www.nvidia.com
Open in
urlscan Pro
2.17.22.9
Public Scan
Submitted URL: https://go.nvidianews.com/MTU2LU9GTi03NDIAAAGRKZdfPLpqEQw8dl9qs6fUzUGqIOkWDDfg5Y5wEz6ywo4WSFKbXdmNcLCr7Ru9JKjJ0p8hFk0=
Effective URL: https://www.nvidia.com/gtc/sessions/large-language-models/?nvid=nv-int-bnr-463583
Submission: On February 10 via manual from BR — Scanned from DE
Effective URL: https://www.nvidia.com/gtc/sessions/large-language-models/?nvid=nv-int-bnr-463583
Submission: On February 10 via manual from BR — Scanned from DE
Form analysis
1 forms found in the DOMhttps://www.nvidia.com/en-us/search/
<form id="searchform" action="https://www.nvidia.com/en-us/search/" class="search-form " __bizdiag="-1902636770" __biza="WJ__">
<input data-search-box-input="" class="search-box-input placeholder" type="text" id="search-terms" name="q" value="" placeholder="Search GTC" autocomplete="off">
<input type="hidden" name="searchPath" id="search-path" value="/gtc/sessions/large-language-models/">
</form>
Text Content
Workshops March 17–21 | AI Conference and & Expo March 18–21 | Keynote March 18 | San Jose, CA and & Virtual * * * My Account * Log In LogOut * EN * EN * 한국어 * 日本語 * 繁中 * 简中 * DE * FR Keynote Session Catalog Agenda * Schedule at a Glance * Topics * Speakers * Workshops & Training * Connect With the Experts Attend * Why Attend * Pricing * Bring Your Teams Sponsor & Exhibitors More * Travel Info * FAQ * Code of Conduct * Inclusion * Privacy Policy * Contact Us * Keynote * Session Catalog * Agenda * Schedule at a Glance * Topics * Speakers * Workshops & Training * Connect With the Experts * Attend * Why Attend * Pricing * Bring Your Teams * Sponsor & Exhibitors * More * Travel Info * FAQ * Code of Conduct * Inclusion * Privacy Policy * Contact Us Log In Register Now Log In Register Register Now * Keynote * Session Catalog * Agenda * Agenda * Schedule at a Glance * Topics * Speakers * Workshops & Training * Connect With the Experts * Attend * Attend * Why Attend * Pricing * Bring Your Teams * Sponsor & Exhibitors * More * More * Travel Info * FAQ * Code of Conduct * Inclusion * Privacy Policy * Contact Us This site requires Javascript in order to view all its content. Please enable Javascript in order to access all the functionality of this web site. Here are the instructions how to enable JavaScript in your web browser. LARGE LANGUAGE MODELS CONFERENCE SESSIONS Explore the latest tools, optimizations, and best practices for large language models (LLMs). View Full Session Catalog FEATURED SESSIONS In-Person THE GOLDILOCKS APPROACH TO LLMS: BALANCING ACCURACY, LATENCY, AND COST FOR OPTIMAL PERFORMANCE [S62163] Elena Agostini | Senior Software Engineer | NVIDIA Nik Spirin | Director, Generative AI and LLMOps Platform | NVIDIA Janaki Vamaraju | Deep Learning Architect and Scientist | NVIDIA Generative AI and large language models (LLMs) are powerful tools for automating enterprise processes. While many organizations have started the process of evaluation and experimentation, there is a still a gap in being able to tune LLM applications for the optimal performance that allows them to be deployed in production. We address the complexity of solution design, guiding you from initial setup using a pre-trained foundation model to achieving state-of-the-art results through information retrieval and customization. We emphasize the importance of defining success criteria — accuracy, latency, and cost — before delving into customization techniques. We'll cover a range of strategies, including accelerated inference, prompt engineering, retrieval augmented generation (RAG), domain adaptation, and fine-tuning. You'll gain insights from real-world customer engagements, transformed into actionable recommendations. This talk is an essential guide for anyone looking to leverage the power of LLMs in their business. Add to Schedule Tuesday Mar 19 | 4:00 PM - 4:50 PM CET Show More In-Person FIRESIDE CHAT WITH KANJUN QIU AND BRYAN CATANZARO: BUILDING PRACTICAL AI AGENTS THAT REASON AND CODE AT SCALE [S62577] Bryan Catanzaro | Vice President, Applied Deep Learning Research | NVIDIA Kanjun Qiu | Chief Executive Officer and Co-Founder | Imbue Join Imbue CEO Kanjun