배움에는 끝이 없습니다

청중 수준: 초심자 기술

전제 조건: 프로그래밍의 기본 개념에 대한 이해
함수, 루프, 사전 및 어레이 등의 Python.

딥 러닝이 컴퓨터 비전 및 자연어 처리 분야의 실습 예제를 어떻게 해결하는지 알아봅니다. 고도로 정확한 결과를 얻기 위해 처음부터 도구와 요령을 익히며 딥러닝 모델을 훈련해 보세요. 또한 무료로 이용 가능한, 사전 준비 트레이닝 첨단 모델을 활용하여 시간을 절약하고 딥 러닝 애플리케이션을 빠르게 시작해 실행하는 방법을 알아봅니다.

이 워크숍 참여 시 학습 내용:
- 딥 러닝 모델 트레이닝에 필요한 기본 기술과 도구 학습
– 일반적인 딥 러닝 데이터 유형 및 모델 아키텍처에 대한 경험 축적
– 모델 정확도 향상을 위해 데이터 증가를 통해 데이터세트 강화
– 모델 간 전이 학습(Transfer Learning)을 활용해 더 적은 데이터와 컴퓨팅 성능으로 효율적인 결과 달성
– 첨단 딥 러닝 프레임워크로 자체 프로젝트를 수행할 수 있는 자신감 획득

Audience Level: Intermediate Technical

Prerequisite(s):
> Familiarity with defensive cybersecurity themes
> Professional data science and/or data analysis experience
> Competency with the Python programming language
> Competency with the Linux command line

Session Abstract:
The NVIDIA Morpheus AI framework lets cybersecurity developers and practitioners harness the power of GPU computing to implement cybersecurity solutions that perform on a scale never before possible. With Morpheus, cybersecurity developers can create optimized AI pipelines for filtering, processing, and classifying large volumes of real-time data. Bringing a new level of information security to data centers, Morpheus enables dynamic protection, real-time telemetry, and adaptive defenses for detecting and remediating cybersecurity threats.

In this workshop you will get hands-on experience solving real-world cybersecurity challenges using the NVIDIA Morpheus AI framework. By the time you complete this workshop you will:

> Be able to build end-to-end Morpheus pipelines to process and analyze massive data streams using NVIDIA GPUs and AI.
> Be able to fine-tune Morpheus's out-of-the-box AI toolkit for your organization's specific needs.
> Understand the key components of the Morpheus framework, how they work together, and how to get started with them on your own.

Audience Level: Beginner Technical

Prerequisite(s): Basic C/C++ competency including familiarity with variable types, loops, conditional statements, functions, and array manipulations.

Session Abstract: The CUDA computing platform enables the acceleration of CPU-only applications to run on the world’s fastest massively parallel GPUs. Experience C/C++ application acceleration by:

> Accelerating CPU-only applications to run their latent parallelism on GPUs
> Utilizing essential CUDA memory management techniques to optimize accelerated applications
> Exposing accelerated application potential for concurrency and exploiting it with CUDA streams
> Leveraging Nsight Systems to guide and check your work

Upon completion, you’ll be able to accelerate and optimize existing C/C++ CPU-only applications using the most essential CUDA techniques and Nsight Systems. You’ll understand an iterative style of CUDA development that will allow you to ship accelerated applications fast.

Audience Level: Intermediate Technical

Prerequisite(s):
> Experience with Python coding and use of library functions and parameters
> Fundamental understanding of a deep learning framework such as TensorFlow, PyTorch, or Keras
> Basic understanding of neural networks

Session Abstract: You’ll learn how to use Transformer-based natural language processing models for text classification tasks, such as categorizing documents. You will also learn how to leverage Transformer-based models for named-entity recognition (NER) tasks and how to analyze various model features, constraints, and characteristics to determine which model is best suited for a particular use case based on metrics, domain specificity, and available resources.

