1. The document discusses Microsoft's artificial intelligence capabilities including cognitive services, bots, Azure machine learning, and tools for building AI applications and managing models.
2. It provides an overview of Microsoft's offerings for AI including cognitive services, bots, Azure machine learning studio for building experiments visually, and Azure machine learning services for experimentation and model management.
3. The document emphasizes that Microsoft's goal is to make AI accessible and useful to every person and organization by providing a broad set of tools, frameworks, and infrastructure for developing, training, deploying, and managing AI applications and models.
1 of 44
More Related Content
Borys Rybak “How to make your data smart with Artificial Intelligence and Machine Learning services”
7. Flexible infrastructure
support for analytics
Best of Microsoft research
and open source
Most comprehensive
data science capabilities
Easy to consume
Artificial Intelligence
Solutions Extensible applications
8. Services
Processing
Frameworks
CPUs
AI Applications
Cognitive Services
Infrastructure
AML Web Services BOT Framework
Model & Experimentation
Management
Spark AI Batch
Training
Storage
COSMOS DB ADLSSQL DB SQL DWBLOB
GPUs
FPGA
IOT
Inferencing
Spark, SQL,
Other Engines
DSVM
EDGE
PROSE Data Wrangling
Machine Learning Toolkits
CNTK Tensorflow ML Server Scikit-Learn Other Libs.
ACS
Docker
15. Roll-your-own with REST APIs
Simple to add: just a few
lines of code required
Integrate into the language
and platform of your choice
Breadth of offerings helps you
find the right API for your app
Bring your own data for your
custom experience
Built by experts in their field
from Microsoft Research, Bing,
and Azure Machine Learning
Quality documentation, sample code,
and community support
Easy Flexible Tested
Why Microsoft
Cognitive Services?
KLUCZ
TWÓRZ
17. Microsoft Cognitive Services
Give your apps a human side
KnowledgeLanguage LabsSearch
Computer Vision
Content Moderator
Emotion
Face
Video
Video Indexer
Project Prague
(gesture)
Project Cuzco (events)
Project Johannesburg
(routing)
Project Nanjing
(isochrones)
Project Abu Dhabi
(distance matrix)
Project Wollongong
(location)
Bing
Autosuggest
Bing Image
Search
Bing News
Search
Bing Video
Search
Bing Web Search
Bing Entity
Search
Academic
Knowledge
Entity Linking
Knowledge
Exploration
Recommendations
QnA Maker
Bing Spell Check
Linguistic Analysis
Text Analytics
Translator Text
& Speech
Web Language
Model
Bing Speech
Speaker
Recognition
21. Customer Service and Support
Make Customer Service Bots more
human and Live Agents more
productive with Enterprise level AI.
I am having trouble setting up a new projector
with my laptop
23. Engage with your users where they already are
Bot Framework: a natural language interface across all conversation channels
Provide information
Perform tasks
Make recommendations
Capture information
OperationalizationInsights
24. Bot Directory
Your Bot Framework
Try, use, and add published bots to the
world’s top conversation experiences.
Developer Portal
Connect your bots to text/sms, Skype,
Slack, Facebook Messenger,
Office 365 mail and other channels.
• Register, connect, publish and
manage your bot through your
bot’s dashboard
• Automatic card normalization
across channels
• Skype channel auto-configured
• Embeddable Web chat control
• Host your bot in your app via
the Direct Line API
• Fast, scalable message routing
• Diagnostic tools
Bot Builder
Tools and services to build great bots
that converse wherever your users are.
• Open source SDK on Github for
Node.js, .NET and REST
• From simple built-in prompts
and command dialogs to
simple to use yet sophisticated
‘FormFlow’ dialogs
• Support for rich attachments
(image, card, video, doc, etc.);
support for calling (Skype)
• Online/offline chat Emulator
• Add bot smarts with Cognitive
Services for language
understanding and more
• Public directory of bots
registered and published with
Microsoft Bot Framework
• Users can try your bot from the
directory via the Web chat
control
• Users can discover and add
your bot to the channels on
which it is configured when the
Directory is made public to end
users
25. Cognitive Toolkit
Unlock deeper learning
A free, easy-to-use, open-source toolkit that
trains deep learning algorithms to learn like the
human brain.
Microsoft Cognitive Toolkit
29. Apps + insights
Social
LOB
Graph
IoT
Image
CRM INGEST STORE PREP & TRAIN MODEL & SERVE
Data orchestration
and monitoring
Data lake
and storage
Hadoop/Spark/SQL
and ML
.
