NVIDIA Wins NeurIPS Awards for Research on Generative AI, Generalist AI Agents

NVIDIA Wins NeurIPS Awards for Research on Generative AI, Generalist AI Agents

Two NVIDIA Research papers — one exploring diffusion-based generative AI models and another on training generalist AI agents — have been honored with NeurIPS 2022 Awards for their contributions to the field of AI and machine learning.

These are among more than 60+ talks, posters and workshops with NVIDIA authors being presented at the NeurIPs conference, taking place this week in New Orleans and next week online.

Synthetic data generation — for images, text or video — is a key theme across several of the NVIDIA-authored papers. Other topics include reinforcement learning, data collection and augmentation, weather models and federated learning.

“AI is an incredibly important technology, and NVIDIA is making fast progress across the gamut — from generative AI to autonomous AI agents,” said Jan Kautz, vice president of learning and perception research at NVIDIA. “In generative AI, we are not only advancing our theoretical understanding of the underlying models, but are also making practical contributions that will reduce the effort of creating realistic virtual worlds and simulations.”

Reimagining the Design of Diffusion-Based Generative Models 

Diffusion-based models have emerged as a groundbreaking technique for generative AI. NVIDIA researchers won an Outstanding Main Track Paper award for work that analyzes the design of diffusion models, proposing improvements that can dramatically improve the efficiency and quality of these models.

The paper breaks down the components of a diffusion model into a modular design, helping developers identify processes that can be adjusted to improve the performance of the entire model. The researchers show that their modifications enable record scores on a metric that assesses the quality of AI-generated images.

Training Generalist AI Agents in a Minecraft-Based Simulation Suite

While researchers have long trained autonomous AI agents in video-game environments such as Starcraft, Dota and Go, these agents are usually specialists in only a few tasks. So NVIDIA researchers turned to Minecraft, the world’s most popular game, to develop a scalable training framework for a generalist agent — one that can successfully execute a wide variety of open-ended tasks.

Dubbed MineDojo, the framework enables an AI agent to learn Minecraft’s flexible gameplay using a massive online database of more than 7,000 wiki pages, millions of Reddit threads and 300,000 hours of recorded gameplay (shown in image at top). The project won an Outstanding Datasets and Benchmarks Paper Award from the NeurIPS committee.

As a proof of concept, the researchers behind MineDojo created a large-scale foundation model, called MineCLIP, that learned to associate YouTube footage of Minecraft gameplay with the video’s transcript, in which the player typically narrates the onscreen action. Using MineCLIP, the team was able to train a reinforcement learning agent capable of performing several tasks in Minecraft without human intervention.

Creating Complex 3D Shapes to Populate Virtual Worlds

Also at NeurIPS is GET3D, a generative AI model that instantly synthesizes 3D shapes based on the category of 2D images it’s trained on, such as buildings, cars or animals. The AI-generated objects have high-fidelity textures and complex geometric details — and are created in a triangle mesh format used in popular graphics software applications. This makes it easy for users to import the shapes into 3D renderers and game engines for further editing.

3D objects generated by GET3D

Named for its ability to Generate Explicit Textured 3D meshes, GET3D was trained on NVIDIA A100 Tensor Core GPUs using around 1 million 2D images of 3D shapes captured from different camera angles. The model can generate around 20 objects a second when running inference on a single NVIDIA GPU.

The AI-generated objects could be used to populate 3D representations of buildings, outdoor spaces or entire cities — digital spaces designed for industries such as gaming, robotics, architecture and social media.

Improving Inverse Rendering Pipelines With Control Over Materials, Lighting

At the most recent CVPR conference, held in New Orleans in June, NVIDIA Research introduced 3D MoMa, an inverse rendering method that enables developers to create 3D objects composed of three distinct parts: a 3D mesh model, materials overlaid on the model, and lighting.

The team has since achieved significant advancements in untangling materials and lighting from the 3D objects — which in turn improves creators’ abilities to edit the AI-generated shapes by swapping materials or adjusting lighting as the object moves around a scene.

The work, which relies on a more realistic shading model that leverages NVIDIA RTX GPU-accelerated ray tracing, is being presented as a poster at NeurIPS.

Enhancing Factual Accuracy of Language Models’ Generated Text 

Another accepted paper at NeurIPS examines a key challenge with pretrained language models: the factual accuracy of AI-generated text.

Language models trained for open-ended text generation often come up with text that includes nonfactual information, since the AI is simply making correlations between words to predict what comes next in a sentence. In the paper, NVIDIA researchers propose techniques to address this limitation, which is necessary before such models can be deployed for real-world applications.

The researchers built the first automatic benchmark to measure the factual accuracy of language models for open-ended text generation, and found that bigger language models with billions of parameters were more factual than smaller ones. The team proposed a new technique, factuality-enhanced training, along with a novel sampling algorithm that together help train language models to generate accurate text — and demonstrated a reduction in the rate of factual errors from 33% to around 15%. 

There are more than 300 NVIDIA researchers around the globe, with teams focused on topics including AI, computer graphics, computer vision, self-driving cars and robotics. Learn more about NVIDIA Research and view NVIDIA’s full list of accepted papers at NeurIPS.

The post NVIDIA Wins NeurIPS Awards for Research on Generative AI, Generalist AI Agents appeared first on NVIDIA Blog.

Read More

MAP Once, Run Anywhere: MONAI Introduces Framework for Deploying Medical Imaging AI Apps

MAP Once, Run Anywhere: MONAI Introduces Framework for Deploying Medical Imaging AI Apps

Delivering AI-accelerated healthcare at scale will take thousands of neural networks working together to cover the breadth of human physiology, diseases and even hospital operations — a significant challenge in today’s smart hospital environment.

MONAI, an open-source medical-imaging AI framework with more than 650,000 downloads, accelerated by NVIDIA, is making it easier to integrate these models into clinical workflows with MONAI Application Packages, or MAPs.

Delivered through MONAI Deploy, a MAP is a way of packaging an AI model that makes it easy to deploy in an existing healthcare ecosystem.