By participating in this workshop, you’ll be able to:
> Understand how text embeddings have rapidly evolved in NLP tasks such as Word2Vec, recurrent neural network (RNN)-based embeddings, and Transformers
> See how Transformer architecture features, especially self-attention, are used to create language models without RNNs
> Use self-supervision to improve the Transformer architecture in BERT, Megatron, and other variants for superior NLP results
> Leverage pre-trained, modern NLP models to solve multiple tasks such as text classification, NER, and question answering
> Manage inference challenges and deploy refined models for live applications

Audience Level: Beginner Technical

Prerequisite(s): Experience with CNNs and Python

Session Abstract: The amount of information moving through our world’s telecommunications infrastructure makes it one of the most complex and dynamic systems that humanity has ever built. In this workshop, you’ll implement multiple AI-based solutions to solve an important telecommunications problem: identifying network intrusions.

In this workshop, you’ll:

> Implement three different anomaly detection techniques: accelerated XGBoost, deep learning-based autoencoders, and generative adversarial networks (GANs)
> Build and compare supervised learning with unsupervised learning-based solutions
> Discuss other use cases within your industry that could benefit from modern computing approaches

Upon completion, you'll be able to detect anomalies within large datasets using supervised and unsupervised machine learning.

Audience Level: Advanced Technical

Prerequisite(s): Intermediate experience writing CUDA C/C++ applications

Session Abstract: Building cluster scale accelerated computing applications that utilize multiple GPUs across multiple nodes requires a distinct set of skills. In this workshop you will learn how to do this by:

> Building efficient cluster-enabled CUDA C++ applications using CUDA-aware MPI and NVSHMEM
> Utilizing NCCL for highly optimized collective communications over multiple GPUs both within and across nodes
> Learning how to identify the best programming approach for your application and multi-node configuration

Upon completion, you will be able to write efficient CUDA C++ applications at cluster scale using a variety of programming models well-suited to many kinds of applications and multi-GPU multi-node configurations.

Audience Level: Beginner Technical

Prerequisite(s): An understanding of fundamental programming concepts in Python such as functions, loops, dictionaries, and arrays.

Session Abstract: Discover how deep learning works through hands-on exercises in computer vision and natural language processing. You’ll train deep learning models from scratch, learning tools and tricks to achieve highly accurate results. You’ll also learn to leverage freely available, state-of-the-art pre-trained models to save time and get your deep learning application up and running quickly.

By participating in this is workshop, you’ll:
> Learn the fundamental techniques and tools required to train a deep learning model
> Gain experience with common deep learning data types and model architectures
> Enhance datasets through data augmentation to improve model accuracy
> Leverage transfer learning between models to achieve efficient results with less data and computation
> Build confidence to take on your own project with a modern deep learning framework

Audience Level: Intermediate Technical

Prerequisite(s):
> Experience with Python coding and use of library functions and parameters
> Fundamental understanding of a deep learning framework such as TensorFlow, PyTorch, or
Keras
> Basic understanding of neural networks

Session Abstract: You’ll learn how to use Transformer-based natural language processing models for text classification tasks, such as categorizing documents. You will also learn how to leverage Transformer-based models for named-entity recognition (NER) tasks and how to analyze various model features, constraints, and characteristics to determine which model is best suited for a particular use case based on metrics, domain specificity, and available resources.

By participating in this workshop, you’ll be able to:
> Understand how text embeddings have rapidly evolved in NLP tasks such as Word2Vec, recurrent neural network (RNN)-based embeddings, and Transformers
> See how Transformer architecture features, especially self-attention, are used to create language models without RNNs
> Use self-supervision to improve the Transformer architecture in BERT, Megatron, and other variants for superior NLP results
> Leverage pre-trained, modern NLP models to solve multiple tasks such as text classification, NER, and question answering
> Manage inference challenges and deploy refined models for live applications

Audience Level: Beginner Technical

Prerequisite(s): Experience with Python, ideally including pandas and NumPy

Session Abstract: RAPIDS is a collection of data science libraries that allows end-to-end GPU acceleration for data science workflows. In this training, you'll:

> Use cuDF and Dask to ingest and manipulate massive datasets directly on the GPU
> Apply a wide variety of GPU-accelerated machine learning algorithms, including XGBoost, cuGRAPH, and cuML, to perform data analysis at massive scale
> Perform multiple analysis tasks on massive datasets in an effort to stave off a simulated epidemic outbreak affecting the UK

Upon completion, you'll be able to load, manipulate, and analyze data orders of magnitude faster than before, enabling more iteration cycles and drastically improving productivity.