IoT
Azure Machine Learning
T H E A I D E V E L O P M E N T L I F E C Y C L E
30. Azure Machine Learning Studio
Platform for data scientists to graphically
build and deploy experiments
Rapid experiment composition
> 100 easily configured modules for
data prep, training, evaluation
Extensibility through R & Python
Serverless training and deployment
Some numbers:
100’s of thousands of deployed models
serving billions of requests
31. What have we learned?
Customers have told us they love the convenience…
…Customers have told us they need:
Greater control over compute & data
More options for model deployment
Which frameworks? ALL OF THEM!
32. Notebooks
IDEs
Azure Machine Learning
Workbench
VS Code Tools for AI
N E W C A P A B I L I T I E S
Experimentation and
Model Management
Services
AZURE MACHINE
LEARNING SERVICES
Spark
SQL Server
Virtual
machines
GPUs
Container
services
SQL Server
Machine Learning
Server
ON-PREMISES
EDGE
Azure IoT
Edge
TRAIN & DEPLOY
OPTIONS
AZURE
33. What Did I Get?
Experimentation Account:
Keep track of projects
Local, scale-up, and scale-out training
Run history tracking
Model Management Account:
Create containers for models
Manage and monitor deployed models
34. Local machine
Scale up to DSVM
Scale out with Spark on HDInsight
Azure Batch AI (Coming Soon)
ML Server
Experiment Everywhere
A ZURE ML
EXPERIMENTATION
Command line tools
IDEs
Notebooks in Workbench
VS Code Tools for AI
35. Manage project dependencies
Manage training jobs locally, scaled-up or
scaled-out
Git based checkpointing and version control
Service side capture of run metrics, output
logs and models
Use your favorite IDE, and any framework
Experimentation service
U S E T H E M O S T P O P U L A R I N N O VAT I O N S
U S E A N Y TO O L
U S E A N Y F R A M E W O R K O R L I B R A R Y
37. Deployment and management of models as
HTTP services
Container-based hosting of real time and
batch processing
Management and monitoring through Azure
Application Insights
First class support for SparkML, Python,
Cognitive Toolkit, TF, R, extensible to support
others (Caffe, MXnet)
Service authoring in Python
Manage models
39. Windows and Mac based
companion for AI development
Full environment set up (Python,
Jupyter, etc)
Embedded notebooks
Run History and Comparison
experience
New data wrangling tools
What Is It?
40. AI Powered Data Wrangling
Rapidly sample, understand,
and prep data
Leverage PROSE and more for
intelligent, data prep by
example
Extend/customize transforms
and featurization through
Python
Generate Python and Pyspark
for execution at scale
41. Machine Learning & AI Portfolio
When to use what?
What engine(s) do you want
to use?
Deployment target
Which experience do you
want?
Build your own or consume pre-
trained models?
Microsoft
ML & AI
products
Build your
own
Azure Machine Learning
Code-first
(On-prem)
ML Server
On-
prem
Hadoop
SQL
Server
(cloud)
AML services (Preview)
SQL
Server
Spark Hadoop Azure
Batch
DSVM Azure
Container
Service
Visual tooling
(cloud)
AML Studio
Consume
Cognitive services, bots
42. The most critical next step
in our pursuit of A.I. is to
agree on an ethical and
empathic framework for
its design.
SATYA NADELLA
Editor's Notes
Building theIntelligent future.
Transforming the Business withArtificial Intelligence.
Artificial Intelligence is term widely used nowadays but it lacks a common definition.
Wikipedia: “Machines showing capabilities that are typically associated with human intelligence”.
AI nie jest nowe. Byliśmy w podróży rozszerzania ludzkich możliwości przez wiele dekad. Głównie w trzech obszarach. Mechanika. Elektronika. Komputery.
Przenieśmy się do roku 1877 i pomyślmy o pierwszych kalkulatorach które potrafiły działać lepiej niż ludzie, nawet jeśli chodzi o opcje mechaniczne jakim był Kalkulator Granta (1877).
Elektronika otworzyła nowe horyzonty dla świata AI. Naturalne doświadczenie jakim było Voder, maszyna zdolna naśladować ludzki głos. Czy też pierwsze próby wnioskowania, rozumowania oraz interakcji Maszyn (Test Turinga).
Następnie komputery przyniosły graficzne interfejsy użytkownika, rozmpoznawanie mowy czy samojeżdżące auta – wczesny 1994.
Więc dlaczego mówimy o SZTUCZNEJ INTELIGENCJI teraz?
