Machine learning and TensorFlow

94
Machine Learning and TensorFlow Artificial Intelligence Present and Future José Papo Gerente de relações com startups e developers Google América Latina @josepapo

Transcript of Machine learning and TensorFlow

Page 1: Machine learning and TensorFlow

Machine Learning and TensorFlowArtificial Intelligence Present and Future

José PapoGerente de relações com startups e developersGoogle América Latina@josepapo

Page 3: Machine learning and TensorFlow

“Machine learning is a core, transformative way by which we’re re-thinking how we’re doing everything”

Sundar PichaiCEO, Google

Page 4: Machine learning and TensorFlow

“Machine learning will cause every successful huge IPO win in 5 years.”

Eric SchmidtExecutive Chairman, Alphabet

Page 5: Machine learning and TensorFlow

Basic Concepts

Page 6: Machine learning and TensorFlow

● Artificial General Intelligence

● Artificial Superintelligence

● Artificial Narrow Intelligence

Artificial Intelligence

Page 7: Machine learning and TensorFlow

Machine Learning (Narrow AI)

Page 8: Machine learning and TensorFlow

Deep Learning (ML on Steroids!!!)

Page 9: Machine learning and TensorFlow
Page 10: Machine learning and TensorFlow

What’s different now from 10 years ago?

WAY MORE DATA

More Compute

Better Algorithms

Page 11: Machine learning and TensorFlow
Page 12: Machine learning and TensorFlow

Machine Learning at Google

Page 13: Machine learning and TensorFlow

Products using Machine Learning

Page 14: Machine learning and TensorFlow
Page 15: Machine learning and TensorFlow
Page 16: Machine learning and TensorFlow
Page 17: Machine learning and TensorFlow
Page 18: Machine learning and TensorFlow
Page 19: Machine learning and TensorFlow

TensorFlow

Page 21: Machine learning and TensorFlow

● Open source Machine Learning library

● Especially useful forDeep Learning

● For research and production

● Apache 2.0 license

Page 22: Machine learning and TensorFlow

Raspberry PiDatacentersYour laptop Android iOS

Portable & Scalable

Page 23: Machine learning and TensorFlow

A multidimensional array.

A graph of operations.

Page 24: Machine learning and TensorFlow

Data Flow Graphs

Computation is defined as a directed acyclic graph (DAG) to optimize an objective function

● Graph is defined in high-level language (Python)● Graph is compiled and optimized● Graph is executed (in parts or fully) on available low

level devices (CPU, GPU)● Data (tensors) flow through the graph● TensorFlow can compute gradients automatically

Page 25: Machine learning and TensorFlow
Page 26: Machine learning and TensorFlow

Image source: Wikimedia

+ =

A Neural Algorithm of Artistic Style http://arxiv.org/abs/1508.06576

?

Page 27: Machine learning and TensorFlow

Image source: Wikimedia

+ =

A Neural Algorithm of Artistic Style http://arxiv.org/abs/1508.06576

Page 28: Machine learning and TensorFlow

Image source: Wikimedia

+ =

A Neural Algorithm of Artistic Style http://arxiv.org/abs/1508.06576

goo.gl/fyDxhC

Page 29: Machine learning and TensorFlow

Most popular ML open source project on GitHub

Page 30: Machine learning and TensorFlow

Cloud, Mobile, Machine Learning

Page 31: Machine learning and TensorFlow

Cloud Machine Learning APIs See, Hear and Understand the world

Page 32: Machine learning and TensorFlow

Cloud Natural Language

Cloud Speech

Cloud Translate

Cloud Vision

Page 33: Machine learning and TensorFlow

FacesFaces, facial landmarks, emotions

OCRRead and extract text, with support for > 10 languages

LabelDetect entities from furniture to transportation

LogosIdentify product logos

Landmarks & Image PropertiesDetect landmarks & dominant color of image

Safe SearchDetect explicit content - adult, violent, medical and spoof

Cloud Vision API

Page 34: Machine learning and TensorFlow

Confidential & ProprietaryGoogle Cloud Platform 34

Cloud Natural Language API

Extract sentence, identify parts of speech and create dependency parse trees for each sentence.

Identify entities and label by types such as person, organization, location, events, products and media.

Understand the overall sentiment of a block of text.

Syntax Analysis Entity Recognition

Sentiment Analysis

Page 35: Machine learning and TensorFlow

Confidential & ProprietaryGoogle Cloud Platform 35

Cloud Speech API

Automatic Speech Recognition (ASR) powered by deep learning neural networking to power your applications like voice search or speech transcription.

Recognizes over 80 languages and variants with an extensive vocabulary.

Returns partial recognition results immediately, as they become available.

Filter inappropriate content in text results.

Audio input can be captured by an application’s microphone or sent from a pre-recorded audio file. Multiple audio file formats are supported, including FLAC, AMR, PCMU and linear-16.

Handles noisy audio from many environments without requiring additional noise cancellation.

Audio files can be uploaded in the request and, in future releases, integrated with Google Cloud Storage.

Automatic Speech Recognition Global Vocabulary Inappropriate Content Filtering

Streaming Recognition

Real-time or Buffered Audio Support Noisy Audio Handling Integrated API

Page 36: Machine learning and TensorFlow

Mobile Vision APIProviding on-device vision for applications

Page 37: Machine learning and TensorFlow

Face APIfaces, facial landmarks, eyes open, smiling

Barcode API1D and 2D barcodes

Text APILatin-based text / structure

Common Mobile Vision APISupport for fast image and video on-device detection and tracking.

NEW!

