Save the date for the AWS Machine Learning Summit: June 2, 2021

On June 2, 2021, don’t miss the opportunity to hear from some of the brightest minds in machine learning (ML) at the free virtual AWS Machine Learning Summit. Machine learning is one of the most disruptive technologies we will encounter in our generation. It’s improving customer experience, creating more efficiencies in operations, and spurring new innovations and discoveries like helping researchers discover new vaccines and aiding autonomous drones to sail our world’s oceans. But we’re just scratching the surface about what is possible. This Summit, which is open to all and free to attend, brings together industry luminaries, AWS customers, and leading ML experts to share the latest in machine learning. You’ll learn about the latest science breakthroughs in ML, how ML is impacting business, best practices in building ML, and how to get started now without prior ML expertise.

Hear from ML leaders from across AWS, Amazon, and the industry, including Swami Sivasubramanian, VP of AI and Machine Learning, AWS; Bratin Saha, VP of Machine Learning, AWS; and Yoelle Maarek, VP of Research, Alexa Shopping, who will share a keynote on how we’re applying customer-obsessed science to advance ML. Andrew Ng, founder and CEO of Landing AI and founder of deeplearning.ai, will join Swami Sivasubramanian in a fireside chat about the future of ML, the skills that are fundamental for the next generation of ML practitioners, and how we can bridge the gap from proof of concept to production in ML. You’ll also get an inside look at trends in deep learning and natural language in a powerhouse fireside chat with Amazon distinguished scientists Alex Smola and Bernhard Schölkopf, and Alexa AI senior principal scientist Dilek Hakkani-Tur.

Pick from over 30 session across four tracks, which all offer something for anyone who is interested in ML. Advanced practitioners and data scientists can learn about scientific breakthroughs and dive deep into the tools for building ML. Business and technical leaders can learn from their peers about implementing organization-wide ML initiatives. And developers can learn how to perform ML without needing any experience.

The science of machine learning

Advanced practitioners will get a technical deep dive into the groundbreaking work that ML scientists within AWS, Amazon, and beyond are doing to advance the science of ML in areas including computer vision, natural language processing, bias, and more. Speakers include two Amazon Scholars, Michael Kearns and Kathleen McKeown. Kearns a professor in the Computer and Information Science department at the University of Pennsylvania, where he holds the National Center Chair. He is co-author of the book “The Ethical Algorithm: The Science of Socially Aware Algorithm Design,” and joined Amazon as a scholar June 2020. McKeown is the Henry and Gertrude Rothschild professor of computer science at Columbia University, and the founding director of the school’s Data Science Institute. She joined Amazon as a scholar in 2019.

The impact of machine learning

Business leaders will learn from AWS customers that are leading the way in ML adoption. Customers including 3M, AstraZeneca, Vanguard, and Latent Space will share how they’re applying ML to create efficiencies, deliver new revenue streams, and launch entirely new products and business models. You’ll get best practices for scaling ML in an organization and showing impact.

How machine learning is done

Data scientists and ML developers will get practical deep dives into tools that can speed up the entire ML lifecycle, from building to training to deploying ML models. Sessions include how to choose the right algorithms, more accurate and speedy data prep, model explainability, and more.

Machine learning: no expertise required

If you’re a developer who wants to apply ML and AI to a use case but doesn’t have the expertise, this track is for you. Learn how to use AWS AI services and other tools to get started with your ML project right away, for use cases including contact center intelligence, personalization, intelligent document processing, business metrics analysis, computer vision, and more.

For more details, visit the website.

About the Author

Laura Jones is a product marketing lead for AWS AI/ML where she focuses on sharing the stories of AWS’s customers and educating organizations on the impact of machine learning. As a Florida native living and surviving in rainy Seattle, she enjoys coffee, attempting to ski and enjoying the great outdoors.

Read More

NVIDIA’s Shalini De Mello Talks Self-Supervised AI, NeurIPS Successes

Shalini De Mello, a principal research scientist at NVIDIA who’s made her mark inventing computer vision technology that contributes to driver safety, finished 2020 with a bang — presenting two posters at the prestigious NeurIPS conference in December.

