In this talk Phil Steitz, Nextiva's CTO, digs into the technology underlying Neural Search and illustrates the concepts with working examples using freely available open source components.
First generation search technologies were focused on matching keywords in search queries with words and phrases in documents. Sophisticated algorithms for indexing, filtering, matching synonyms, and applying other rules-based heuristics in search pipelines were later developed, and systems such as Apache Lucene, Solr, and Elasticsearch power most of the text-based search in web applications today. In the last few years, advances in AI have led to the evolution of a new approach to search, which some call neural search.
Neural search has three big advantages over the first generation technologies. First, it does not require explicit programming. Second, instead of focusing on keywords and fixed rules, it operates on representations that capture more context and meaning. Third, it can be used to search any kind of digital media. Neural search can enable queries in one media to drive search in a different one or queries in one language to drive search across documents in a different language.
Follow Nextiva Technology on Eventbrite and be the first to signup for the next technical discussions: → [ Ссылка ]
Subscribe to Nextiva for more videos like this one: [ Ссылка ]
Stay connected with Nextiva:
YouTube ([ Ссылка ])
Facebook ([ Ссылка ])
Instagram ([ Ссылка ])
LinkedIn ([ Ссылка ])
Pinterest ([ Ссылка ])
Twitter ([ Ссылка ])
#programming #artificialintelligence #machinelearning
![](https://s2.save4k.ru/pic/OIsgoC9uWDc/maxresdefault.jpg)