Audience Level: Intermediate Technical

Prerequisite(s): Python programming experience, basic understanding of neural networks, and a fundamental understanding of a deep learning framework such as TensorFlow or PyTorch

Session Abstract: Learn how to quickly build and deploy production quality conversational AI applications with real-time transcription and natural language processing (NLP) capabilities.

Conversational AI pipelines are complex and expensive to develop from scratch. In this course, you'll learn to build a conversational AI service using the NVIDIA Riva framework. Riva provides a complete, GPU-accelerated software stack, making it easy for developers to quickly create, deploy, and run end-to-end, real-time conversational AI applications that can understand terminology unique to each company and its customers. The Riva framework includes pretrained conversational AI models, tools, and optimized services for speech, vision, and natural language understanding (NLU) tasks. With Riva, developers can create customized language-based AI services for intelligent virtual assistants, virtual customer service agents, real-time transcription, multi-user diarization,chatbots and much more.

In this course, you will integrate NVIDIA Riva automatic speech recognition (ASR) and named entity recognition (NER) models with a web-based application to produce transcriptions of audio inputs with highlighted relevant text. You'll then customize the NER model using NVIDIA TAO Toolkit to provide different targeted highlights for the application. Finally, you'll explore production-level deployment performance and scaling considerations of Riva services with Helm Charts and Kubernetes clusters.

You'll learn how to:

> Deploy and enable pretrained ASR and NER models on Riva for a Conversational AI Application
> Explore the Riva ASR service with audio examples
> Fine-tune and deploy a domain-specific NER model with TAO Toolkit
> Deploy a production-level conversational AI application with a Helm Chart for scaling in Kubernetes clusters

Upon completion, you’ll be able to build Conversational AI applications and deploy them as services.

Audience Level: Beginner Technical

Prerequisite(s): Experience with CNNs and Python

Session Abstract: The amount of information moving through our world’s telecommunications infrastructure makes it one of the most complex and dynamic systems that humanity has ever built. In this workshop, you’ll implement multiple AI-based solutions to solve an important telecommunications problem: identifying network intrusions.

In this workshop, you’ll:

> Implement three different anomaly detection techniques: accelerated XGBoost, deep learning-based autoencoders, and generative adversarial networks (GANs)
> Build and compare supervised learning with unsupervised learning-based solutions
> Discuss other use cases within your industry that could benefit from modern computing approaches

Upon completion, you'll be able to detect anomalies within large datasets using supervised and unsupervised machine learning.

Audience Level: Beginner Technical

Prerequisite(s): An understanding of fundamental programming concepts in Python such as functions, loops, dictionaries, and arrays.

Session Abstract: Discover how deep learning works through hands-on exercises in computer vision and natural language processing. You’ll train deep learning models from scratch, learning tools and tricks to achieve highly accurate results. You’ll also learn to leverage freely available, state-of-the-art pre-trained models to save time and get your deep learning application up and running quickly.

By participating in this is workshop, you’ll:
> Learn the fundamental techniques and tools required to train a deep learning model
> Gain experience with common deep learning data types and model architectures
> Enhance datasets through data augmentation to improve model accuracy
> Leverage transfer learning between models to achieve efficient results with less data and computation
> Build confidence to take on your own project with a modern deep learning framework

Audience Level: Intermediate Technical

Prerequisite(s):
>Experience with Python coding and use of library functions and parameters
>Fundamental understanding of a deep learning framework such as TensorFlow, PyTorch, or
Keras
>Basic understanding of neural networks

Session Abstract: You’ll learn how to use Transformer-based natural language processing models for text classification tasks, such as categorizing documents. You will also learn how to leverage Transformer-based models for named-entity recognition (NER) tasks and how to analyze various model features, constraints, and characteristics to determine which model is best suited for a particular use case based on metrics, domain specificity, and available resources.

By participating in this workshop, you’ll be able to:
> Understand how text embeddings have rapidly evolved in NLP tasks such as Word2Vec, recurrent neural network (RNN)-based embeddings, and Transformers
> See how Transformer architecture features, especially self-attention, are used to create language models without RNNs
> Use self-supervision to improve the Transformer architecture in BERT, Megatron, and other variants for superior NLP results
> Leverage pre-trained, modern NLP models to solve multiple tasks such as text classification, NER, and question answering
> Manage inference challenges and deploy refined models for live applications

Audience Level: Beginner Technical

Prerequisite(s): Experience with Python; basic understanding of data processing and deep learning.