A 10-year NVIDIA veteran, De Mello works on self-supervised and few-shot learning, 3D reconstruction, viewpoint estimation and human-computer interaction.

She told NVIDIA AI Podcast host Noah Kravitz about her NeurIPS submissions on reconstructing 3D meshes and self-learning transformations for improving head and gaze redirection — both significant challenges for computer vision.

De Mello’s first poster demonstrates how she and her team successfully manage to recreate 3D models in motion without requiring annotations of 3D mesh, 2D keypoints or camera pose — even on such kinetic figures as animals in the wild.

The second poster takes on the issue of datasets in which large portions are unlabeled — focusing specifically on datasets consisting of images of human faces with many variables, including lighting, reflections and head and gaze orientation. De Mello achieved an architecture that could self-learn these variations and control them.

De Mello intends to continue focusing on creating self-supervising AI systems that require less data to achieve the same quality output, which she envisions ultimately helping to reduce bias in AI algorithms.

Key Points From This Episode:

  • Early in her career at NVIDIA, De Mello noticed that technologies for looking inside the car cabin weren’t as mature as the algorithms for automotive vision outside the car. She focused her research on the former, leading to the creation of NVIDIA’s DRIVE IX product for AI-based automotive interfaces in cars.
  • While science has been a lifelong passion, De Mello discovered an appreciation for art and found the perfect blend of the two in signal and image processing. She could immediately see the effects of AI on visual content.

Tweetables:

“We as humans are able to learn effectively with less data — how can we make learning systems do the same? This is a fundamental question to answer for the viability of AI” [29:29]

“Looking back at my career, the one thing I’ve learned is that it’s really important to follow your passion” [32:37]

You Might Also Like:

Behind the Scenes at NeurIPS with NVIDIA and CalTech’s Anima Anandkumar

Anima Anandkumar, NVIDIA’s director of machine learning research and Bren professor at CalTech’s CMS Department, joins AI Podcast host Noah Kravitz to talk about NeurIPS 2020 and to discuss what she sees as the future of AI.

MIT’s Jonathan Frankle on “The Lottery Hypothesis”

Jonathan Frankle, a Ph.D. student at MIT, discusses a paper he co-authored focusing on “The Lottery Hypothesis,” which promises to help advance our understanding of why neural networks, and deep learning, works so well.

NVIDIA’s Neda Cvijetic Explains the Science Behind Self-Driving Cars

Neda Cvijetic, senior manager of autonomous vehicles at NVIDIA, leads the NVIDIA DRIVE Labs series of videos and blogs that break down the science behind autonomous vehicles. She takes NVIDIA AI Podcast Noah Kravitz behind the wheel of a (metaphorical) self-driving car.

Tune in to the AI Podcast

Get the AI Podcast through iTunes, Google Podcasts, Google Play, Castbox, DoggCatcher, Overcast, PlayerFM, Pocket Casts, Podbay, PodBean, PodCruncher, PodKicker, Soundcloud, Spotify, Stitcher and TuneIn. If your favorite isn’t listed here, drop us a note.

Tune in to the Apple Podcast Tune in to the Google Podcast Tune in to the Spotify Podcast

Make the AI Podcast Better

Have a few minutes to spare? Fill out this listener survey. Your answers will help us make a better podcast.

The post NVIDIA’s Shalini De Mello Talks Self-Supervised AI, NeurIPS Successes appeared first on The Official NVIDIA Blog.

Read More

What Will NVIDIA CEO Jensen Huang Cook Up This Time at NVIDIA GTC?

Don’t blink. Accelerated computing is moving innovation forward faster than ever.

And there’s no way to get smarter, quicker, about how it’s changing your world than to tune in to NVIDIA CEO Jensen Huang’s GTC keynote Monday, April 12, starting at 8:30 a.m. PT.

The keynote, delivered again from the kitchen in Huang’s home, will kick off a conference with more than 1,500 sessions covering just about every innovation — from quantum computing to AI — that benefits from moving faster.

Factories of the Future and More….

In his address, Huang will share the company’s vision for the future of computing from silicon to software to services, and from the edge to the data center to the cloud.