Sam postęp technologiczny przez te wszystkie lata zdawał się rozwijać liniowo, jednakże 3 rzeczy dramatycznie się zmieniły – szybkość innowacji:
Zaawansowanie w algorytmach AI (DNN, czyli deep learning z sieciami neuronowymi)
Potężna moc obliczeniowa dostępna w chmurze.
Big data umożliwiające trenować się tym algorytmom (poteżny zestaw danych do nauki).
Buduj niezwykłe kompetencje za pomocą zestawu Cortana Skills Kit
Nasze narzędzia i technologie umożliwiają tworzenie głęboko zintegrowanych, atrakcyjnych środowisk pracy z Cortaną.
Buduj niezwykłe kompetencje za pomocą zestawu Cortana Skills Kit
Nasze narzędzia i technologie umożliwiają tworzenie głęboko zintegrowanych, atrakcyjnych środowisk pracy z Cortaną.
Rozmawiaj – Rozmawiaj z użytkownikami
Aplikacje – Uruchamiaj kod aplikacji klienckich na urządzeniach
Złożoność – Łańcuch umiejętności do rozwiązywania złożonych problemów
Niestandardowość – Twórz niestandardowe umiejętności za pomocą profilu Cortany
Strumień – Przesyłaj strumieniowo dźwięk do użytkowników
Zasoby – Używaj istniejących zasobów do szybkiego rozpoczęcia tworzenia umiejętności
Microsoft is forging ahead to make FPGA processing power available to external Azure developers for data-intensive tasks like deep-neural-networking tasks.
Why choose these APIs? They work, and it’s easy.
Easy: The APIs are easy to implement because of the simple REST calls. Being REST APIs, there’s a common way to implement and you can get started with all of them for free simply by going to one place, one website, www.microsoft.com/cognitive. (You don’t have to hunt around to different places.)
Flexible: We’ve got a breadth of intelligence and knowledge APIs so developers will be able to find what intelligence feature they need; and importantly, they all work on whatever language, framework, or platform developers choose. So, devs can integrated into their apps—iOS, Android, Windows—using their own tools they know and love (such as python or node.js, etc.).
Tested: Tap into an ever-growing collection of powerful AI algorithms developed by experts. Developers can trust the quality and expertise build into each API by experts in their field from Microsoft’s Research organization, Bing, and Azure machine learning and these capabilities are used across many Microsoft first party products such as Cortana, Bing and Skype.
Project Oxford
Boty, oprogramowanie Office czy usługi kognitywne (czy wszystko naraz) mogą ulepszyć Twoją aplikację i sprawić, że będzie wyjątkowa.
Chatboty – nieludzkie
Agenci – ludzcy (uczucia, etc.)
Frictionless human-like conversations
Seamless integration that enables contextual dialogue
Intelligence built on deep reinforcement learning
Darmowy, łatwy w użyciu, otwarty zestaw narzędzi, który trenuje algorytmy głębokiego uczenia się, w taki sposób jak uczy się ludzki mózg.
For whom like to do some research about these teams’ work, LeNet (1998), AlexNet (2012) , GoogleNet(2014), VGGNet (2014), ResNet(2015) are worth to look. Each of these network architectures have unique approach to different problems. For example, AlexNet has parallel two CNN line trained on two GPUs with cross-connections, GoogleNet has inception modules ,ResNet has residual connections.
According to the paper published in 2015, 152-layer ResNet was the deepest network trained on ImageNet at that time. And as promised it has lower parameter than of VGG Net which is 8x times smaller in depth. This has quite impact on faster training performance.
This improvements results in winning the 1st place in ILSVRC classification competition on ImageNet with 3.57% top 5 error.
https://medium.com/@bakiiii/microsoft-presents-deep-residual-networks-d0ebd3fe5887
For those in the deep learning field, this approach seem familiar. Yes you are right, it is actually similar principle introduced with Long Short Term Memory (LSTM) cells. Long Short Term Memory
Dziś z przyjemnością ogłaszam, że nasz zespół badawczy osiągnął 5,1-procentowy poziom błędu dzięki naszemu systemowi rozpoznawania mowy, nowy kamień milowy w branży, znacznie przewyższający dokładność, jaką osiągnęliśmy w ubiegłym roku.
Centrala telefoniczna to zbiór nagranych rozmów telefonicznych, z których społeczność badaczy mowy korzysta od ponad 20 lat w celu testowania systemów rozpoznawania mowy. Zadanie polega na transkrypcji rozmów między nieznajomymi, którzy dyskutują na takie tematy, jak sport i polityka.
https://catalog.ldc.upenn.edu/ldc97s62
STOPA BŁĘDU.