Big Data to Artificial Intelligence

94,000 people

What is AI? 1950’S. Large computing services and large sets of data. Ai learns from data and can predict things. 1994.  Why should I be using AI: Personalized customer interactions, efficiency and agility, spotting patterns, and adding structure to data. Machine learning for gas. 
Mimic the neurons in the brain. Output layer, activated neurons, input layer. 7500 projects are using deep learning to help users do what they need to do. Maek AI easy, fast, and useful for enterprises and developers. 1.2 million data scientists at kaggle. Infrastructure, API building blocks
Google cloud ai speed through clod tpu. Video scene transitions. Diagoflow – chatbots. Use chatbot before the phone call. Batch processing to stream processing. Google BigQuery. all big query is cinrpied.Predictive analytics let you answer questions based on deep patterns in data. Streaming data, AdWords, google analytics, and google transfer service.. Look into Google BigQuery and it is serverless with unlimited capacity and ease of sharing. Data pre-processing is least fun. Extact, transformation. 
Machine learning – tool for ai
Big data
data mining – bring in to analyze and machine learning
Serverless – don’t have to set up a system, just write a query.
Smarter Marketing with Google Cloud Data Analytics and AI
Google data prep
Data studio to build dashboards in a visual
Data Science Workbooks
Seroko open source tool
Love = trust + anticipation
Language processing – wordeveck
Cloud Vision
OCR – Optcial character recognition
Engineering blog at giphy cloud vision/box
Video – see what happens in every scene
Year in search 2016 google – use these
Scene Level annotations
cloud speech – 100 lang
speech timestamps – audio file when a word is spoken
Cloud translation using python
Cloud natural language
Entity-based sentiment
Use cloud speech to transcribe and translation and translate language. Cloud natrual language from those calls.
Reimaging conversational experiences with cloud speech and dialogflow
Demon on Google next youtube
machine learning is using many examples to answer questions. Training and predictive answers.
Tensorlow is googles machine learning resource. Tensore are mult-dimensional rays, flow through graphs. 
Embedding spaces – represent data
Wide and Deep Google Research papper
memorization relevance, generalization diversity
Pandas Data Exploration
Cloud Machine Learning Engine
play with data, tune parameters, don’t want to bogged down in ops. 
Nobody knows anything
Machine learning artist style to innovate pictures. Tool to create something new.
Every company will be a data company. 80% of all internet traffic will be video. 
Take speech in video and transcript into different languages with speech api in google cloud
Translation is live chat
Data prep – histograms, 
stratisfy sample – find outliers
Create data clean pipelines – remove duplicates
count rows by country transformation
Google Cloud next 17
Mash together data points with python join
12,000 openings – job search problem
Job discovery 
skills, abbreviation, python 
data resources, broad customer research, understanding 
programming, mapping, and navigation for jobs you want
Get emotional reaction from a technology result
Supervised learning – classification
Problem of over fitting – generalize the model
Deep neural networks step 2: testing; artificial neural networks
input out x
out is y
mathematical is the function that translates itM
Artificial mathematical algorithms in white papers
calculus – chain roll, differentiation
linera algebra
quality data and quality analytics
 17 SECONDS TO handle a petabyte of sales data – modern data warehouse
google data studio good for visualizartoin
Streaming dtat

Leave a Reply

Your email address will not be published. Required fields are marked *