2023 Intel On Technology Innovation Conference Live Broadcast | Let more people understand the charm of AI
Foreword:
The 2023 Intel On Technology Innovation Conference will be held on December 19th | China Station |
助力开发者,让 AI 无处不在
, the conference is aimed at developers in the era of intelligent computing. Chinese experts will explain in depth Intel's latest generation computing platform that accelerates AI capabilities, supports open, multi-architecture software solutions and tools, and shapes future technology and application innovation.
Main areas:
AI:
GenAI 在 IA
Deployment and optimization on accelerated platforms On Intel platforms including clients, edge computing and data centers, various Intel® CPU、GPU
and NPU
other deep learning computing engines can be selected for different business scenarios, coupled with development OpenVINO™
and IPEX, ITEX, ITREX, BigDL-LLM, SynapseAI®, Optimum
optimization tool chains such as Accelerate the continuous parameter optimization and rapid deployment of generative artificial intelligence, allowing developers to flexibly and conveniently develop and deliver consumer and enterprise-level LLM/AIGC
business software.
Here I will pick one at random and expand on it.
Using OpenVINO™ to implement edge-side generative AI
OpenVINO
As Intel AI
a software development tool, it provides national AI ecological and technology integration support, including the industry's most popular AI framework models PyTorch、TensorFlow、OBBXRuntime、PaddlePaddle
. It also supports custom Chinese models. Developers OpenVINO
only need to focus on the construction of upper-layer AI applications while using them. OpenVINO
It will provide technical interfaces based on the underlying instruction set and hardware optimization of various Intel hardware platforms to unify the upper layer across multiple heterogeneous platforms. Development instructions.
The second part is superimposed on Diffusion
the part where the fine-tuning effect of the model is increased.
There is also an image-to-text model Pix2sturct
that can be used to extract basic information from the image and perform text understanding of the screenshot content.
About speech-related models:
SoftVC
It is a combinationRNN
andTransforner
model that can be used forAI
song generation, sound cloning, cheers and other application sound effects.BArc
It is commonTTS
, that is,Text to Speech
a model that can simulate various sounds such as music, noise, sound effects, etc. with high fidelity (approximately providing100
a variety of human voice models),AI
which is also a relatively widespread model in the field.Distil Whisper
It is a popular model currently and is mainly used for speech recognition input.
OpenVINO™ learning summary:
OpenVINO™
It is an open source application across various industries AI
. It provides an efficient model development tool suite to deploy generative and text generation applications through cases, which can help developers debug the model and control the accuracy of the model.
New generation AI PC computing platform:
Developing applications for PC
computing platforms AI
requires powerful hardware performance, optimized energy consumption, advanced tools, and an open ecosystem. In this topic, you will gain an in-depth understanding of how Intel's new generation AI PC
hardware platforms (such as Intel® Core™ Ultra
processors and Intel Radar™ graphics) and their excellent connectivity can help you develop AI applications and games more easily. We will also showcase Intel's open software and hardware ecosystem (such as OpenVINO™
Intel and Intel OneAPI
) and the latest solutions launched with partners to help you understand the latest developments in client development.
New generation of Xeon platform:
The new Xeon full-stack empowers cloud computing and accelerates innovation. The AI
application of cloud computing and the wave of the large model era are revolutionaryly redefining cloud computing, and a new generation of Intel® Xeon® Scalable processors have emerged. Its built-in Intel® AMX
acceleration engine brings ultimate acceleration performance to AI, and many other built-in accelerators (QAT, IAA, DSA, DLB)
can accelerate and optimize performance of various cloud services; at the same time, with the support of Intel® SGX
and Intel® TDX “
Double Insurance, it provides a safe operating environment It can protect business; Intel also has rich full-stack software optimization (Kernel, Hypervisor, AI FW and lib[oneDNN], IPEX 等)
, empowers various cloud software, fully unleashes the potential of Xeon, and accelerates cloud computing and AI
innovation through software and hardware collaboration.
Intel® Xeon® Empowers Cloud Data Security with Confidential Computing
The digital age has brought great convenience to everyone's life, and has also greatly improved the operating efficiency of enterprises. It has become a huge driving force for economic growth. However, another aspect of digitalization is security issues that cannot be underestimated.
Through this set of data, it can be seen that with the security risk issues in the digital field and the security problems they bring, personal privacy and data leakage and abuse caused by the lack of data security protection. Therefore, in recent years, many countries and regions around the world have introduced relevant laws and regulations to strictly regulate and guide data security, especially privacy protection and other issues. For example: California’s Consumer Privacy Act, and our country also introduced the Information Security Protection Law this year.
Achieving data security protection requires building a data security system that covers the entire life cycle of data. From a technical perspective, the life cycle of data includes three processes:
- Transmission process in the network
- Stored procedures on disk
- Processor or memory usage
There are two major technical categories for the protection of private data:
1. Categories based on cryptography such as: multi-party secure computing, homomorphic encryption, differential privacy
2. Building a trusted execution environment based on hardware assistance This environment can be called TEE and is more suitable for deployment need
- The confidentiality of the data use process must be ensured that the data is not leaked during use.
- The integrity of the data usage process must ensure that the user's data has not been tampered with during processing.
- The integrity of the data access program, ensuring that the program is secure and does not contain malicious code.
Intel® TDX Technology: https://www.intel.com/content/www/us/en/developer/tools/trust-domain-extensions/overview.html
Edge cloud collaboration:
Intel, a leader in cloud-native development and edge intelligence, collaborates with its modern "edge-cloud collaboration" application development experience, technology, products and platforms accumulated over many years of working with innovative partners to empower generative large models AI
. Core hybrid AI
solutions to accelerate their commercialization. OpenVINO™
Provide easy-to-use AI
model optimization and deployment methods, allowing AI
models to be efficiently deployed on Intel platforms from edge to cloud; edge security technology makes edge-cloud collaboration AI
solutions more reliable; Intel infrastructure processors are also used in AI
clusters and edges AI
Improve data center performance in other user scenarios.
advanced technology:
Join us as we pursue transformative solutions to the computing industry's biggest challenges and learn about the technology developments that will impact our work over the next 2-5 years and beyond. Topics include the future of Moore's Law, the future of chip design, quantum computing, the future of AI, neuromorphic computing, smart computing for "double carbon" and more.
If you want to know more about it, you can watch the live review: https://marketing.intel.cn/innovation#/
Summary of live viewing of Intel·on Innovation Technology Conference:
To briefly summarize this live broadcast, this Intel·on Innovation Technology Conference focuses on artificial intelligence, a new generation of AI PC computing platforms, a new generation of Xeon platforms, etc. The new Xeon full-stack empowers cloud computing, accelerates innovative AI applications, and shapes future technology and application innovations. This live broadcast was very rewarding. I hope that the digital economy and AI can make our lives more convenient.