It’s easy to think that Large Language Models make the search engines of old obsolete—why would you want to present users with a list of possibly applicable results, if you can prepare a tailored-made response, answering the user directly, precisely and confidently? besides rather frequently incorrectly. How do LLMs of present stack up against the current known approaches, like keyword or semantic search? Can we have the best of both worlds? During this session, we’ll dig into the performance of LLMs, erstwhile answering knowledge-based questions. We will besides research any methods of leveraging the benefits of all approaches, to truly get what the user asked for.
GeeCON 2024: Zbyszko Papierski - LLM — Zabójca wyszukiwarek?
It’s easy to think that Large Language Models make the search engines of old obsolete—why would you want to present users with a list of possibly applicable results, if you can prepare a tailored-made response, answering the user directly, precisely and confidently? besides rather frequently incorrectly. How do LLMs of present stack up against the current known approaches, like keyword or semantic search? Can we have the best of both worlds? During this session, we’ll dig into the performance of LLMs, erstwhile answering knowledge-based questions. We will besides research any methods of leveraging the benefits of all approaches, to truly get what the user asked for.
