Algorithms

Top 12 Talks from NVIDIA GTC 2022 to Get You Started

Banner depicting different images from GTC

NVIDIA GTC 2022 is a global developer conference that explores the latest technologies that drive industry transformation, provides participants direct access to experts, and features groundbreaking research and solutions for our world’s challenges.

This year, there are more than 900 sessions throughout the four-day event. Technologists from emerging startups and top enterprises such as Deloitte, Microsoft, Sony, and more, are set to share experiences, discoveries, and implementations related to pushing the boundaries and getting the most value out of recent technological advancements.

Out of these almost a thousand sessions at GTC, which ones should you start with? If you’re looking for a comprehensive overview of the AI landscape today, then we recommend the following 12 sessions that talk about general AI strategy, and the exciting spaces of edge AI and deep learning.

1. Achieving AI Success with World-Class AI Infrastructure: Insights from AI High Performers

What is it about:

This session brings together AI implementers who’ve deployed AI at scale using NVIDIA DGX systems. Discover learnings and best practices that enabled these companies to achieve the highest bottom-line impact and ROI from AI.

Who are the speakers:

  • Tony Paikeday, Senior Director, Artificial Intelligence Systems, NVIDIA
  • Mohamed Sidahmed, Deep Learning and AI R&D Manager, Shell
  • Prashant Kukde, Assistant Vice President, Conversational AI, RingCentral
  • Woomyoung Park, Conversational AI Leader, CLOVA CIC, NAVER
  • Angelica Demarchi Munhoz, Software Development Manager, SiDi
  • Vitor Rolim, Senior Software Developer (ML/NLP), SiDi

Why watch it:

Implementing AI requires resources that not only include the technology itself but also the skills and processes that ensure its long-term success. In this session, you’ll learn from the best practices of AI experts including RingCentral, Shell, and NAVER, to help inform your next AI project and secure better performance and returns on investment. You can view the session on GTC here.

2. Spend More Time on ML, Less Time on Ops (Presented by Weights & Biases) 

What is it about:

In traditional software, the git “diff” is used to see what code changes led to strange behavior. Can you do the same for machine learning? Learn the best practices for setting up your experimentation pipeline to easily and systematically track, compare, and share experiments over time. This way, you can spend more time on ML — and less on ops. 

Who is the speaker: 

  • Andrew Truong, Machine Learning Engineer, Weights & Biases

Why watch it:

Behind the scenes, one of the critical components of seamless AI implementations is the operations that guide the AI lifecycle. Specific to machine learning, this session from Weights & Biases will help you navigate the experimentation pipeline which is an important foundation for successful ML projects. You can view the session on GTC here.

3. Why Synthetic Data is Important for Your Business: Strategies and Implementations Across Industries

What is it about:

Focused on 3D synthetic data generation spanning partners and customers, this panel will showcase the underlying value of OV for high-fidelity, accurate data generation across different use cases and industries. 

Who are the speakers:

  • Rev Lebaredian, VP Omniverse & Simulation Technology, NVIDIA
  • Gerard Andrews, Product Marketing, NVIDIA
  • Yashar Behzado, Phd., CEO & Founder, Synthesis AI
  • Pedro Urbina, Software Developer Manager, Microsoft
  • Nikita Jaipuria, Technical Expert – AI Based Modeling For DAT, Ford
  • Gil Elbaz, CTO and Co-founder, Datagen

Why watch it:

Without enough data, it’s close to impossible to deploy useful AI models, especially for real-world use cases. This session, which features companies such as Datagen and Synthesis AI, looks at 3D synthetic data generation that can be helpful in building valuable AI solutions. You can view the session on GTC here.

4. How to Improve Model Efficiency with Hardware-Aware Neural Architecture Search (Presented by Deci AI)  

What is it about:

Running successful and efficient inference at scale requires meeting such performance criteria as accuracy, latency, throughput, and model size, among others. Neural architecture search (NAS) holds the power to automate the cumbersome deep learning model development process, as well as quickly and efficiently generate deep neural networks that are designed to meet specific production constraints. Deci’s AutoNAC (Automated Neural Architecture Construction) technology does this by finding the best algorithm that takes into account all of the many parameters that are required to create powerful and efficient deep learning models. We’ll cover the evolution of NAS technology and recent advances that are making NAS viable for industry applications and commercial use. We’ll outline the algorithmic optimization process with case studies and best practices for achieving best-in-class accuracy and latency results on NVIDIA T4 GPU, Jetson Nano, and Xavier NX devices.