A highlight: Huang will detail NVIDIA’s vision for manufacturing and you’ll get a chance to meet “Dave,” who is exploring the Factory of the Future.

Be on the Hunt for Some Surprises

And, to have a little quick fun, we’ve added a few surprises – so be on the lookout. Watch the @NVIDIAGTC Twitter handle for clues and more details.

Stick Around

There’s no need to register for GTC to watch the keynote. But if you’re inspired, it’s a great way to explore all the trends Huang will touch on at GTC — and more.

For more than a decade, GTC has been the place to see innovations that have changed the world. More than 100,000 developers, researchers and IT professionals have already registered to join this year’s conference.

Registration is free and open to all.

Where to Watch

Mark the date — April 12 at 8:30 a.m. PT — on your calendar. Here’s where you can watch live:

U.S.:

Latin America:

Asia:

See you there.

 

The post What Will NVIDIA CEO Jensen Huang Cook Up This Time at NVIDIA GTC? appeared first on The Official NVIDIA Blog.

Read More

Use computer vision to detect crop disease through image analysis with Amazon Rekognition Custom Labels

Currently, many diseases affect farming and lead to significant economic losses due to reduction of yield and loss of quality produce. In many cases, the health condition of a crop or a plant is often assessed by the condition of its leaves. For farmers, it is crucial to identify these symptoms early. Early identification is key to controlling diseases before they spread too far. However, manually identifying if a leaf is infected, the type of the infection, and the required disease control solution is a hard problem to solve. Current methods can be error prone and very costly. This is where an automated machine learning (ML) solution for computer vision (CV) can help. Typically, building complex machine learning models require hundreds of thousands of labeled images, along with expertise in data science. In this post, we showcase how you can build an end-to-end disease detection, identification, and resolution recommendation solution using Amazon Rekognition Custom Labels.

Amazon Rekognition is a fully managed service that provides CV capabilities for analyzing images and video at scale, using deep learning technology without requiring ML expertise. Amazon Rekognition Custom Labels, an automated ML feature of Amazon Rekognition, lets you quickly train custom CV models specific to your business needs, simply by bringing labeled images.

Solution overview

We create a custom model to detect the plant leaf disease. To create our custom model, we follow these steps:

  1. Create a project in Amazon Rekognition Custom Labels.
  2. Create a dataset with images containing multiple types of plant leaf diseases.
  3. Train the model and evaluate the performance.
  4. Test the new custom model using the automatically generated API endpoint.

Amazon Rekognition Custom Labels lets you manage the ML model training process on the Amazon Rekognition console, which simplifies the end-to-end model development and inference process.

Creating your project

To create your plant leaf disease detection project, complete the following steps:

  1. On the Amazon Rekognition console, choose Custom Labels.
  2. Choose Get Started.
  3. For Project name, enter plant-leaf-disease-detection.
  4. Choose Create project.

You can also create a project on the Projects page. You can access the Projects page via the navigation pane.

Creating your dataset

To create your leaf disease detection model, you first need to create a dataset to train the model with. For this post, our dataset is composed of three categories of plant leaf disease images: bacterial leaf blight, brown spots, and leaf smut.

The following images show examples of bacterial leaf blight.

The following images show examples of brown spots.

The following images show examples of leaf smut.

We sourced our images from UCI, Citation (Prajapati HB, Shah JP, Dabhi VK. Detection and classification of rice plant diseases. Intelligent Decision Technologies. 2017 Jan 1;11(3):357-73, doi: 10.3233/IDT-170301) (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.)

To create your dataset, complete the following steps:

  1. Create an Amazon Simple Storage Service (Amazon S3) bucket.

For this post, I create an S3 bucket called plan-leaf-disease-data.

  1. Create three folders inside this bucket called Bacterial-Leaf-Blight, Brown-Spot, and Leaf-Smut to store images of each disease category.

  1. Upload each category of image files in their respective bucket.
  2. On the Amazon Rekognition console, under Datasets, choose Create dataset.
  3. Select Import images from Amazon S3 bucket.

  1. For S3 folder location, enter the S3 bucket path.
  2. For automatic labeling, select Automatically attach a label to my images based on the folder they’re stored in.