“If someone wanted to deploy several AI models in an imaging department to help experts identify a dozen different conditions, or partially automate the creation of medical imaging reports, it would take an untenable amount of time and resources to get the right hardware and software infrastructure for each one,” said Dr. Ryan Moore at Cincinnati Children’s Hospital. “It used to be possible, but not feasible.”

MAPs simplify the process. When a developer packages an app using the MONAI Deploy Application software development kit, hospitals can easily run it on premises or in the cloud. The MAPs specification also integrates with healthcare IT standards such as DICOM for medical imaging interoperability.

“Until now, most AI models would remain in an R&D loop, rarely reaching patient care,” said Jorge Cardoso, chief technology officer at the London Medical Imaging & AI Centre for Value-Based Healthcare. “MONAI Deploy will help break that loop, making impactful clinical AI a more frequent reality.”

MONAI Deploy Adopted by Hospitals, Healthcare Startups

Healthcare institutions, academic medical centers and AI software developers around the world worldwide are adopting MONAI Deploy, including:

  • Cincinnati Children’s Hospital: The academic medical center is creating a MAP for an AI model that automates total cardiac volume segmentation from CT images, aiding pediatric heart transplant patients in a project funded by the National Institutes of Health.
  • National Health Service in England: The NHS Trusts have deployed its MONAI-based AI Deployment Engine platform, known as AIDE, across four hospitals to provide AI-enabled disease-detection tools to healthcare professionals serving 5 million patients a year.
  • Qure.ai: A member of the NVIDIA Inception program for startups, Qure.ai develops medical imaging AI models for use cases including lung cancer, traumatic brain injuries and tuberculosis. The company is using MAPs to package its solutions for deployment, accelerating its time to clinical impact.
  • SimBioSys: The Chicago-based Inception startup builds 3D virtual representations of patients’ tumors and is using MAPs for precision medicine AI applications that can help predict how a patient will respond to a specific treatment.
  • University of California, San Francisco: UCSF is developing MAPs for several AI models, with applications including hip fracture detection, liver and brain tumor segmentation, and knee and breast cancer classification.

Putting Medical Imaging AI on the MAP

The MAP specification was developed by the MONAI Deploy working group, a team of experts from more than a dozen medical imaging institutions, to benefit AI app developers as well as the clinical and infrastructure platforms that run AI apps.

For developers, MAPs can help accelerate AI model evolution by helping researchers easily package and test their models in a clinical environment. This allows them to collect real-world feedback that helps improve the AI.

For cloud service providers, supporting MAPs — which were designed using cloud-native technologies — enables researchers and companies using MONAI Deploy to run AI applications on their platform, either by using containers or with native app integration. Cloud platforms integrating MONAI Deploy and MAPs include:

  • Amazon HealthLake Imaging: The MAP connector has been integrated with the HealthLake Imaging service, allowing clinicians to view, process and segment medical images in real time.
  • Google Cloud: Google Cloud’s Medical Imaging Suite, designed to make healthcare imaging data more accessible, interoperable and useful, has integrated MONAI into its platform to enable clinicians to deploy AI-assisted annotation tools that help automate the highly manual and repetitive task of labeling medical images.
  • Nuance Precision Imaging Network, powered by Microsoft Azure: Nuance and NVIDIA recently announced a partnership bringing together MONAI and the Nuance Precision Imaging Network, a cloud platform that provides more than 12,000 healthcare facilities with access to AI-powered tools and insights.
  • Oracle Cloud Infrastructure: Oracle and NVIDIA recently announced a collaboration to bring accelerated compute solutions for healthcare, including MONAI Deploy, to Oracle Cloud Infrastructure. Developers can start building MAPs with MONAI Deploy today using NVIDIA containers on the Oracle Cloud Marketplace.

Get started with MONAI and discover how NVIDIA is helping build AI-powered medical imaging ecosystems at this week’s RSNA conference.

The post MAP Once, Run Anywhere: MONAI Introduces Framework for Deploying Medical Imaging AI Apps appeared first on NVIDIA Blog.

Read More

NVIDIA Partners With NHS Trusts to Deploy AI Platform in UK Hospitals

NVIDIA Partners With NHS Trusts to Deploy AI Platform in UK Hospitals

A consortium of 10 National Health Service Trusts — the publicly funded healthcare system in England — is now deploying the MONAI-based AIDE platform across four of its hospitals, providing AI-enabled disease-detection tools to healthcare professionals serving 5 million patients a year.

AIDE, short for AI Deployment Engine, is expected to be rolled out next year across 11 NHS hospitals serving 18 million patients, bringing AI capabilities to clinicians. It’s built on MONAI, an open-source medical imaging AI framework co-developed by NVIDIA and the AI Centre, which allows AI applications to interface with hospital systems.

Together, MONAI and AIDE enable safe and effective validation, deployment and evaluation of medical imaging AI models, which the NHS will apply in diagnosing and treating cancers, stroke, dementia and other conditions. The platform is being deployed at the following facilities: Guy’s and St Thomas’s, King’s College Hospital, East Kent Hospital University and University College London Hospitals NHS Foundation Trusts.

“Deployment of this infrastructure for clinical AI tools is a hugely exciting step in integrating AI into healthcare services,” said James Teo, professor of neurology and data science at King’s College Hospital NHS. “These platforms will provide a scalable way for clinicians to deploy healthcare AI tools to support decision-making to improve the speed and precision of patient care. This is the start of a digital transformation journey with strong, safe and open foundations.”

MONAI Making Hospital Integration Easier

Introduced in 2019, MONAI is reducing the complexity of medical workflows from R&D to the clinic. It allows developers to easily build and deploy AI applications, resulting in a model ready for clinical integration, and making it easier to interpret medical exams and unlock new levels of knowledge about patients.

