Watson Machine Learning - Details

Try free tutorial

Build, deploy, test, retrain and monitor a predictive machine learning model.

Watson Machine Learning key features

Open source and IBM tool support

Bring popular tools, libraries and frameworks to deploy including Apache Spark MLlib, scikit-learn, XGBoost, TensorFlow, IBM SPSS® Modeler and IBM Watson® Studio.

Learn more about open source

One-click model deployment

Deploy machine learning algorithms to production at scale across any cloud and on premises with a single click.

Explore one-click deployment

Integrated AutoAI model deployment

Train and save a model, then deploy AutoAI models while viewing deployment details and test results.

Test AutoAI now

Seamless model management

Use model lineage to track versions and events, including model create, model deployed, model re-evaluated, model retrained and model set as an active model.

Train and deploy now

Hands-on machine learning

Explore multiple machine-learning and deep-learning capabilities in this one-hour lab.

Start learning

Easier deep learning experiments

Train numerous models to identify the right combination of data in conjunction with hyperparameters to optimize the performance of your neural networks.

See deep learning in action

Deploy Python functions

Deploy any Python function to a REST endpoint easily. Preprocess data, perform error handling and manage multiple API calls.

Go live with Python

Run decision optimization models

Deploy your decision optimization and associated master data once and then submit job requests via REST API or the Python client.

Build apps with optimization

Hyperparameter optimization

Explore a search space of hyperparameters automaticallly, build a series of models and compare the models using metrics of interest.

Learn and optimize hyperparameters

Leverage REST API and Python client for easier integration

Use the Python library or REST API with IBM Watson Machine Learning to train models, create deployments and score new data.

Push your model with REST API

Flexibly specify a model type and runtimes

Save your choice of a machine learning model that maps to the supported tools, algorithms and frameworks.

Use your preferred machine learning tools

Explainable models enabled by AutoAI

Examine estimators, model types and metrics used by AutoAI, including algorithms trained on your data sets and feature transformations.

Take a closer look at AutoAI

Put the power of AI model deployment in your hands

Create a real-time object detection app

Try the object detection app tutorial

Use Watson™ Machine Learning to train your own custom model to detect objects in real time — without substantial computing power and time.

Build a product recommendation engine

Try the interactive purchase tutorial

Use Jupyter Notebooks with IBM Watson Studio to build an interactive recommendation engine PixieApp and deploy it with Watson Machine Learning.

Monitor model performance

Get the code and view the demo

Deploy and monitor machine learning models using German credit data with Watson Machine Learning and IBM Watson OpenScale™.

Automate model building with AutoAI

Try the AutoAI tutorial

Build models, view a leaderboard, compare pipelines and deploy selected models with Watson Studio and Watson Machine Learning.

Predict a product purchase

Learn how to automate prediction

Predict whether a customer is likely to buy a tent from an outdoor equipment store with AutoAI using Watson Studio and Watson Machine Learning.

Use a Python notebook to deploy decision optimization

See how decision optimization gets deployed (05:04)

Deploy your decision optimization model with Watson Machine Learning using a Jupyter notebook to access machine learning services and monitor jobs.

Watson Machine Learning Server

Take the next step on your AI journey.

Read the white paper (PDF, 588 KB)

Watson Machine Learning use cases

Use case: Get started with AutoAI training

Problem

It can take months for data scientists to reach accurate predictions and monitor models in production.
Data scientists who code and know algorithms are in short supply.
Citizen data scientists and analysts need a rapid on-ramp.

Solution

AutoAI helps data scientists rapidly develop candidate pipelines, select top performing models on the leaderboard and deploy models with Watson Machine Learning Cloud. Model monitoring is easier and faster, and the overall process can be reduced from weeks to hours.

Try the AutoAI tutorial →

Use case: Bring open source data science to production

Problem

Businesses are concerned with lack of governance in open source-based models.
Visibility and understanding of model deployment status is hard to gain.
It is difficult to share outcomes with analysts and subject matter experts.

Solution

With Watson Machine Learning Server, you can deploy a machine learning model to IBM Watson Machine Learning using the Python client. You can scale workloads and deploy assets with a seamless user experience.

Explore deployment using Python →

Use case: Bring your Watson Studio Desktop to deployment

Problem

You need to advance from individual data science projects to team-based projects.
Your desktop does not have enough computing power to execute functions.
Execution and deployment of assets from Watson Studio Desktop is the next logical step.

Solution

Watson Machine Learning Server can be connected to Watson Studio Desktop to provide computational power as a runtime, or as a deployment space with multiple user management.

Connect to the Watson Machine Learning Server →

Use case: Build and scale your AI models at scale

Problem

You are moving from departmental to cross-business unit projects that involve many data scientists.
Progressing from experimentation to production is a high priority.
You need flexibility to build and deploy models across multiple clouds.

Solution

Together with Watson Studio, Watson Machine Learning Local helps you build, train and deploy models at scale in your private cloud or the public cloud of your choice including AWS, Azure, and others.

Explore Watson Machine Learning Local →

Use case: Deploy your decision optimization models

Problem

You are unable to deploy outcomes from a range of scenarios for a business problem.
You are challenged in running optimization models via API for applications.
You want to blend machine learning and decision optimization using the same deployment mechanism.

Solution

With Watson Machine Learning, you can deploy your decision optimization prescriptive model and associated data. This can be achieved using the Watson Machine Learning REST API or by using the Watson Machine Learning Python client.

Learn how to deploy decision optimization →

Use case: Build and scale models on a data and AI platform

Problem

Your data science approach has not caught up with the agile practice of DevOps and app development.
There are multiple tools and approaches in cloud and AI that impede productivity of team members.
Simplifying decisions and predictions at scale is critical.

Solution

IBM Cloud Pak™ for Data, a cloud-native data and AI platform, together with Watson Studio Premium, delivers a seamless experience for running and optimizing machine learning and decision optimization models.

See IBM Watson Studio Premium for IBM Cloud Pak for Data →

Try the free hands-on lab

Explore multiple machine learning and deep learning capabilities of IBM Watson Studio.

Try the lab

Try or buy now

Whether you want to try Watson Machine Learning with a free 30-day trial before you buy, or you're ready to explore pricing and deployment plans — IBM makes it simple for you to take the next step.

Free 30-day trial

ㅤㅤ ㅤㅤ ㅤㅤ

View pricing and buy