Session Abstract: Learn how to design, train, test, and deploy building blocks of a hardware-accelerated industrial inspection pipeline.

Explore how to build a deep learning model to automate the verification of capacitors in NVIDIA's printed circuit board (PCB) using a real production dataset. This can lower the verification cost and increase the production throughput across a variety of manufacturing use cases. You'll learn how to:

> Extract meaningful insights from the provided dataset using Pandas DataFrame and NumPy library
> Apply transfer-learning to a deep learning classification model known as InceptionV3
> Fine-tune the deep learning model and set up evaluation metrics

Upon completion, you'll be able to design, train, test, and deploy building blocks of a hardware-accelerated industrial inspection pipeline.

Audience Level: Advanced Technical

Prerequisite(s):
> Good understanding of PyTorch, deep learning and data parallel training concepts
> Practice with multi-GPU training and natural language processing are useful, but optional

Session Abstract: Very large deep neural networks (DNNs), whether applied to natural language processing (e.g., GPT-3), computer vision (e.g., huge Vision Transformers), or speech AI (e.g., Wave2Vec 2) have certain properties that set them apart from their smaller counterparts. As DNNs become larger and are trained on progressively larger datasets, they can adapt to new tasks with just a handful of training examples, accelerating the route toward general artificial intelligence. Training models that contain tens to hundreds of billions of parameters on vast datasets isn’t trivial and requires a unique combination of AI, high-performance computing (HPC), and systems knowledge. The goal of this course is to demonstrate how to train the largest of neural networks and deploy them to production.

In this workshop, you’ll learn how to:

> Train neural networks across multiple servers
> Use techniques such as activation checkpointing, gradient accumulation, and various forms of model parallelism to overcome the challenges associated with large-model memory footprint
> Capture and understand training performance characteristics to optimize model architecture
> Deploy very large multi-GPU models to production using NVIDIA Triton™ Inference Server

Audience Level: Intermediate Technical

Prerequisite(s): Python programming experience, basic understanding of neural networks, and a fundamental understanding of a deep learning framework such as TensorFlow or PyTorch

Session Abstract: Learn how to quickly build and deploy production quality conversational AI applications with real-time transcription and natural language processing (NLP) capabilities.

Conversational AI pipelines are complex and expensive to develop from scratch. In this course, you'll learn to build a conversational AI service using the NVIDIA Riva framework. Riva provides a complete, GPU-accelerated software stack, making it easy for developers to quickly create, deploy, and run end-to-end, real-time conversational AI applications that can understand terminology unique to each company and its customers. The Riva framework includes pretrained conversational AI models, tools, and optimized services for speech, vision, and natural language understanding (NLU) tasks.  With Riva, developers can create customized language-based AI services for intelligent virtual assistants, virtual customer service agents, real-time transcription, multi-user diarization,chatbots and much more.

In this course, you will integrate NVIDIA Riva automatic speech recognition (ASR) and named entity recognition (NER) models with a web-based application to produce transcriptions of audio inputs with highlighted relevant text.  You'll then customize the NER model using  NVIDIA TAO Toolkit to provide different targeted highlights for the application. Finally, you'll explore production-level deployment performance and scaling considerations of Riva services with Helm Charts and Kubernetes clusters.

You'll learn how to:

> Deploy and enable pretrained ASR and NER models on Riva for a Conversational AI Application
> Explore the Riva ASR service with audio examples
> Fine-tune and deploy a domain-specific NER model with TAO Toolkit
> Deploy a production-level conversational AI application with a Helm Chart for scaling in Kubernetes clusters

Upon completion, you’ll be able to build Conversational AI applications and deploy them as services.

Audience Level: Intermediate Technical

Prerequisite(s):
> Familiarity with defensive cybersecurity themes
> Professional data science and/or data analysis experience
> Competency with the Python programming language
> Competency with the Linux command line

Session Abstract: The NVIDIA Morpheus AI framework lets cybersecurity developers and practitioners harness the power of GPU computing to implement cybersecurity solutions that perform on a scale never before possible. With Morpheus, cybersecurity developers can create optimized AI pipelines for filtering, processing, and classifying large volumes of real-time data. Bringing a new level of information security to data centers, Morpheus enables dynamic protection, real-time telemetry, and adaptive defenses for detecting and remediating cybersecurity threats.