MONAI provides deep learning infrastructure and workflows optimized for medical imaging. MONAI, with more than 650,000 downloads, is used by leading healthcare institutions Guy’s and St Thomas’ Hospital and King’s College Hospital in the U.K., for its ability to harness the power and potential of medical imaging data to simplify and streamline the process for building AI models.

“Across the healthcare ecosystem, researchers, hospitals and startups are realizing the power of incorporating a streamlined AI pipeline into their work,” said Haris Shuaib, AI transformation lead at the AI Centre. “The open-source MONAI ecosystem is standardizing hundreds of AI algorithms for maximum interoperability and impact, enabling their deployment in just a few weeks instead of three-to-six months.”

Built in collaboration with the AI Centre for Value Based Healthcare — a consortium of universities, hospitals and industry partners led by King’s College London and Guy’s and St Thomas’ NHS Foundation Trust — AIDE brings the capabilities of AI to clinicians. This solution equips clinicians with improved information about patients, making healthcare data more accessible and interoperable, in order to improve patient care.

The AI Centre has already developed algorithms to improve diagnosis of COVID-19, breast cancer, brain tumor, stroke detection and dementia risk. AIDE connects approved AI algorithms to a patient’s medical record seamlessly and securely, with the data never leaving the hospital trust.

Once the clinical data has been analyzed, the results are sent back to the electronic healthcare record to support clinical decision-making. This provides another valuable data point for clinical multidisciplinary teams when reviewing patients’ cases. It’s hoped that AIDE can support speeding up this process to benefit patients.

“The AI Centre has done invaluable work towards integrating AI into national healthcare. Deploying MONAI is a critical milestone in our journey to enable the use of safe and robot AI innovations within the clinic,” said Professor Sebastien Ourselin, deputy director of the AI Centre. “This could only be achieved through our strong partnerships between academic and industry leaders like NVIDIA.”

The code for AIDE will be made open source and published on GitHub on Dec. 7. AIDE will be displayed in the South Hall of the McCormick Place convention center in Chicago as part of the RSNA Imaging AI in Practice demonstration.

Get started with MONAI and watch the NVIDIA RSNA special address.

The post NVIDIA Partners With NHS Trusts to Deploy AI Platform in UK Hospitals appeared first on NVIDIA Blog.

Read More

Introducing Amazon Kendra tabular search for HTML Documents

Introducing Amazon Kendra tabular search for HTML Documents

Amazon Kendra is an intelligent search service powered by machine learning (ML). Kendra reimagines enterprise search for your websites and applications so your employees and customers can easily find the content they’re looking for, even when it’s scattered across multiple locations and content repositories within your organization.

Amazon Kendra users can now quickly find the information they need from tables on a webpage (HTML tables) using Amazon Kendra tabular search. Tables contains useful information in structured format so it can be easily interpreted by making visual associations between row and column headers. With Amazon Kendra tabular search, you can now get specific information from the cell or certain rows and columns relevant to your query, as well as preview of the table.

In this post, we provide an example of how to use Amazon Kendra tabular search.

Tabular search in Amazon Kendra

Let’s say you have a webpage in HTML format that contains a table with inflation rates and annual changes in the US from 2012–2021, as shown in the following screenshot.

When you search for “Inflation rate in US”, Amazon Kendra presents the top three rows in the preview and up to five columns, as shown in the following screenshot. You can then see if this article has the relevant details that you’re looking for and decide to either use this information or open the link to get additional details. Amazon Kendra tabular search can also handle merged rows.

Let’s do another search and get specific information from the table by asking “What was the annual change of inflation rate in 2017?”. As shown in the following screenshot, Amazon Kendra tabular search highlights the specific cell that contains the answer to your question.

Now let’s search for “Which year had top inflation rate?”, Amazon Kendra searches the table, sorts the results, and gives you the year that had the highest inflation rate.

Amazon Kendra can also find the range of column information that you’re looking for. For example, let’s search for “Inflation rate from 2012 and 2014.” Amazon Kendra displays the rows and columns between 2012–2014 in the preview.

Get started with Amazon Kendra tabular search

Amazon Kendra tabular search is turned on by default and no special configuration is required to enable it. For newer documents, Amazon Kendra tabular search will work by default. For existing HTML pages that contain tables, you can either update the document and sync (if you only have a few documents), or reach out to AWS Support .

To test tabular search on your internal or external webpage, complete the following steps:

  1. Create an index.
  2. Add data sources by using the web crawler or downloading the HTML page and uploading it to an Amazon Simple Storage Service (Amazon S3) bucket.
  3. Go to the Search Indexed Content tab and test it out.

Limitations and considerations

Keep the following in mind when using this feature:

  • In this release, Amazon Kendra only supports HTML formatted tables or HTML tables within the table tag. This doesn’t include nested tables or other forms of tables.
  • Amazon Kendra can search through tables up to 30 columns and 60 rows, and up to 500 total table cells. If you have a table with a higher numbers of rows, columns, or table cells, Amazon Kendra will not search within that table.
  • Amazon Kendra doesn’t display tabular search results if the confidence score of query result for the column and row is very low. You can look at the confidence score within ScoreAttributes using the QueryResultItem API.

Conclusion

With Amazon Kendra’s tabular search for HTML in Amazon Kendra, you can now search across both unstructured data from various data sources and structured data in the form of tables. This further enhances the user experiences and you can get factual responses from your natural language query as well as from the tables. The table preview with Kendra’s suggested answers allows you to quickly asses if the HTML document table contains relevant information you are looking for, thereby saving time.

Amazon Kendra tabular search is available in the following AWS regions during launch: US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Ireland), Asia Pacific (Sydney), Asia Pacific (Singapore), Canada (Central) and AWS GovCloud (US-West).

To learn more about Amazon Kendra, visit the Amazon Kendra product page.


About the authors

Vikas Shah is an Enterprise Solutions Architect at Amazon web services. He is a technology enthusiast who enjoys helping customers find innovative solutions to complex business challenges. His areas of interest are ML, IoT, robotics and storage. In his spare time, Vikas enjoys building robots, hiking, and traveling.