This creates data labeling of the images as folder names.

You can now see the generated S3 bucket permissions policy.

  1. Copy the JSON policy.

  1. Navigate to the S3 bucket.
  2. On the Permission tab, under Bucket policy, choose Edit.
  3. Enter the JSON policy you copied.
  4. Chose Save changes.

  1. Choose Submit.

You can see that image labeling is organized based on the folder name.

Training your model

After you label your images, you’re ready to train your model.

  1. Choose Train Model.
  2. For Choose project, choose your project plant-leaf-disease-detection.
  3. For Choose training dataset, choose your dataset plant-leaf-disease-dataset.

As part of model training, Amazon Rekognition Custom Labels requires a labeled test dataset. Amazon Rekognition Custom Labels uses the test dataset to verify how well your trained model predicts the correct labels and generates evaluation metrics. Images in the test dataset are not used to train your model and should represent the same types of images you use with your model to analyze.

  1. For Create test set, select how you want to create your test dataset.

Amazon Rekognition Custom Labels provides three options:

  • Choose an existing test dataset
  • Create a new test dataset
  • Split training dataset

For this post, we select Split training dataset and let Amazon Rekognition hold back 20% of the images for testing and use the remaining 80% of the images to train the model.

Our model took approximately 1 hour to train. The training time required for your model depends on many factors, including the number of images provided in the dataset and the complexity of the model.

When training is complete, Amazon Rekognition Custom Labels outputs key quality metrics, including F1 score, precision, recall, and the assumed threshold for each label. For more information about metrics, see Metrics for Evaluating Your Model.

Our evaluation results show that our model has a precision of 1.0 for Bacterial-Leaf-Blight and Brown-Spot, which means that no objects were mistakenly identified (false positives) in our test set. Our model also didn’t miss any objects in our test set (false negatives), which is reflected in our recall score of 1. You can often use the F1 score as an overall quality score because it takes both precision and recall into account. Finally, we see that our assumed threshold to generate the F1 score, precision, and recall metrics each category is 0.62, 0.69, and 0.54 for Bacterial-Leaf-Blight, Brown-Spot, and Leaf-Smut, respectively. By default, our model returns predictions above this assumed threshold.

We can also choose View test results to see how our model performed on each test image. The following screenshot shows an example of a correctly identified image of bacterial leaf blight during the model testing (true positive).

Testing your model

Your plant disease detection model is now ready for use. Amazon Rekognition Custom Labels provides the API calls for starting, using, and stopping your model; you don’t need to manage any infrastructure. For more information, see Starting or Stopping an Amazon Rekognition Custom Labels Model (Console).

In addition to using the API, you can also use the Custom Labels Demonstration. This CloudFormation template enables you to set up a custom, password-protected UI where you can start and stop your models and run demonstration inferences.

Once deployed, the application can be accessed using a web browser using the address specified in url output from the CloudFormation stack created during deployment of the solution.

  1. Choose Start the model.

  1. Provide the inference unit required. For this example, let’s give a value of 1.

You’re charged for the amount of time, in minutes, that the model is running. For more information, see Inference hours.

It might take a while to start.

  1. Choose the model name.

  1. Choose Upload.

A window opens for you to choose the plant leaf image from your local drive.

The model detects the disease in the uploaded leaf image along with confidence score. It also gives the pest control recommendation based on the type of disease.

Cleaning up

To avoid incurring unnecessary charges, delete the resources used in this walkthrough when not in use. For instructions, see the following:

Conclusion

In this post, we showed you how to create an object detection model with Amazon Rekognition Custom Labels. This feature makes it easy to train a custom model that can detect an object class without needing to specify other objects or losing accuracy in its results.

For more information about using custom labels, see What Is Amazon Rekognition Custom Labels?

About the Authors

Dhiraj Thakur is a Solutions Architect with Amazon Web Services. He works with AWS customers and partners to provide guidance on enterprise cloud adoption, migration, and strategy. He is passionate about technology and enjoys building and experimenting in the analytics and AI/ML space.