In this workshop you will get hands-on experience solving real-world cybersecurity challenges using the NVIDIA Morpheus AI framework. By the time you complete this workshop you will:

> Be able to build end-to-end Morpheus pipelines to process and analyze massive data streams using NVIDIA GPUs and AI.
> Be able to fine-tune Morpheus's out-of-the-box AI toolkit for your organization's specific needs.
> Understand the key components of the Morpheus framework, how they work together, and how to get started with them on your own.

모델 병렬화 : 라지 스케일 뉴럴 네트워크 구축 및 배포

수업 레벨 : 중급

전제 조건 :
> PyTorch, 딥 러닝과 데이터 병렬 학습 개념에 대한 이해
> 멀티 GPU 학습과 자연어 처리 경험 (선택)

세션 개요 : 자연어 처리(예:GPT-3) , 컴퓨터 비전(예:Huge Vision Transformers) , 그리고 음성 AI (예:Wave2Vec 2)에 적용되는 매우 큰 심층 신경망(DNN)은 더 작은 모델들과 구별되는 특정학 속성을 가지고 있습니다. DNN이 점점 커지고 점차적으로 더 큰 데이터세트에서 학습됨에 따라, DNN은 많은 학습 예제를 통해 새로운 작업에 잘 적응할 수 있으며, 일반 인공 지능을 향한 길을 가속화하고 있습니다. 거대한 데이터 세트에서 수백 또는 수천억개의 파라미터를 가지고 있는 학습 모델은 사소한 것이 아니며, AI, 고성능 컴퓨팅(HPC)그리고 시스템 지식의 고유한 조합이 필요합니다. 이번 과정의 목표는 가장 큰 뉴럴 네트워크를 학습하고 어떻게 양산 환경에 배포할 수 있는지를 설명하는 것입니다.

이번 워크숍을 통해 다음을 배울 수 있습니다:

> 여러개의 서버에 걸쳐 뉴럴 네트워크 학습하기
> 액티베이션 체크포인트, 그레디언트 누적, 또는 다양한 형태의 모델 병렬화 기법을 활용하여 거대 모델의 메모리 차지 공간과 연관된 다양한 과제를 해결하기
> 학습 성능 특성을 이해하고 모델 구조를 최적화하기
> NVIDIA Triton™ Inference Server을 활용하여 매우 큰 규모의 멀티 GPU 모델을 실 운영 환경에 구축하기

Audience Level: Beginner Technical

Prerequisite(s): Experience with Python, ideally including pandas and NumPy

Session Abstract: RAPIDS is a collection of data science libraries that allows end-to-end GPU acceleration for data science workflows. In this training, you'll:

> Use cuDF and Dask to ingest and manipulate massive datasets directly on the GPU
> Apply a wide variety of GPU-accelerated machine learning algorithms, including XGBoost, cuGRAPH, and cuML, to perform data analysis at massive scale
> Perform multiple analysis tasks on massive datasets in an effort to stave off a simulated epidemic outbreak affecting the UK

Upon completion, you'll be able to load, manipulate, and analyze data orders of magnitude faster than before, enabling more iteration cycles and drastically improving productivity.

Audience Level: Beginner Technical

Prerequisite(s): Experience with Python and deep networks

Session Abstract: Learn how to identify anomalies and failures in time-series data, estimate the remaining useful life of the corresponding parts, and use this information to map anomalies to failure conditions.

You'll learn how to:

> Leverage predictive maintenance to manage failures and avoid costly unplanned downtimes
> Identify key challenges around identifying anomalies that can lead to costly breakdowns
> Use time-series data to predict outcomes using machine learning classification models with XGBoost
> Apply predictive maintenance procedures by using a long short-term memory ( LSTM)-based model to predict device failure
> Experiment with autoencoders to detect anomalies by using the time-series sequences from the previous steps

Upon completion, you’ll understand how to use AI to predict the condition of equipment and estimate when maintenance should be performed.