Read More

Enterprise administrative controls, simple sign-up, and expanded programming language support for Amazon CodeWhisperer

Enterprise administrative controls, simple sign-up, and expanded programming language support for Amazon CodeWhisperer

Amazon CodeWhisperer is a machine learning (ML)-powered service that helps improve developer productivity by generating code recommendations based on developers’ prior code and comments. Today, we are excited to announce that AWS administrators can now enable CodeWhisperer for their organization with single sign-n (SSO) authentication. Administrators can easily integrate CodeWhisperer with their existing workforce identity solutions, provide access to users and groups, and configure organization-wide settings. Additionally, individual users who don’t have AWS accounts can now use CodeWhisperer using their personal email with AWS Builder ID. The sign-up process takes only a few minutes and enables developers to start using CodeWhisperer immediately without any waitlist. We’re also expanding programming language support for CodeWhisperer. In addition to Python, Java, and JavaScript, developers can now use CodeWhisperer to accelerate development of their C# and TypeScript projects.

In this post, we discuss enterprise administrative controls, the new AWS Builder ID sign-up for CodeWhisperer, and support for new programming languages.

Enable CodeWhisperer for your organization

CodeWhisperer is now available on the AWS Management Console. Any user with an AWS administrator role can enable CodeWhisperer, add and remove users, and centrally manage settings for your organization via the console.

As a prerequisite, your AWS administrators have to set up SSO via AWS IAM Identity Center (successor to AWS Single Sign-On), if not already enabled for your organization. IAM Identity Center enables you to use your organization’s SSO to access AWS services by integrating your existing workforce identity solution with AWS. After SSO authentication is set up, your administrators can enable CodeWhisperer and assign access to users and groups, as shown in the following screenshot.

Set up CodeWhisperer

In addition to managing users, AWS administrators can also configure settings for the reference tracker and data sharing. The CodeWhisperer reference tracker detects whether a code recommendation might be similar to particular CodeWhisperer training data and can provide those references to you. CodeWhisperer learns, in part, from open-source projects. Sometimes, a suggestion it’s giving you may be similar to a specific piece of training data. The reference tracker setting enables administrators to decide whether CodeWhisperer is allowed to offer suggestions in such cases. When allowed, CodeWhisperer will also provide references, so that you can learn more about where the training data comes from. AWS administrators can also opt out of data sharing for the purpose of CodeWhisperer service improvement on behalf of your organization (see AI services opt-out policies). Once configured by the administrator, the settings are applied across your organization.

Developers who were given access can start using CodeWhisperer in their preferred IDE by simply logging in using their SSO login credentials. CodeWhisperer is available as part of the AWS Toolkit extensions for major IDEs, including JetBrains, Visual Studio Code, and AWS Cloud9.

In your preferred IDE, choose the SSO login option and follow the prompts to get authenticated and start getting recommendations from CodeWhisperer, as shown in the following screenshots.

connect using AWS IAM

confirm your input

Sign up within minutes using your personal email

If you’re an individual developer who doesn’t have access to an AWS account, you can use your personal email to sign up and enable CodeWhisperer in your preferred IDE. The sign-up process takes only a few minutes.

We’re introducing a new method of authentication with AWS Builder ID. AWS Builder ID is a new form of authentication that allows you to sign up securely with just your personal email and a password. After you create an AWS Builder account, simply log in and enable CodeWhisperer for your IDE, as shown in the following screenshot. For more information, see AWS Builder ID docs.

sign up using personal email

Build apps faster with TypeScript and C# programming languages

Keeping up with multiple programming languages, frameworks, and software libraries is an arduous task even for the most experienced developers. Looking up correct programming syntax and searching code snippets from web to programming tasks takes a significant amount of time, especially if you consider the cost of distractions.

CodeWhisperer provides ready-to-use real-time recommendations in your IDE to help you finish your coding tasks faster. Today, we’re expanding our support to include TypeScript and C# programming languages, in addition to Python, Java, and JavaScript.

CodeWhisperer understands your intent and provides recommendations based on the most commonly used best practices for a programming language. The following example shows how CodeWhisperer can generate the entire function in TypeScript to render JSON to a table.

TypeScript to render JSON to a table

CodeWhisperer also makes it easy for developers to use AWS services by providing code recommendations for AWS application programming interfaces (APIs) across the most popular services, including Amazon Elastic Compute Cloud (Amazon EC2), AWS Lambda, and Amazon Simple Storage Service (Amazon S3). We also offer a reference tracker with our recommendations that provides valuable information about the similarity of the recommendation to particular CodeWhisperer training data. Furthermore, we have implemented techniques to detect and filter biased code that might be unfair. The following example shows how CodeWhisperer can generate an entire function based on prompts provided in C#.

CodeWhisperer generates entire function based on prompts provided in C#

Get started with CodeWhisperer

During the preview period, CodeWhisperer is available to all developers across the world for free. To access the service in preview, you can enable it for your organization using the console, or you can use the AWS Builder ID to get started as an individual developer. For more information about the service, visit Amazon CodeWhisperer.


About the Authors

Bharadwaj Tanikella is a Senior Product Manager for Amazon CodeWhisperer. He has a background in Machine Learning, both as a developer and a Product Manager. In his spare time he loves to bike, read non-fiction and learning new languages.

Ankur Desai is a Principal Product Manager within the AWS AI Services team.

Read More

Optimize hyperparameters with Amazon SageMaker Automatic Model Tuning

Optimize hyperparameters with Amazon SageMaker Automatic Model Tuning

Machine learning (ML) models are taking the world by storm. Their performance relies on using the right training data and choosing the right model and algorithm. But it doesn’t end here. Typically, algorithms defer some design decisions to the ML practitioner to adopt for their specific data and task. These deferred design decisions manifest themselves as hyperparameters.

What does that name mean? The result of ML training, the model, can be largely seen as a collection of parameters that are learned during training. Therefore, the parameters that are used to configure the ML training process are then called hyperparameters—parameters describing the creation of parameters. At any rate, they are of very practical use, such as the number of epochs to train, the learning rate, the max depth of a decision tree, and so forth. And we pay much attention to them because they have a major impact on the ultimate performance of your model.

