Cerebras Systems builds the fastest AI accelerators in the industry. In this talk we will review how the size and scope of massive natural language processing (NLP) presents fundamental challenges to legacy compute and to traditional cloud providers. We will explore the importance of guaranteed node to node latency in large clusters, how that can’t be achieved in the cloud, and how it prevents linear and even deterministic scaling. We will examine the complexity of distributing NLP models over hundreds or thousands of GPUs and show how quickly and easily a cluster of Cerebras CS-2s is set up, and how linear scaling can be achieved over millions of compute cores with Cerebras technology. And finally, we will show how innovative customers are using clusters of Cerebras CS-2s to train large language models in order to solve both basic and applied scientific challenges, including understanding the COVID-19 replication mechanism, epigenetic language modelling for drug discovery, and in the development of clean energy. This enables researchers to test ideas that may otherwise languish for lack of resources and, ultimately, reduces the cost of curiosity.

In this keynote, Dr. Cédric Bourrasset, AI Distinguished Expert at Atos, will reveal how Atos pioneered the successful architecture, build, and delivery of large-scale AI infrastructures. He will present a live demonstration of Atos-driven technology to illustrate new AI-driven endpoints featuring GPU and IPU workflow capabilities, featuring a global customer case study to elaborate on the current complex challenges faced by designing and manufacturing large-scale AI computing platforms. He will also leverage over 15 years of personal experience in designing and manufacturing supercomputing systems.

The true potential of AI rests on super-human learning capacity, and on the ability to selectively draw on that learning. Both of these properties – scale and selectivity – challenge the design of AI computers and the tools used to program them. A rich pool of new ideas is emerging, driven by a new breed of computing company, according to Graphcore co-founder Simon Knowles. At the AI Hardware Summit, Phil Brown, VP Scaled Systems Product discusses the creation of the Intelligence Processing Unit (IPU) – a new type of processor, specifically designed for AI computation. He looks ahead, towards the development of AIs with super-human cognition, and explores the nature of computation systems needed to make powerful AI an economic everyday reality.

Cerebras Systems builds the fastest AI accelerators in the industry. In this talk we will review how the size and scope of massive natural language processing (NLP) presents fundamental challenges to legacy compute and to traditional cloud providers. We will explore the importance of guaranteed node to node latency in large clusters, how that can’t be achieved in the cloud, and how it prevents linear and even deterministic scaling. We will examine the complexity of distributing NLP models over hundreds or thousands of GPUs and show how quickly and easily a cluster of Cerebras CS-2s is set up, and how linear scaling can be achieved over millions of compute cores with Cerebras technology. And finally, we will show how innovative customers are using clusters of Cerebras CS-2s to train large language models in order to solve both basic and applied scientific challenges, including understanding the COVID-19 replication mechanism, epigenetic language modelling for drug discovery, and in the development of clean energy. This enables researchers to test ideas that may otherwise languish for lack of resources and, ultimately, reduces the cost of curiosity.  ​

AI Hardware Summit attendees are invited to attend the an extended networking session where they can meet attendees from across both events. The Meet & Greet is a perfect opportunity to reconnect with peers, expand your network, and discuss the state of ML across the cloud-edge continuum!

State-of-the-art large language models (LLMs) are empowering organizations to unlock the most critical insights in their unstructured data. Despite the opportunities, the cost and complexity of developing these models internally makes them impractical for most organizations to develop on their own. SambaNova overcomes these challenges with production-ready, pre-trained LLMs delivered through a full-stack solution, which can be further adapted through unlimited fine-tuning or pre-training within an organization’s own environment. In this workshop, we will showcase prototyping solutions for enterprise semantic search, legal compliance analysis, and call center service analysis, built on top of pretrained LLMs available through SambaNova Dataflow-as-a-Service. This is followed by live trials on the demo to showcase the exciting potential of large language models.

Developer workshops are restricted to machine learning practitioners from research institutions and enterprises who are interested in learning how to port code onto novel AI platforms and want to get hands-on access to hardware and SDKs. 

Workshops are application only and subject to eligibility and availability. The workshops are free, and lunch, shared networking sessions, and access to the Meet and Greet function and keynote is included in the developer pass. If you're a machine learning engineer / AI application developer, please apply using the form in the registration section of the website or by emailing [email protected]. There are approximately 30 spaces available.

During this workshop, attendees will be updated on the state of the art in computer vision use cases and learn how to build deep learning models for object detection, while improving model performance. Atos' expert host will provide best practices on team organization to facilitate success. Finally, attendees will learn how to implement an engineering strategy - build a project template & data versioning, experiment tracking & feedback, leverage testing & boost model promotion.

Developer workshops are restricted to machine learning practitioners from research institutions and enterprises who are interested in learning how to port code onto novel AI platforms and want to get hands-on access to hardware and SDKs. 

Workshops are application only and subject to eligibility and availability. The workshops are free, and lunch, shared networking sessions, and access to the Meet and Greet function and keynote is included in the developer pass. If you're a machine learning engineer / AI application developer, please apply using the form in the registration section of the website or by emailing [email protected]. There are approximately 30 spaces available.