 

 

Sameer Goel is a Solutions Architect in Seattle, who drives customer success by building prototypes on cutting-edge initiatives. Prior to joining AWS, Sameer graduated with a master’s degree from NEU Boston, with a concentration in data science. He enjoys building and experimenting with AI/ML projects on Raspberry Pi.

Read More

NVIDIA-Powered Systems Ready to Bask in Ice Lake

Data-hungry workloads such as machine learning and data analytics have become commonplace. To cope with these compute-intensive tasks, enterprises need accelerated servers that are optimized for high performance.

Intel’s 3rd Gen Intel Xeon Scalable processors (code-named “Ice Lake”), launched today, are based on a new architecture that enables a major leap in performance and scalability. These new systems are an ideal platform for enterprise accelerated computing, when enhanced with NVIDIA GPUs and networking, and include features that are well-suited for GPU-accelerated applications.

Ice Lake platform benefits for accelerated computing.

The move to PCIe Gen 4 doubles the data transfer rate from the prior generation, and now matches the native speed of NVIDIA Ampere architecture-based GPUs, such as the NVIDIA A100 Tensor Core GPU. This speeds throughput to and from the GPU, which is especially important to machine learning workloads that involve vast amounts of training data. This also improves transfer speeds for data-intensive tasks like 3D design for NVIDIA RTX Virtual Workstations accelerated by the powerful NVIDIA A40 data center GPU and others.

Faster PCIe performance also accelerates GPU direct memory access transfers. Faster I/O communication of video data between the GPU and GPUDirect for Video-enabled devices delivers a powerful solution for live broadcasts.

The higher data rate additionally enables networking speeds of 200Gb/s, such as in the NVIDIA ConnectX family of HDR 200Gb/s InfiniBand adapters and 200Gb/s Ethernet NICs, as well as the upcoming NDR 400Gb/s InfiniBand adapter technology.

The Ice Lake platform supports 64 PCIe lanes, so more hardware accelerators – including GPUs and networking cards – can be installed in the same server, enabling a greater density of acceleration per host. This also means that greater user density can be achieved for multimedia-rich VDI environments accelerated by the latest NVIDIA GPUs and NVIDIA Virtual PC software.

These enhancements allow for unprecedented scaling of GPU acceleration. Enterprises can tackle the biggest jobs by using more GPUs within a host, as well as more effectively connecting GPUs across multiple hosts.

Intel has also made Ice Lake’s memory subsystem more performant. The number of DDR4 memory channels has increased from six to eight and the data transfer rate for memory now has a maximum speed at 3,200 MHz.  This allows for greater bandwidth of data transfer from main memory to the GPU and networking, which can increase throughput for data-intensive workloads.

Finally, the processor itself has improved in ways that will benefit accelerated computing workloads. The 10-15 percent increase in instructions per clock can lead to an overall performance improvement of up to 40 percent for the CPU portion of accelerated workloads. There are also more cores — up to 76 in the Platinum 9xxx variant. This will enable a greater density of virtual desktop sessions per host, so that GPU investments in a server can go further.

We’re excited to see partners already announcing new Ice Lake systems accelerated by NVIDIA GPUs, including Dell Technologies with the Dell EMC PowerEdge R750xa, purpose built for GPU acceleration, and new Lenovo ThinkSystem Servers, built on 3rd Gen Intel Xeon Scalable processors and PCIe Gen4, with several models powered by NVIDIA GPUs.

Intel’s new Ice Lake platform, with accelerator hardware, is a great choice for enterprise customers who plan to update their data center. Its new architectural enhancements enable enterprises to run accelerated applications with better performance and at data center scale and our mutual customers will be able to quickly experience its benefits.

Visit the NVIDIA Qualified Server Catalog to see a list of GPU-accelerated server models with Ice Lake CPUs, and be sure to check back as more systems are added.

The post NVIDIA-Powered Systems Ready to Bask in Ice Lake appeared first on The Official NVIDIA Blog.

Read More

World of Difference: GTC to Spotlight AI Developers in Emerging Markets

Startups don’t just come from Silicon Valley — they hail from Senegal, Saudi Arabia, Pakistan, and beyond. And hundreds will take the stage at the GPU Technology Conference.