Just like turning a knob on a radio receiver to find the right frequency, each hyperparameter should be carefully tuned to optimize performance. Searching the hyperparameter space for the optimal values is referred to as hyperparameter tuning or hyperparameter optimization (HPO), and should result in a model that gives accurate predictions.

In this post, we set up and run our first HPO job using Amazon SageMaker Automatic Model Tuning (AMT). We learn about the methods available to explore the results, and create some insightful visualizations of our HPO trials and the exploration of the hyperparameter space!

Amazon SageMaker Automatic Model Tuning

As an ML practitioner using SageMaker AMT, you can focus on the following:

  • Providing a training job
  • Defining the right objective metric matching your task
  • Scoping the hyperparameter search space

SageMaker AMT takes care of the rest, and you don’t need to think about the infrastructure, orchestrating training jobs, and improving hyperparameter selection.

Let’s start by using SageMaker AMT for our first simple HPO job, to train and tune an XGBoost algorithm. We want your AMT journey to be hands-on and practical, so we have shared the example in the following GitHub repository. This post covers the 1_tuning_of_builtin_xgboost.ipynb notebook.

In an upcoming post, we’ll extend the notion of just finding the best hyperparameters and include learning about the search space and to what hyperparameter ranges a model is sensitive. We’ll also show how to turn a one-shot tuning activity into a multi-step conversation with the ML practitioner, to learn together. Stay tuned (pun intended)!

Prerequisites

This post is for anyone interested in learning about HPO and doesn’t require prior knowledge of the topic. Basic familiarity with ML concepts and Python programming is helpful though. For the best learning experience, we highly recommend following along by running each step in the notebook in parallel to reading this post. And at the end of the notebook, you also get to try out an interactive visualization that makes the tuning results come alive.

Solution overview

We’re going to build an end-to-end setup to run our first HPO job using SageMaker AMT. When our tuning job is complete, we look at some of the methods available to explore the results, both via the AWS Management Console and programmatically via the AWS SDKs and APIs.

First, we familiarize ourselves with the environment and SageMaker Training by running a standalone training job, without any tuning for now. We use the XGBoost algorithm, one of many algorithms provided as a SageMaker built-in algorithm (no training script required!).

We see how SageMaker Training operates in the following ways:

  • Starts and stops an instance
  • Provisions the necessary container
  • Copies the training and validation data onto the instance
  • Runs the training
  • Collects metrics and logs
  • Collects and stores the trained model

Then we move to AMT and run an HPO job:

  • We set up and launch our tuning job with AMT
  • We dive into the methods available to extract detailed performance metrics and metadata for each training job, which enables us to learn more about the optimal values in our hyperparameter space
  • We show you how to view the results of the trials
  • We provide you with tools to visualize data in a series of charts that reveal valuable insights into our hyperparameter space

Train a SageMaker built-in XGBoost algorithm

It all starts with training a model. In doing so, we get a sense of how SageMaker Training works.

We want to take advantage of the speed and ease of use offered by the SageMaker built-in algorithms. All we need are a few steps to get started with training:

  1. Prepare and load the data – We download and prepare our dataset as input for XGBoost and upload it to our Amazon Simple Storage Service (Amazon S3) bucket.
  2. Select our built-in algorithm’s image URI – SageMaker uses this URI to fetch our training container, which in our case contains a ready-to-go XGBoost training script. Several algorithm versions are supported.
  3. Define the hyperparameters – SageMaker provides an interface to define the hyperparameters for our built-in algorithm. These are the same hyperparameters as used by the open-source version.
  4. Construct the estimator – We define the training parameters such as instance type and number of instances.
  5. Call the fit() function – We start our training job.

The following diagram shows how these steps work together.

SageMaker training overview

Provide the data

To run ML training, we need to provide data. We provide our training and validation data to SageMaker via Amazon S3.

In our example, for simplicity, we use the SageMaker default bucket to store our data. But feel free to customize the following values to your preference:

sm_sess = sagemaker.session.Session([..])

BUCKET = sm_sess.default_bucket()
PREFIX = 'amt-visualize-demo'
output_path = f's3://{BUCKET}/{PREFIX}/output'

In the notebook, we use a public dataset and store the data locally in the data directory. We then upload our training and validation data to Amazon S3. Later, we also define pointers to these locations to pass them to SageMaker Training.

# acquire and prepare the data (not shown here)
# store the data locally
[..]
train_data.to_csv('data/train.csv', index=False, header=False)
valid_data.to_csv('data/valid.csv', index=False, header=False)
[..]
# upload the local files to S3
boto_sess.resource('s3').Bucket(BUCKET).Object(os.path.join(PREFIX, 'data/train/train.csv')).upload_file('data/train.csv')
boto_sess.resource('s3').Bucket(BUCKET).Object(os.path.join(PREFIX, 'data/valid/valid.csv')).upload_file('data/valid.csv')

In this post, we concentrate on introducing HPO. For illustration, we use a specific dataset and task, so that we can obtain measurements of objective metrics that we then use to optimize the selection of hyperparameters. However, for the overall post neither the data nor the task matter. To present you with a complete picture, let us briefly describe what we do: we train an XGBoost model that should classify handwritten digits from the
Optical Recognition of Handwritten Digits Data Set [1] via Scikit-learn. XGBoost is an excellent algorithm for structured data and can even be applied to the Digits dataset. The values are 8×8 images, as in the following example showing a
0 a
5 and a
4.

Select the XGBoost image URI

After choosing our built-in algorithm (XGBoost), we must retrieve the image URI and pass this to SageMaker to load onto our training instance. For this step, we review the available versions. Here we’ve decided to use version 1.5.1, which offers the latest version of the algorithm. Depending on the task, ML practitioners may write their own training script that, for example, includes data preparation steps. But this isn’t necessary in our case.

