍ͨ͠ํ

Slide 135

Slide 135 text

30 8݄ WED

Slide 136

Slide 136 text

30 8݄ WED Open Source Machine Learning Server 03 JPIOUG Meetup 19:30 @Shibuya

Slide 137

Slide 137 text

30 8݄ WED Open Source Machine Learning Server 03 JPIOUG Meetup 19:30 @Shibuya ʲΏΔืʳLT͍ͨ͠ํ

Slide 138

Slide 138 text

jpioug.org

Slide 139

Slide 139 text

Open Source Machine Learning Server 15 Machine Learning minutes! Thank You