Natural Language Search for Solr and Elasticsearch
Details
Why don't search engines just understand? Open source search engines, while effective at processing and retrieving text at scale, are otherwise naïve and have no practical way to parse or understand human language.
This week, our speaker Max Irwin will take you through a practical tour of open source NLP tools that can be readily applied to improving search quality, all-the-while keeping your existing Solr or Elasticsearch backed applications.
Beginning with a problem statement of language problems and pain points with Lucene-based search engines, Max will talk about how to relieve them with open source tooling and techniques, including:
- Precise autocomplete via knowledge graph extraction
- Query intent detection
- Improving BM25's effectiveness with coreference resolution
- Lemmatization for fixing the woes of stemming
- Linguistic Payloading for additional relevance signals
As a bonus, Max will showcase these applications made available in a new open source search service, which can be used as a drop-in NLP augmentation for your existing Solr or Elasticsearch stacks.