Qiu and Bryan Catanzaro, vice president of applied deep learning research at NVIDIA, for a discussion on AI agents and the next chapter in AI innovation. Launched in October 2022, imbue recently announced a $210 million-plus Series B fundraiser at $1 billion valuation to create the next wave of AI: agents that can reason and code. Imbue’s research is focused on solving AI’s lack of reasoning abilities. Kanjun and Bryan will discuss the current barriers to building agents, how AI marks a revolution in human-computer interaction, and the future of the personal computer. Kanjun will also speak to Imbue’s recent research — including papers on the training process of large-scale SSL methods; CARBS, a hyperparameter optimizer that automatically reproduces Chinchilla scaling laws; and work analyzing a request for comments on AI policy from the U.S. Department of Commerce. Add to Schedule Wednesday Mar 20 | 5:00 PM - 5:50 PM CET Show More Virtual CUSTOMIZING FOUNDATION LARGE LANGUAGE MODELS IN DIVERSE LANGUAGES WITH NVIDIA NEMO [S62743] Miguel Martinez | Senior Deep Learning Data Scientist | NVIDIA Meriem Bendris | Senior Deep Learning Data Scientist | NVIDIA Dora Csillag | Senior Solutions Architect - GenAI&Inference | NVIDIA Sergio Perez Perez | Solution Architect | NVIDIA We'll focus on customizing foundation large language models (LLMs) for languages other than English. We'll go through techniques like prompt-engineering, prompt-tuning, parameter-efficient fine-tuning, and supervised instruction fine-tuning (SFT), enabling LLMs to adapt to diverse use cases. We'll showcase some of these techniques using NVIDIA NeMo Framework for both NVIDIA Foundation Models and other community models, such as Llama-2. Finally, we'll demonstrate how to efficiently deploy the customized models using NVIDIA TensorRT-LLM and NVIDIA Triton Inference Server. Add to Schedule Tuesday Mar 19 | 12:00 PM - 12:50 PM CET Show More In-Person NAVIGATING THE LARGE LANGUAGE MODELS FRONTIER: PRACTICAL STRATEGIES FOR BUILDING ENTERPRISE APPLICATIONS POWERED BY LLMS [S62752] Kari Briski | Vice President Generative AI Software Product Management | NVIDIA Farshad Saberi Movahed | Senior Strategic Alliances Manager - NLP | NVIDIA Harrison Chase | Chief Executive Officer | LangChain Jerry Liu | Chief Executive Officer | LlamaIndex Arvind Jain | Chief Executive Officer | Glean Our panel of experts will talk about the best practices for building robust large language model (LLM)-based enterprise applications that deliver value and efficiency. Products such as ChatGPT have demonstrated the unprecedented power of LLMs in processing information and generating content. But harnessing LLMs for building enterprise applications introduces a spectrum of intricate challenges. They include, but aren't limited to, managing the behavior of LLMs (e.g., avoiding hallucination), adapting LLMs to domain-specific tasks while pre-trained on very general domain corpora, interacting with agents to execute some specific tasks, latency, security, and so on. We'll explore how enterprises can address these challenges and exploit the full potential of LLMs for their applications. Add to Schedule Tuesday Mar 19 | 6:00 PM - 6:50 PM CET Show More In-Person WHAT’S NEXT IN GENERATIVE AI [S62430] Manuvir Das | Vice President of Enterprise Computing | NVIDIA Brad Lightcap | Chief Operating Officer | OpenAI First, we'll explore OpenAI’s perspective on the direction of the market, gaining insights into their vision for the future. Then we'll examine how these AI tools are currently being used in practical applications, providing real-world examples to illustrate their effectiveness. We'll also discuss the apps that are expected to emerge in the near future, revolutionizing industries and unlocking new possibilities. As we explore the impact of these tools, we'll also address key issues and societal implications that arise from their widespread adoption. Finally, we'll explore how OpenAI is operationalizing these AI tools at scale, uncovering the strategies and best practices employed to maximize their potential. Add to Schedule Tuesday Mar 19 | 5:00 PM - 5:50 PM CET Show More In-Person THE SMALL