GTC, running April 12-16, will spotlight developers and startups advancing AI in Africa, Latin America, Southeast Asia, and the Middle East. Registration is free, and provides access to 1,500+ talks, as well as dozens of hands-on training sessions, demos and networking events.

Several panels and talks will focus on supporting developer ecosystems in emerging markets and opening access for communities to solve pressing regional problems with AI.

NVIDIA Inception, an acceleration platform for AI and data science startups, will host an Emerging Markets Pavilion where attendees can catch on-demand lightning talks from startup founders in healthcare, retail, energy and financial services. And developers from around the world will have access to online training programs through the NVIDIA Deep Learning Institute.

Beyond GTC, NVIDIA is exploring opportunities and pathways to reach data science and deep learning developers around the world. We’re working with groups like the data science competition platform Zindi to sponsor AI hackathons in Africa — and so are our NVIDIA Inception members, like Instadeep, an AI startup with offices in Tunisia, Nigeria, Kenya, England and France.

Programs like these, including the NVIDIA Developer Program, aim to support the next generation of developers, innovators and leaders with the resources to drive AI breakthroughs worldwide.

Focus on Emerging Developer Communities

While AI developers and startup founders come from diverse backgrounds and places, not all receive equivalent support and opportunities. At GTC, speakers from NVIDIA, Amazon Web Services, Google and Microsoft will join nonprofit founders and startup CEOs to discuss how we can bolster developer ecosystems in emerging markets.

Session topics include:

Startups Star in the NVIDIA Inception Pavilion

The NVIDIA Inception program includes more than 7,500 AI and data science startups from around the world. More than 300 will present at GTC.

It all kicks off after NVIDIA CEO Jensen Huang’s opening keynote on April 12, with a panel led by Jeff Herbst, our VP of business development and head of NVIDIA Inception.

The panel, AI Startups: NVIDIA Inception Insights and Trends from Around the World, will discuss efforts and challenges to nurture a broad cohort of young companies, including those from underserved and underrepresented markets. In addition to reps from NVIDIA, the panel will include Noga Tal, global director of partnerships at Microsoft for Startups; Maribel Lopez, co-founder of the Emerging Technology Research Council; and Badr Idrissi, CEO of Atlan Space, a Morocco-based NVIDIA Inception member.

Hosted by NVIDIA Inception, a virtual Emerging Markets Pavilion will feature global startups including:

Visit the GTC site to learn more and register.

The post World of Difference: GTC to Spotlight AI Developers in Emerging Markets appeared first on The Official NVIDIA Blog.

Read More

Join AWS at NVIDIA GTC 21, April 12–16

Starting Monday, April 12, 2021, the NVIDIA GPU Technology Conference (GTC) is offering online sessions for you to learn AWS best practices to accomplish your machine learning (ML), virtual workstations, high performance computing (HPC), and Internet of Things (IoT) goals faster and more easily.

Amazon Elastic Compute Cloud (Amazon EC2) instances powered by NVIDIA GPUs deliver the scalable performance needed for fast ML training, cost-effective ML inference, flexible remote virtual workstations, and powerful HPC computations. At the edge, you can use AWS IoT Greengrass and Amazon SageMaker Neo to extend a wide range of AWS Cloud services and ML inference to NVIDIA-based edge devices so the devices can act locally on the data they generate.

AWS is a Global Diamond Sponsor of the conference.

Available sessions

ML infrastructure:

ML with Amazon SageMaker:

ML deep dive:

High performance computing:

Internet of Things:

Edge computing with AWS Wavelength:

Automotive:

Computer vision with AWS Panorama:

Game tech:

Visit AWS at NVIDIA GTC 21 for more details and register for free for access to this content during the week of April 12, 2021. See you there!

About the Author

Geoff Murase is a Senior Product Marketing Manager for AWS EC2 accelerated computing instances, helping customers meet their compute needs by providing access to hardware-based compute accelerators such as Graphics Processing Units (GPUs) or Field Programmable Gate Arrays (FPGAs). In his spare time, he enjoys playing basketball and biking with his family.

Read More