BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/Chicago BEGIN:DAYLIGHT DTSTART:20250309T030000 TZOFFSETFROM:-0600 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:CDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20251102T010000 TZOFFSETFROM:-0500 TZOFFSETTO:-0600 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:CST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20250312T200934Z UID:CFD2E7A3-6BFC-44BF-BA16-7BA827280528 DTSTART;TZID=America/Chicago:20250312T120000 DTEND;TZID=America/Chicago:20250312T130000 DESCRIPTION:Garrett Hall\, a Research Engineer at the Southwest Research In stitute\, will deliver an introductory presentation on Large Language Mode ls. This talk is the first in a two-part series and covers several fundame ntal concepts\, including tokenization\, vector embeddings\, and positiona l encoding.\n\nTokenization is the process of converting words or phrases into numerical values that machine learning models can understand. By brea king down text into smaller units called tokens\, the model can more effec tively process and analyze the data.\n\nVector embeddings are a crucial ne xt step\, where these tokens are transformed into dense vector representat ions. These vectors capture semantic meaning\, enabling the model to under stand relationships between words based on their contextual usage. Embeddi ngs essentially map tokens into high-dimensional space where similar words are located closer together.\n\nPositional encoding provides additional i nformation about the order of the words in a sentence\, establishing a fou ndation for sentence structure. It embeds positional information within th e tokenized data so that the model can recognize the sequence and context of words\, which is essential for understanding the meaning of the text as a whole.\n\nFinally\, the presentation will illustrate Retrieval-Augmente d Generation (RAG) processes. RAG combines retrieval-based and generative models to enhance the generation of relevant and accurate text by incorpor ating external information sources. This section will demonstrate how the preceding concepts of tokenization\, embeddings\, and positional encoding come together in RAG to create more coherent and contextually appropriate text.\n\nCookies and refreshments will be served.\n\nTalk is restricted to US citizens.\n\nRegistration required by COB Monday 3/10 for admittance t o SwRI grounds on day of event.\n\nSpeaker(s): Garrett\n\nBldg: Building 5 1\, 6220 Culebra Rd\, San Antonio\, Texas\, United States\, 78238 LOCATION:Bldg: Building 51\, 6220 Culebra Rd\, San Antonio\, Texas\, United States\, 78238 ORGANIZER:garrett.hall@swri.org SEQUENCE:35 SUMMARY:IEEE AESS - Part I: LLM Basics URL;VALUE=URI:https://events.vtools.ieee.org/m/472758 X-ALT-DESC:Description: <br /><p class="mb-2 whitespace-pre-wrap">Garrett H all\, a Research Engineer at the Southwest Research Institute\, will deliv er an introductory presentation on Large Language Models. This talk is the first in a two-part series and covers several fundamental concepts\, incl uding tokenization\, vector embeddings\, and positional encoding.</p>\n<p class="mb-2 whitespace-pre-wrap">Tokenization is the process of converting words or phrases into numerical values that machine learning models can u nderstand. By breaking down text into smaller units called tokens\, the mo del can more effectively process and analyze the data.</p>\n<p class="mb-2 whitespace-pre-wrap">Vector embeddings are a crucial next step\, where th ese tokens are transformed into dense vector representations. These vector s capture semantic meaning\, enabling the model to understand relationship s between words based on their contextual usage. Embeddings essentially ma p tokens into high-dimensional space where similar words are located close r together.</p>\n<p class="mb-2 whitespace-pre-wrap">Positional encoding p rovides additional information about the order of the words in a sentence\ , establishing a foundation for sentence structure. It embeds positional i nformation within the tokenized data so that the model can recognize the s equence and context of words\, which is essential for understanding the me aning of the text as a whole.</p>\n<p class="mb-2 whitespace-pre-wrap">Fin ally\, the presentation will illustrate Retrieval-Augmented Generation (RA G) processes. RAG combines retrieval-based and generative models to enhanc e the generation of relevant and accurate text by incorporating external i nformation sources. This section will demonstrate how the preceding concep ts of tokenization\, embeddings\, and positional encoding come together in RAG to create more coherent and contextually appropriate text.</p>\n<p cl ass="mb-2 whitespace-pre-wrap">&nbsp\;</p>\n<p class="mb-2 whitespace-pre- wrap">Cookies and refreshments will be served.</p>\n<p class="mb-2 whitesp ace-pre-wrap">Talk is restricted to US citizens.</p>\n<p class="mb-2 white space-pre-wrap">Registration required by COB Monday 3/10 for admittance to SwRI grounds on day of event.</p> END:VEVENT END:VCALENDAR