MODELS REVOLUTION [S61190] Sébastien Bubeck | Vice President | Microsoft GenAI Large language models (LLMs) have taken the field of AI by storm. But how large do they really need to be? I'll discuss the phi series of models from Microsoft Research, which exhibit many of the striking emergent properties of LLMs despite having a mere 1 billion parameters. Add to Schedule Tuesday Mar 19 | 5:00 PM - 5:50 PM CET Show More In-Person THE UNSOLVED CHALLENGES OF LLMS IN OPEN-ENDED WEB TASKS: A CASE STUDY [S62589] Nicolas Chapados | Vice President, Research | ServiceNow, Inc. Alexandre Lacoste | Staff Research Scientist | ServiceNow, Inc. We investigate the challenges associated with developing goal-driven AI agents capable of performing open-ended tasks in a web environment using zero-shot learning. Our primary focus is on harnessing the capabilities of large language models (LLMs) in the context of web navigation through HTML-based user interfaces. We evaluate the MiniWoB benchmark and show that it's a suitable yet challenging platform for assessing an agent's ability to comprehend and solve tasks without prior human demonstrations. Our main contribution encompasses a set of extensive experiments where we compare and contrast various agent design considerations, such as action space, observation space, and the choice of LLM, with the aim of shedding light on the bottlenecks and limitations of LLM-based zero-shot learning in this domain, in order to foster research in this area. Add to Schedule Wednesday Mar 20 | 12:00 AM - 12:50 AM CET Show More In-Person LARGE LANGUAGE MODELS: PAST, PRESENT, AND FUTURE [S62922] Thomas Scialom | Research Scientist | Meta This talk will discuss the recent history of LLMs, deep dive on Llama-2 RLHF, and share my vision of the future. Add to Schedule Wednesday Mar 20 | 11:00 PM - 11:25 PM CET Show More Engage * AI Startups * Connect With the Experts * Demos * DLI Training & Workshops Discover * Keynote * NVIDIA On-Demand * Sponsors More * Code of Conduct * FAQ * Inclusion * Privacy Policy * Contact Us Follow GTC NVIDIA United States * Privacy Policy * Manage My Privacy * Do Not Sell or Share My Data * Legal * Accessibility * Corporate Policies * Product Security * Contact Copyright © 2024 NVIDIA Corporation NVIDIA uses cookies to enable and improve the use of the website. Please see our Cookie Policy for more information. NVIDIA uses cookies to enable and improve the use of the website. GPC signal detected and only ‘Required’ cookies have been enabled. To update your communication preferences please visit the Preference Center. Please see our Cookie Policy for more information. Reject Cookies Accept Cookies Manage Cookies Cookie Settings NVIDIA websites use cookies to deliver and improve the visitor experience. Learn more about the cookies we use on our Cookie Policy page. Required Cookies These cookies are required for our sites to function and cannot be turned off. Performance Cookies Performance Cookies These cookies provide information to help us improve your web experience by monitoring the performance of our website and collecting anonymous data on how you use it. Advertising Cookies Advertising Cookies Set by our advertising partners, these cookies are used to build a profile of your interests and show you relevant ads on other sites. They do not store personal information, but are based on uniquely identifying your browser and internet device. * PERSONALIZATION COOKIES Switch Label label These cookies are used to better understand and optimize your web experience, such as pages visited or purchases made through our e-store. These cookies and the information they collect may be managed by other companies, and the information collected by these cookies may be used to build a profile of your interests and show you relevant advertising on other sites. They do not store direct personally identifiable information, but are based on uniquely identifying your browser and internet device. Cookie Details Back Button Cookie List Search Icon Filter Icon Clear checkbox label label Apply Cancel Consent Leg.Interest checkbox label label checkbox label label checkbox label label Decline All Save and Accept