The Numenta team will be part of Intel’s booth (booth #810) at VMware Explore Las Vegas on August 26-28, 2024. Our staff software engineer Jared Weiss will be presenting on August 28 at 3:30pm PT on revolutionizing AI deployment with scalable, private, CPU-based solutions. We will also be giving a demo of our AI platform NuPIC’s document retrieval solution at the booth throughout the week.
VMware Explore 2024 features expert-led business and technical sessions, an extensive ecosystem of the top cloud partners, a thriving marketplace of ISVs, and several networking events across the VMware community. Learn more here.
—-
PRESENTATION
Beyond GPUs: Revolutionizing AI Deployment with Scalable, Private, CPU-Based Solutions
While organizations are racing to implement AI into their offerings and throughout their business to stay competitive, the challenges of AI deployment remain too high a hurdle for many. AI projects can be highly complex and costly, difficult to scale, and largely dependent upon the availability of GPUs. Companies need something that removes the complexity and provides scalable, secure and cost-effective AI solutions to power a new generation of AI-enabled applications.
The Numenta Platform for Intelligent Computing (NuPIC™) is a unique neuroscience-based AI software platform that makes it easy to deploy large language models (LLMs) with extreme efficiency and complete privacy. Best of all, NuPIC can run entirely on CPUs, which gives incredible flexibility in resource utilization and allocation. When running on AMX-enabled Intel Xeon CPUs, NuPIC achieves higher throughput, lower latencies, and higher accuracies than LLMs on GPUs.
In this session, you will discover how NuPIC makes CPUs the optimal choice for running diverse AI workloads, enabling you to efficiently power your applications with unparalleled performance, scalability and privacy – no AI expertise required.
DEMO
From RAG to Rich Insights: AI-Driven Document Retrieval powered by NuPIC with Intel Xeon
As businesses scale, effectively managing an ever-growing repository of documents presents a significant challenge. As the volume of documents continues to increase, from emails and reports to legal documents and technical manuals, finding specific information quickly and accurately becomes increasingly difficult, impacting agent productivity and time spent on revenue-generating activities. Businesses need a robust solution that can navigate through unstructured data and extract relevant information in real-time. The Numenta Platform for Intelligent Computing (NuPIC™) is a unique AI software platform that uses neuroscience principles to process large amounts of language data quickly and accurately.
With easy and completely private deployment in VMWare, NuPIC can run large language models (LLMs) entirely on AMX-enabled Intel Xeon CPUs and achieve high throughput, low latencies, and higher accuracies than traditional LLMs on GPUs. In this demo, we walk through how a company can leverage NuPIC’s retrieval-augmented generation (RAG) solution to continually index a growing collection of documents, then build a question-answering system based on the knowledge base. We’ll highlight how NuPIC allows you to run various types of models in parallel on a single CPU, significantly reducing operational costs and making it easier to deploy AI applications at scale.
Jared Weiss • Staff Software Engineer