If you want to write your own training script, then stay tuned, we’ve got you covered in our next post! We’ll show you how to run SageMaker Training jobs with your own custom training scripts.

For now, we need the correct image URI by specifying the algorithm, AWS Region, and version number:

xgboost_container = sagemaker.image_uris.retrieve('xgboost', region, '1.5-1')

That’s it. Now we have a reference to the XGBoost algorithm.

Define the hyperparameters

Now we define our hyperparameters. These values configure how our model will be trained, and eventually influence how the model performs against the objective metric we’re measuring against, such as accuracy in our case. Note that nothing about the following block of code is specific to SageMaker. We’re actually using the open-source version of XGBoost, just provided by and optimized for SageMaker.

Although each of these hyperparameters are configurable and adjustable, the objective metric multi:softmax is determined by our dataset and the type of problem we’re solving for. In our case, the Digits dataset contains multiple labels (an observation of a handwritten digit could be 0 or 1,2,3,4,5,6,7,8,9), meaning it is a multi-class classification problem.

hyperparameters = {
    'num_class': 10,
    'max_depth': 5,
    'eta':0.2,
    'alpha': 0.2,
    'objective':'multi:softmax',
    'eval_metric':'accuracy',
    'num_round':200,
    'early_stopping_rounds': 5
}

For more information about the other hyperparameters, refer to XGBoost Hyperparameters.

Construct the estimator

We configure the training on an estimator object, which is a high-level interface for SageMaker Training.

Next, we define the number of instances to train on, the instance type (CPU-based or GPU-based), and the size of the attached storage:

estimator = sagemaker.estimator.Estimator(
    image_uri=xgboost_container, 
    hyperparameters=hyperparameters,
    role=role,
    instance_count=1, 
    instance_type='ml.m5.large', 
    volume_size=5, # 5 GB 
    output_path=output_path
)

We now have the infrastructure configuration that we need to get started. SageMaker Training will take care of the rest.

Call the fit() function

Remember the data we uploaded to Amazon S3 earlier? Now we create references to it:

s3_input_train = TrainingInput(s3_data=f's3://{BUCKET}/{PREFIX}/data/train', content_type='csv')
s3_input_valid = TrainingInput(s3_data=f's3://{BUCKET}/{PREFIX}/data/valid', content_type='csv')

A call to fit() launches our training. We pass in the references to the training data we just created to point SageMaker Training to our training and validation data:

estimator.fit({'train': s3_input_train, 'validation': s3_input_valid})

Note that to run HPO later on, we don’t actually need to call fit() here. We just need the estimator object later on for HPO, and could just jump to creating our HPO job. But because we want to learn about SageMaker Training and see how to run a single training job, we call it here and review the output.

After the training starts, we start to see the output below the cells, as shown in the following screenshot. The output is available in Amazon CloudWatch as well as in this notebook.

The black text is log output from SageMaker itself, showing the steps involved in training orchestration, such as starting the instance and loading the training image. The blue text is output directly from the training instance itself. We can observe the process of loading and parsing the training data, and visually see the training progress and the improvement in the objective metric directly from the training script running on the instance.

Output from fit() function in Jupyter Notebook

Also note that at the end of the output job, the training duration in seconds and billable seconds are shown.

Finally, we see that SageMaker uploads our training model to the S3 output path defined on the estimator object. The model is ready to be deployed for inference.

In a future post, we’ll create our own training container and define our training metrics to emit. You’ll see how SageMaker is agnostic of what container you pass it for training. This is very handy for when you want to get started quickly with a built-in algorithm, but then later decide to pass your own custom training script!

Inspect current and previous training jobs

So far, we have worked from our notebook with our code and submitted training jobs to SageMaker. Let’s switch perspectives and leave the notebook for a moment to check out what this looks like on the SageMaker console.

Console view of SageMaker Training jobs

SageMaker keeps a historic record of training jobs it ran, their configurations such as hyperparameters, algorithms, data input, the billable time, and the results. In the list in the preceding screenshot, you see the most recent training jobs filtered for XGBoost. The highlighted training job is the job we just trained in the notebook, whose output you saw earlier. Let’s dive into this individual training job to get more information.

The following screenshot shows the console view of our training job.

Console view of a single SageMaker Training job

We can review the information we received as cell output to our fit() function in the individual training job within the SageMaker console, along with the parameters and metadata we defined in our estimator.

Recall the log output from the training instance we saw earlier. We can access the logs of our training job here too, by scrolling to the Monitor section and choosing View logs.

Console View of monitoring tab in training job

This shows us the instance logs inside CloudWatch.

Console view of training instance logs in CloudWatch

Also remember the hyperparameters we specified in our notebook for the training job. We see them here in the same UI of the training job as well.

Console view of hyperparameters of SageMaker Training job

In fact, the details and metadata we specified earlier for our training job and estimator can be found on this page on the SageMaker console. We have a helpful record of the settings used for the training, such as what training container was used and the locations of the training and validation datasets.

You might be asking at this point, why exactly is this relevant for hyperparameter optimization? It’s because you can search, inspect, and dive deeper into those HPO trials that we’re interested in. Maybe the ones with the best results, or the ones that show interesting behavior. We’ll leave it to you what you define as “interesting.” It gives us a common interface for inspecting our training jobs, and you can use it with SageMaker Search.

Although SageMaker AMT orchestrates the HPO jobs, the HPO trials are all launched as individual SageMaker Training jobs and can be accessed as such.

With training covered, let’s get tuning!

Train and tune a SageMaker built-in XGBoost algorithm

To tune our XGBoost model, we’re going to reuse our existing hyperparameters and define ranges of values we want to explore for them. Think of this as extending the borders of exploration within our hyperparameter search space. Our tuning job will sample from the search space and run training jobs for new combinations of values. The following code shows how to specify the hyperparameter ranges that SageMaker AMT should sample from:

from sagemaker.tuner import IntegerParameter, ContinuousParameter, HyperparameterTuner

hpt_ranges = {
    'alpha': ContinuousParameter(0.01, .5),
    'eta': ContinuousParameter(0.1, .5),
    'min_child_weight': ContinuousParameter(0., 2.),
    'max_depth': IntegerParameter(1, 10)
}

The ranges for an individual hyperparameter are specified by their type, like ContinuousParameter. For more information and tips on choosing these parameter ranges, refer to Tune an XGBoost Model.