Who is the speaker:

  • Yonatan Geifman, CEO and Co-Founder, Deci

Why watch it:

Developing deep learning models has challenges that can range from large compute demands to the complexity of algorithms, making them hard to reach production. NAS can automate the development process and generate DNNs that take into consideration multiple parameters and production constraints. This session will discuss case studies and best practices for achieving best-in-class accuracy and latency results on NVIDIA devices. You can register for the session on GTC here.

5. Transforming AI and ML at the Edge with Microsoft and NVIDIA (Presented by Microsoft Azure)    

What is it about:

NVIDIA and Microsoft are working together to transform AI and machine learning, leveraging the power of the GPU at the edge combined with Azure AI services. Discover how to accelerate edge AI, simplify AI and machine learning training and deployment at scale, and make AI more accessible to non-developers and data scientists. Also get a glimpse of what’s to come next with edge AI and IoT innovation from Microsoft.

Who is the speaker:

  • Christa St. Pierre, Group Manager, Azure Edge, Microsoft

Why watch it:

Deep learning applications are typically hardware intensive, requiring GPUs that provide power and ensure computation of neural networks goes smoothly. But how do you do it on the edge? In this session, NVIDIA and Microsoft will discuss how they work together to accelerate and simplify edge AI. You can view the session on GTC here.

6. Building Secure and High-Performance Edge Cloud Platforms

What is it about:

Modern workloads fuel the demand for high-performance IT infrastructures at the network’s edge. Streaming video, gaming, and IoT performance are driven by the latency that edge platforms provide. To deliver on the promise of edge computing, StackPath has re-imagined the stack by integrating NVIDIA GPU and DPU acceleration technologies into its platform. Learn about StackPath’s journey to build, accelerate, and protect edge computing platforms.

Who is the speaker:

  • Ashok Ganesan, Chief Product Officer, StackPath

Why watch it:

The emerging trend of edge AI can open a lot of opportunities and benefits not only for companies but also for end-users. But moving high-performance computing at the edge brings with it challenges—from device limitations to privacy. Watch this session for a general overview of how to build, accelerate, and protect edge computing platforms. You can view the session on GTC here.

7. An End-to-end Walkthrough for Deploying Deep Learning Models on Jetson (Presented by Deci AI)

What is it about: 

Join us for a technical session packed with practical tips and tricks from model selection and training tools to running successful inference at the edge. We’ll demonstrate how to benchmark different models, leverage training best practices, easily implement TensorRT-based compilation and quantization, all while using the latest open-source libraries and other free community tools. You’ll gain practical knowledge of how to cut the guesswork, quickly gain state-of-the-art performance, maximize your Jetson devices’ compute power, and boost runtime for any AI-based application.

Who are the speakers:

  • Nadav Cohen, VP Product, Deci AI
  • Ofer Baratz, Deep Learning Product Manager, Deci AI

Why watch it:

Before we head into the next batch of sessions that feature real-world uses cases of edge AI and deep learning, here’s a technical overview for model deployment on NVIDIA Jetson devices. Catch this session by our team to learn actionable tips and tricks that will help you to successfully deploy your deep learning models into production—with exceptional performance that maximizes hardware and deliver the best runtime. You can register for the session on GTC here.

8. Accelerating Intelligent Spaces with Edge Computing

What is it about:

2021 saw massive growth in the demand for edge computing. From smart hospitals and cities to cashierless shops to self-driving cars, edge AI is needed more than ever. This session looks at the different considerations for deploying AI at the edge and how it can be used for various use cases.

Who is the speaker:

  • Justin Boitano, VP/GM Enterprise and Edge Computing, NVIDIA

Why watch it:

Discover the emerging use cases of deploying AI at the edge during this session. From cashierless shopping to self-driving cars, you’ll get a general introduction on how edge AI can improve the way humans and machines collaborate, better allocate work, enhance the process of product design, and more. You can view the session on GTC here.

9. AI at Scale in Agriculture Using 3D Deep Learning Technology

What is it about:

Autonomous robots in agriculture need to be efficient and agile in order to cope with ever-changing conditions and the chaos of the plants, where no two are the same. This session explains the fundamentals of scalable AI applications that enable regular operators to easily create and deploy computer vision algorithms in a no-code environment, and how 3D deep learning technology is becoming a breakthrough within agriculture. 