Page 38: Machine learning and TensorFlow

Face APIPhoto credit developers.google.com/vision

Page 39: Machine learning and TensorFlow

Text DetectionLatin based languageUnderstand text structure

Photo credit Getty Images

Page 40: Machine learning and TensorFlow

Barcode Detection1D barcodes

EAN-13/8 UPC-A/E Code-39/93/128 ITF Codabar

2D barcodes QR Code Data Matrix PDF-417 AZTEC

UPC

DataMatrix

QR Code

PDF 417

Video and image credit Google

Page 41: Machine learning and TensorFlow

Machine Learning DemocratizationUse Cases in Latin America

Page 42: Machine learning and TensorFlow
Page 43: Machine learning and TensorFlow
Page 44: Machine learning and TensorFlow
Page 45: Machine learning and TensorFlow
Page 46: Machine learning and TensorFlow
Page 47: Machine learning and TensorFlow
Page 48: Machine learning and TensorFlow
Page 49: Machine learning and TensorFlow
Page 50: Machine learning and TensorFlow
Page 51: Machine learning and TensorFlow
Page 52: Machine learning and TensorFlow
Page 53: Machine learning and TensorFlow
Page 54: Machine learning and TensorFlow
Page 55: Machine learning and TensorFlow
Page 56: Machine learning and TensorFlow
Page 57: Machine learning and TensorFlow
Page 58: Machine learning and TensorFlow

ACESSO UNIVERSAL A MEDICINA DE QUALIDADE

Page 59: Machine learning and TensorFlow

Machine Learning

Page 60: Machine learning and TensorFlow

AGENDA

• Otimização do broadcast

• Otimização do processo billing

• Personal Cloud

Machine Learning

Page 61: Machine learning and TensorFlow

operacional

comercial

usuários

3perspectiva

s

Page 62: Machine learning and TensorFlow

Otimização do broadcast

• Reduzir a quantidade de envio de mensagens de estímulo mantendo a mesma taxa de retorno.

comercial Desafio

Proposta• Identificar o comportamento ou características dos

usuários mais propensos a responder ao estímulo.

Page 63: Machine learning and TensorFlow

Otimização do broadcast

• Text Mining para tratamento das frases, classificando-as, como por exemplo, pela ideia transmitida.

• Análise de modelos preditivos para seleção dos clientes mais propensos.

comercial

Processo de análise

Page 64: Machine learning and TensorFlow

Otimização do broadcast

• Prever quem não irá responder a nossa oferta nos dá a possibilidade de pensarmos em algo diferente para este usuário e desta forma conhecê-lo um pouco mais.

• Redução de média 40% nos envios de broadcast.

comercial

Resultado

Page 65: Machine learning and TensorFlow

Otimização do billing

• Aumentar o sucesso nas cobranças dos serviços prestados.

Desafio

Proposta

• Identificar os clientes mais propensos em determinados horários.

operacional

Page 66: Machine learning and TensorFlow

Otimização do billing

• Tratamento e enriquecimento da base de dados com BigQuery.

• Análise de modelos preditivos para criação de escore de crédito.

Processo de análise

operacional

Page 67: Machine learning and TensorFlow

Otimização do billingProposição de uso

Page 68: Machine learning and TensorFlow

Otimização do billing

• Redução de custos com infraestrutura de TI, uso mais inteligente de recursos.

• Melhora de 42% em média na acertividade do billing.

Resultado

operacional

Page 69: Machine learning and TensorFlow

Personal Cloud

• Detectar objetos e faces dentro das fotos dos usuários do Personal Cloud para possibilitar busca e criação de álbuns de forma automática

Desafio

Proposta

• Utilização da API do Google Cloud Vision.

usuários

Busca por tags e álbuns automáticos

Page 70: Machine learning and TensorFlow

Otimização do broadcast

Processo de análiseusuário

s pé

dedo

bolsa

óculos

praia

Page 71: Machine learning and TensorFlow

Don’t Think Outside The Box, Think Like There is NO

BOX!

Page 72: Machine learning and TensorFlow
Page 73: Machine learning and TensorFlow
Page 74: Machine learning and TensorFlow
Page 75: Machine learning and TensorFlow
Page 76: Machine learning and TensorFlow
Page 77: Machine learning and TensorFlow
Page 78: Machine learning and TensorFlow
Page 79: Machine learning and TensorFlow
Page 80: Machine learning and TensorFlow
Page 81: Machine learning and TensorFlow
Page 82: Machine learning and TensorFlow
Page 83: Machine learning and TensorFlow
Page 84: Machine learning and TensorFlow
Page 85: Machine learning and TensorFlow
Page 86: Machine learning and TensorFlow
Page 87: Machine learning and TensorFlow
Page 88: Machine learning and TensorFlow
Page 89: Machine learning and TensorFlow
Page 90: Machine learning and TensorFlow
Page 91: Machine learning and TensorFlow
Page 92: Machine learning and TensorFlow
Page 93: Machine learning and TensorFlow

tensorflow.orggithub.com/tensorflow

Want to learn more?Udacity class on Deep Learning, goo.gl/iHssII

Guides, codelabs, videosMNIST for Beginners, goo.gl/tx8R2bTF Learn Quickstart, goo.gl/uiefRnTensorFlow for Poets, goo.gl/bVjFILML Recipes, goo.gl/KewA03TensorFlow and Deep Learning without a PhD, goo.gl/pHeXe7

Learn More...

Page 94: Machine learning and TensorFlow

OBRIGADO!slideshare.net/jpapo

José Papo @josepapo