We haven’t run any experiments yet, so we don’t know the ranges of good values for our hyperparameters. Therefore, we start with an educated guess, using our knowledge of algorithms and our documentation of the hyperparameters for the built-in algorithms. This defines a starting point to define the search space.

Then we run a tuning job sampling from hyperparameters in the defined ranges. As a result, we can see which hyperparameter ranges yield good results. With this knowledge, we can refine the search space’s boundaries by narrowing or widening which hyperparameter ranges to use. We demonstrate how to learn from the trials in the next and final section, where we investigate and visualize the results.

In our next post, we’ll continue our journey and dive deeper. In addition, we’ll learn that there are several strategies that we can use to explore our search space. We’ll run subsequent HPO jobs to find even more performant values for our hyperparameters, while comparing these different strategies. We’ll also see how to run a warm start with SageMaker AMT to use the knowledge gained from previously explored search spaces in our exploration beyond those initial boundaries.

For this post, we focus on how to analyze and visualize the results of a single HPO job using the Bayesian search strategy, which is likely to be a good starting point.

If you follow along in the linked notebook, note that we pass the same estimator that we used for our single, built-in XGBoost training job. This estimator object acts as a template for new training jobs that AMT creates. AMT will then vary the hyperparameters inside the ranges we defined.

By specifying that we want to maximize our objective metric, validation:accuracy, we’re telling SageMaker AMT to look for these metrics in the training instance logs and pick hyperparameter values that it believes will maximize the accuracy metric on our validation data. We picked an appropriate objective metric for XGBoost from our documentation.

Additionally, we can take advantage of parallelization with max_parallel_jobs. This can be a powerful tool, especially for strategies whose trials are selected independently, without considering (learning from) the outcomes of previous trials. We’ll explore these other strategies and parameters further in our next post. For this post, we use Bayesian, which is an excellent default strategy.

We also define max_jobs to define how many trials to run in total. Feel free to deviate from our example and use a smaller number to save money.

n_jobs = 50
n_parallel_jobs = 3

tuner_parameters = {
    'estimator': estimator, # The same estimator object we defined above
    'base_tuning_job_name': 'bayesian',
    'objective_metric_name': 'validation:accuracy',
    'objective_type': 'Maximize',
    'hyperparameter_ranges': hpt_ranges,
    'strategy': 'Bayesian',
    'max_jobs': n_jobs,
    'max_parallel_jobs': n_parallel_jobs
}

We once again call fit(), the same way as when we launched a single training job earlier in the post. But this time on the tuner object, not the estimator object. This kicks off the tuning job, and in turn AMT starts training jobs.

tuner = HyperparameterTuner(**tuner_parameters)
tuner.fit({'train': s3_input_train, 'validation': s3_input_valid}, wait=False)
tuner_name = tuner.describe()['HyperParameterTuningJobName']
print(f'tuning job submitted: {tuner_name}.')

The following diagram expands on our previous architecture by including HPO with SageMaker AMT.

Overview of SageMaker Training and hyperparameter optimization with SageMaker AMT

We see that our HPO job has been submitted. Depending on the number of trials, defined by n_jobs and the level of parallelization, this may take some time. For our example, it may take up to 30 minutes for 50 trials with only a parallelization level of 3.

tuning job submitted: bayesian-221102-2053.

When this tuning job is finished, let’s explore the information available to us on the SageMaker console.

Investigate AMT jobs on the console

Let’s find our tuning job on the SageMaker console by choosing Training in the navigation pane and then Hyperparameter tuning jobs. This gives us a list of our AMT jobs, as shown in the following screenshot. Here we locate our bayesian-221102-2053 tuning job and find that it’s now complete.

Console view of the Hyperparameter tuning jobs page. Image shows the list view of tuning jobs, containing our 1 tuning entry

Let’s have a closer look at the results of this HPO job.

We have explored extracting the results programmatically in the notebook. First via the SageMaker Python SDK, which is a higher level open-source Python library, providing a dedicated API to SageMaker. Then through Boto3, which provides us with lower-level APIs to SageMaker and other AWS services.

Using the SageMaker Python SDK, we can obtain the results of our HPO job:

sagemaker.HyperparameterTuningJobAnalytics(tuner_name).dataframe()[:10]

This allowed us to analyze the results of each of our trials in a Pandas DataFrame, as seen in the following screenshot.

Pandas table in Jupyter Notebook showing results and metadata from the trails ran for our HPO job

Now let’s switch perspectives again and see what the results look like on the SageMaker console. Then we’ll look at our custom visualizations.

On the same page, choosing our bayesian-221102-2053 tuning job provides us with a list of trials that were run for our tuning job. Each HPO trial here is a SageMaker Training job. Recall earlier when we trained our single XGBoost model and investigated the training job in the SageMaker console. We can do the same thing for our trials here.

As we investigate our trials, we see that bayesian-221102-2053-048-b59ec7b4 created the best performing model, with a validation accuracy of approximately 89.815%. Let’s explore what hyperparameters led to this performance by choosing the Best training job tab.

Console view of a single tuning job, showing a list of training jobs ran

We can see a detailed view of the best hyperparameters evaluated.

Console view of a single tuning job, showing the details of the best training job

We can immediately see what hyperparameter values led to this superior performance. However, we want to know more. Can you guess what? We see that alpha takes on an approximate value of 0.052456 and, likewise, eta is set to 0.433495. This tells us that these values worked well, but it tells us little about the hyperparameter space itself. For example, we might wonder whether 0.433495 for eta was the highest value tested, or whether there’s room for growth and model improvement by selecting higher values.