Who is the speaker:

  • Jonathan Berte, Founder & CEO, Robovision

Why watch it:

Agritech is one of the popular and increasing applications of edge AI that utilizes computer vision and 3D deep learning technology. In this session, Robovision shares the fundamentals of building scalable AI applications that can help operators to be more efficient and agile in face of ever-changing conditions. You can view the session on GTC here.

10. The Role of AI in Smart and Sustainable Urbanization 

What is it about:

It’s becoming increasingly important to ensure that advancements and innovations in urban planning are also achieving cities’ sustainability goals. Learn how AI is playing a big role in making urbanization smarter, but also more sustainable.

Who are the speakers:

  • Shilpa Kolhatkar, Global Head of AI Nations Business Development, NVIDIA
  • Amen Mashariki, Principal Scientist, NVIDIA
  • Donnel Baird, CEO, Blocpower
  • Dr. Oualid Ali, President, Future Cities Council, Canada
  • Daniel Zarrilli, Special Advisor, Climate & Sustainability, Columbia University

Why watch it:

Considering sustainability today is paramount to having better living and environmental conditions tomorrow. Discover from the public sector how AI can make the process of urbanization smarter and more sustainable, allowing cities to reach sustainability goals and increase the quality of life of people. You can view the session on GTC here.

11. Connect with Experts: AI at the Edge for Autonomous Machines, Robotics, and Intelligent Video Analytics

What is it about:

Developing solutions for vision, autonomous machines, or robotics? Share your challenges with NVIDIA experts in embedded edge AI development, the Jetson platform, and SDKs including ISAAC, DeepStream, and TLT.

Who are the speakers:

  • Richard Love, Marketing Manager, NVIDIA
  • Vipul Amin, System Architect, NVIDIA
  • Teresa Conceicao, Solutions Architect – Robotics and Omniverse Enterprise, NVIDIA
  • Maycon da Silva Carvalho, Field Applications Engineer – Jetson, NVIDIA
  • Abubakr Karali, Senior Solutions Architect, NVIDIA
  • Vincent Nguyen Quang Do, System Architect, NVIDIA
  • Ekaterina Sirazitdinova, Data Scientist – Deep Learning, Computer Vision and Video Analytics, NVIDIA

Why watch it:

Our last pick for deep learning use cases at the edge is also a session where you can get advice from NVIDIA experts. Join this session if you face challenges in vision, autonomous machines, or robotics that you want help with. You can view the session on GTC here.

12. AI For Good: Intersecting Humanity and Intelligent Systems

What is it about:

Along with the power of AI, machine learning, and data science, are complex ethical and social questions associated with bias, fairness, and transparency of algorithmic intelligence. Catch this session to enrich your understanding of human-AI interaction, and how these questions will touch upon AI research, entrepreneurship, education, and policies. 

Who are the speakers:

  • Nidhiya V Raj, AI Startups Lead – South Asia, NVIDIA
  • Darlington Akogo, Founder, karaAgro
  • Beena Ammanath, Executive Director, Deloitte Global AI Institute, Deloitte
  • Sara Smolley, Co-founder and Head of Partnerships, VoiceItt

Why watch it:

We’re wrapping up our list with an important trend in AI. More focus is being directed to the ethical and social implications of AI, driving businesses to consider these factors when building solutions. This session is a great overview of AI’s connection with us, humans, and how with a better approach to technology, companies can make society more equitable for all. You can view the session on GTC here.

Any sessions sparked your interest? You can check out the full list of sessions on the NVIDIA GTC page. Make sure to register so you get complete access to all the event has to offer including research posters, interactive panels, demos, podcasts, technical sessions, and more.

You May Also Like

Top 10 List of Large Language Models Reshaping the Open-Source Arena

From Top-k to Beam Search: Everything You Need to Know About LLM Decoding Strategies

High Performance AI - Deci Gen AI Deve Platform and Deci Nano

Introducing Deci’s Gen AI Development Platform and Deci-Nano

The latest deep learning insights, tips, and best practices delivered to your inbox.

Share
Add Your Heading Text Here
				
					from transformers import AutoFeatureExtractor, AutoModelForImageClassification

extractor = AutoFeatureExtractor.from_pretrained("microsoft/resnet-50")

model = AutoModelForImageClassification.from_pretrained("microsoft/resnet-50")