We consistently see avoiding integrating machine learning with the whole org is at the root of machine learning adoption problems.
Haystack Is Back! Weekly online Haystack LIVE talks, teaming up with Berlin Buzzwords and MICES for an awesome conference week and a whole raft of online training now available
To build great search, you need a great search team. Here's how you can define what roles you'll need, collaborate effectively and some pointers on how to find the right people.
Search is important, but evaluating how relevant your search is can be tricky, if you don't have meaningful metrics. These are 5 ways to quantify search, so you can understand and track how your search is performing.
Get certified to use Machine Learning to improve Elasticsearch search relevance, from the team pioneering Information Retrieval with Elasticsearch.
Learn to improve search quality from the best experts in the field. From TFIDF, to Semantic Search, to Learning to Rank with Solr!
Learn to improve search quality from the best experts in the field. From TFIDF, to Semantic Search, to Learning to Rank with Elasticsearch!
Why you need a search strategy as part of your business strategy and why effective leadership is important to instil the right approach across your whole organization.
Think Like a Relevance Engineer training for Solr & Elasticsearch now available online!
How we're adapting to continue empowering search teams with online consultancy & training
Does your pet feature actually help users? Feedback debt, like technical debt, eventually needs to be repaid.
Eric Pugh talking about "Streaming Expressions in Solr Revolutionizes How You Think About Your Data" and Nathan Day on using statistics to efficiently bootstrap your collection of human judgements.
Now that you are using Quepid to collaborate with members of your team by sharing your cases with other folks, you need a way to view basic information for *ALL* of your owned and shared cases and navigate to those cases within Quepid. That is why we have the Multi Case Dashboard.
Understand what the common evaluation metrics are telling you about the relevance of search results, so you can pick the best fit for your search situation.
Announcing Search Insights 2020, a new free report from The Search Network covering many aspects of search and relevancy, with several contributions from the experts at OSC.
Elasticsearch Learning to Rank - Search as an Machine Learning Problem at the Triangle Search Technology Meetup
We will discuss how search can be treated as a machine learning problem. 'Learning to Rank' takes the step to returning optimized results to users based on patterns in usage behavior. We will talk through where Learning to Rank has shined, as well as the limitations of a machine learning-based solution to improve search relevance.
Migrating search engines? Want to prove the migration succeeded without worrying about lost customers, revenue, and the like. With Quepid, you can take a snapshot of the state of your products existing search results with a small amount of work. Then you'll always have this snapshot to compare your new search solution to.
In which we learn all the different competing ways of deploying Tika as a HTTP service!
While Quepid has been open source for six months, this is the first release that you can deploy yourself!
Deploying Solr in complicated multi-region environments & Search at BBC Online at the London Lucene/Solr Meetup
Konstantin Perikov of EPAM with a review of different ways to deploy Solr in complicated multi region environments and Andy Webb on Search at BBC Online
Human beings search more and more. Continuing to treat search as some mystical voodoo is untenable and must stop.
Will my talk get chosen? Will attendees find it interesting? A chance to answer some of the questions you might have about submitting a talk to Haystack, the Search Relevance Conference.
Traditional search engines and word2vec have their limitations for search relevance. Pt 2 of our series on BERT.
In which we take advantage of Solr's scripting update capability to turn Solr into an OCRing blackbox!
A relevant search result helps users take further steps, make good decisions, and supports their ability to formulate better, more informed, queries.
In which we deal with learning that sometimes you don't get to use the latest version of Tesseract...
In this post, we unwrap the mystery behind two popular search relevance metrics, and discuss their pros and cons. Our subjects for this exercise are Normalized Discounted Cumulative Gain, and Expected Reciprocal Rank, commonly acronymified as nDCG and ERR. We'll start with some refresher background, visualize what these metrics actually look like, and paint a picture of how each can be either helpful or misleading, depending on the situation. Afterwards, you'll have a better understanding of their behavior and which ones to use when (and why).
How to convert the _default managed schema in Solr v8.3.0 to the classic schema of yesterday
We explore what the basic unit of your search system should be. What *should* your document be? What 'result' in the SERP UI best satisfies your users?
In which we deploy Tika and Tesseract as a API in Solr, exposed via the /extract handler!
In which we explore how to deploy Tika and Tesseract as a stand alone service. And tie it into Solr via Payloads...
It's time for Tika Tuesdays! Three years ago I started messing around with OCRing documents with Tika, and today that process is relatively straightforward. This weekly series will share what I've learned.
My notes taken during the Haystack EU 2018 in Berlin, Germany, on October 28th.