For that, we need to zoom out, and take a much wider view to see how other values for our hyperparameters performed. One way to look at a lot of data at once is to plot our hyperparameter values from our HPO trials on a chart. That way we see how these values performed relatively. In the next section, we pull this data from SageMaker and visualize it.

Visualize our trials

The SageMaker SDK provides us with the data for our exploration, and the notebooks give you a peek into that. But there are many ways to utilize and visualize it. In this post, we share a sample using the Altair statistical visualization library, which we use to produce a more visual overview of our trials. These are found in the amtviz package, which we are providing as part of the sample:

from amtviz import visualize_tuning_job
visualize_tuning_job(tuner, trials_only=True)

The power of these visualizations becomes immediately apparent when plotting our trials’ validation accuracy (y-axis) over time (x-axis). The following chart on the left shows validation accuracy over time. We can clearly see the model performance improving as we run more trials over time. This is a direct and expected outcome of running HPO with a Bayesian strategy. In our next post, we see how this compares to other strategies and observe that this doesn’t need to be the case for all strategies.

Two Charts showing HPO trails. Left Chart shows validation accuracy over time. Right chart shows density chart for validation accuracy values

After reviewing the overall progress over time, now let’s look at our hyperparameter space.

The following charts show validation accuracy on the y-axis, with each chart showing max_depth, alpha, eta, and min_child_weight on the x-axis, respectively. We’ve plotted our entire HPO job into each chart. Each point is a single trial, and each chart contains all 50 trials, but separated for each hyperparameter. This means that our best performing trial, #48, is represented by exactly one blue dot in each of these charts (which we have highlighted for you in the following figure). We can visually compare its performance within the context of all other 49 trials. So, let’s look closely.

Fascinating! We see immediately which regions of our defined ranges in our hyperparameter space are most performant! Thinking back to our eta value, it’s clear now that sampling values closer to 0 yielded worse performance, whereas moving closer to our border, 0.5, yields better results. The reverse appears to be true for alpha, and max_depth appears to have a more limited set of preferred values. Looking at max_depth, you can also see how using a Bayesian strategy instructs SageMaker AMT to sample more frequently values it learned worked well in the past.

Four charts showing validation accuracy on the y-axis, with each chart showing max_depth, alpha, eta, min_child_weight on the x-axis respectively. Each data point represents a single HPO trial

Looking at our eta value, we might wonder whether it’s worth exploring more to the right, perhaps beyond 0.45? Does it continue to trail off to lower accuracy, or do we need more data here? This wondering is part of the purpose of running our first HPO job. It provides us with insights into which areas of the hyperparameter space we should explore further.

If you’re keen to know more, and are as excited as we are by this introduction to the topic, then stay tuned for our next post, where we’ll talk more about the different HPO strategies, compare them against each other, and practice training with our own Python script.

Clean up

To avoid incurring unwanted costs when you’re done experimenting with HPO, you must remove all files in your S3 bucket with the prefix amt-visualize-demo and also shut down Studio resources.

Run the following code in your notebook to remove all S3 files from this post.

!aws s3 rm s3://{BUCKET}/amt-visualize-demo --recursive

If you wish to keep the datasets or the model artifacts, you may modify the prefix in the code to amt-visualize-demo/data to only delete the data or amt-visualize-demo/output to only delete the model artifacts.

Conclusion

In this post, we trained and tuned a model using the SageMaker built-in version of the XGBoost algorithm. By using HPO with SageMaker AMT, we learned about the hyperparameters that work well for this particular algorithm and dataset.

We saw several ways to review the outcomes of our hyperparameter tuning job. Starting with extracting the hyperparameters of the best trial, we also learned how to gain a deeper understanding of how our trials had progressed over time and what hyperparameter values are impactful.

Using the SageMaker console, we also saw how to dive deeper into individual training runs and review their logs.

We then zoomed out to view all our trials together, and review their performance in relation to other trials and hyperparameters.

We learned that based on the observations from each trial, we were able to navigate the hyperparameter space to see that tiny changes to our hyperparameter values can have a huge impact on our model performance. With SageMaker AMT, we can run hyperparameter optimization to find good hyperparameter values efficiently and maximize model performance.

In the future, we’ll look into different HPO strategies offered by SageMaker AMT and how to use our own custom training code. Let us know in the comments if you have a question or want to suggest an area that we should cover in upcoming posts.

Until then, we wish you and your models happy learning and tuning!

References

Citations:

[1] Dua, D. and Graff, C. (2019). UCI Machine Learning Repository [http://archive.ics.uci.edu/ml]. Irvine, CA: University of California, School of Information and Computer Science.


About the authors

Andrew Ellul is a Solutions Architect with Amazon Web Services. He works with small and medium-sized businesses in Germany. Outside of work, Andrew enjoys exploring nature on foot or by bike.

Elina Lesyk is a Solutions Architect located in Munich. Her focus is on enterprise customers from the Financial Services Industry. In her free time, Elina likes learning guitar theory in Spanish to cross-learn and going for a run.

Mariano Kamp is a Principal Solutions Architect with Amazon Web Services. He works with financial services customers in Germany on machine learning. In his spare time, Mariano enjoys hiking with his wife.

Read More

DeepMind’s latest research at NeurIPS 2022

NeurIPS is the world’s largest conference in artificial intelligence (AI) and machine learning (ML), and we’re proud to support the event as Diamond sponsors, helping foster the exchange of research advances in the AI and ML community. Teams from across DeepMind are presenting 47 papers, including 35 external collaborations in virtual panels and poster sessions.Read More

DeepMind’s latest research at NeurIPS 2022

DeepMind’s latest research at NeurIPS 2022

NeurIPS is the world’s largest conference in artificial intelligence (AI) and machine learning (ML), and we’re proud to support the event as Diamond sponsors, helping foster the exchange of research advances in the AI and ML community. Teams from across DeepMind are presenting 47 papers, including 35 external collaborations in virtual panels and poster sessions.Read More