Machine Learning on Azure Domain specific pretrained models

Machine Learning on Azure Domain specific pretrained models To simplify solution development … Vision Speech Language Search Visual Studio Code Azure Notebooks Jupyter Command line Py. Torch Tensor. Flow Scikit-Learn ONNX Familiar Data Science tools To simplify model development Popular frameworks To build advanced deep learning solutions Productive services To empower data science and development teams Azure Databricks Azure Machine Learning VMs CPU GPU FPGA Cloud Edge Powerful infrastructure To accelerate deep learning Flexible Deployment Deploy models on Cloud, on-premise, at the edge On Premise

Azure Cognitive Services

Serving more than 1, 300, 000 students and 10, 000 educators from over 1, 100 institutions in 69 countries worldwide.

Azure Databricks

Why Spark? Open-source data processing engine built around speed, ease of use, and sophisticated analytics In memory engine that is up to 100 times faster than Hadoop § Largest open-source data project with 1000+ contributors § Highly extensible with support for Scala, Java and Python alongside Spark SQL, Graph. X, Streaming and Machine Learning Library (Mllib) § §

What is Azure Databricks? Azure Databricks Collaborative Workspace Machine learning models Io. T / streaming data DATA ENGINEER BUSINESS ANALYST DATA SCIENTIST Deploy Production Jobs & Workflows BI tools Cloud storage MULTI-STAGE PIPELINES JOB SCHEDULER NOTIFICATION & LOGS Data warehouses Optimized Databricks Runtime Engine Hadoop storage DATABRICKS I/O APACHE SPARK Data exports SERVERLESS Rest APIs Data warehouses Enhance Productivity Build on secure & trusted cloud Scale without limits

Databricks intro

Azure Machine Learning Service

What can I do with Azure Machine Learning?

AML service Build, train, and deploy models to the cloud and edge Set of Azure Cloud Services Python and R SDK Build and train models faster Operationalize models efficiently Automated machine learning Auto-scaling cloud compute Tool/IDE agnostic Python SDK Open source frameworks support Machine learning life cycle management Machine learning pipelines for CI/CD Deploy models to the cloud and edge

AML Service Workspace

Azure Machine Learning GUI

Ops Integrated with Azure Dev. Ops

Automated ML Benefits Overview Azure Automated ML lets you Automate the exploration process Use resources more efficiently Optimize model for desired outcome Control resource budget Apply it to different models and learning domains Pick training frameworks of choice Visualize all configurations in one place

Automated Hyperparameter Tuning How it works Hyperparameter Tuning runs in Azure ML Launch multiple parallel training runs (A) Generate new runs • Which parameter configuration to explore? #1 #2 ? ? ? … (B) Manage resource usage of active runs • How long to execute a run?

Automated Hyperparameter Tuning Manage Active Jobs Evaluate training runs for specified primary metric Use resources to explore new configurations Early terminate poor performing training runs. Early termination policies include: Early terminate poorly performing runs

Azure Data Explorer

The Platform Azure Data Explorer in a sentence Any appendonly stream of records High volume High velocity High variance (structured, semistructured, free-text) Purposely built © Microsoft Corporation Relational query model: Filter, aggregate, join, calculated columns, … Fullymanaged Rapid iterations to explore the data Paa. S, Vanilla, Database

Azure Data Explorer use cases Io. T applications Discover and address performance issues with machines, equipment, and devices in real-time to optimize production quality and productivity. Big data logging platform Enhance customer experiences using digital platforms. Spot trends, patterns, or anomalies within billions of lines of log data to make near instant corrections to improve performance. Saa. S applications Build multi-tenant Saa. S applications embedded with interactive analytics. Monitor the performance of the application, improve products, and provide business owners insights to boost business outcomes. © Microsoft Corporation “Azure Data Explorer has improved our analysis capabilities for our product tremendously…The scalability and performance allow us to deeply analyze our collected data and retrieve valuable insights. ” Nik Shampur Software Development Lead

Azure Data Explorer Fast and fully managed data analytics service Fully managed for efficiency Optimized for streaming data Designed for data exploration Focus on insights, not the infrastructure for fast time to value Get near-instant insights from fast-flowing data Run ad-hoc queries using the intuitive query language No infrastructure to manage; provision the service, choose the SKU for your workload, and create database. Scale linearly up to 200 MB per second per node with highly performant, low latency ingestion. Returns results from 1 Billion records < 1 second without modifying the data or metadata © Microsoft Corporation

Azure Data Explorer overview 1. Capability for many data types, formats, and sources Structured (numbers), semi-structured (JSONXML), and free text 2. Batch or streaming ingestion Use managed ingestion pipeline or queue a request for pull ingestion Azure Portal Cloud Servers Apps (Via API) Logstash Plugin Power BI Kafka Sync Stream Io. T Hub 3. 4. Compute and storage isolation • Independent scale out / scale in • Persistent data in Azure Blob Storage • Caching for low-latency on compute Multiple options to support data consumption Use out-of-the box tools and connectors or use APIs/SDKs for custom solution © Microsoft Corporation Active Data Connections Event Hub Io. T Data Lake / Blob Query, Control Commands Create, Manage Event Grid Data Management ADX Web UI Engine Batch Grafana Azure Data Explorer Customer Data Lake Azure Storage Azure ODBC / JDBC Apps OSS Ingested Data Applications

Intuitive querying Simple and powerful • Rich rational query language (filter, aggregate, join, calculated columns, and more) • Built-in full-text search, time series, user analytics, and machine learning operators • Out-of-the box visualization (render) • Easy-to-use syntax + Microsoft Intelli. Sense • Highly recognizable hierarchical schema entities Comprehensive • Built for querying over structured, semi-structured and unstructured data simultaneously Extensible • In-line Python • SQL © Microsoft Corporation Designed for data exploration

Thank you