Under the theme, "Search Capabilities", the presentations and discussions explored how artificial intelligence (AI) can be used to enhance the relevance of search tools and how search has developed to focus more on user experiences.
Among the speakers, LexisNexis South Africa’s DevOps Team Lead, Morné Olderwagen, outlined the journey LexisNexis has been on to ensure that its legal search results meet client needs for relevance, with the aid of AI and natural language processing (NLP).
“LexisNexis works with diverse content sets and all of this legal information requires commentary from professional authors. Each of these content sets contains its own set of related metadata, such as a case law document with case name, case number, the judges involved, the parties, what sort of topics were covered and then the entire judgment. People search this for their research when trying to plan for a case in which their client is involved,” Olderwagen explained.
With separate types of content within one research platform, the LexisNexis tech team manipulates relevance using field boosting and functional scores to ensure that clients are able to find their documents quickly and efficiently. Olderwagen said LexisNexis also supplies specific, focused search forms for clients to search against individual fields, with the related meta data fields available for individual match queries.
He said to measure the performance of its tech development which is constantly evolving, LexisNexis prefers to use a normalised discounted cumulative gain (NDCG) as its measurement system to evaluate results sets before and after changes.
Olderwagen said the LexisNexis tech development team is exploring the use of a third-party plug-in called Elasticsearch Learning to Rank to manipulate the order in which search results are returned.
“This utilises error correction to get a result set that will be better for your customer and finetunes it to that effect,” he said.
Intuitive search functionality is critical to a number of leading LexisNexis solutions, such as LexisLibrary
for legal research and Lexis Practical Guidance
for the “how-to” of legal practice. LexisNexis is deploying AI, including NLP, visualisations and robot process automation, to draw knowledge, actions and conclusions, while enriching the data to serve clients and make the practice of law more efficient.
Also on the line-up at the meetup was Elasticsearch Engineer at Knowledge Focus, Neil Maderthaner, who provided practical scenarios and demonstrations of Painless, the simple, secure scripting language designed specifically for use with Elasticsearch. Maderthaner has over 20 years of experience working with almost all the prominent search-based products in the industry and is one of the first certified Elastic Engineers in South Africa.
In addition, development engineers Thomas Parry and Nicholas Waring from tech development giant, Derivco, gave a joint presentation on F5 and Elasticsearch and how this is used to spot hackers.
The Elastic Durban Meetup was sponsored by LexisNexis which has been using Elasticsearch as its Enterprise Search platform for over a decade under the leadership of Olderwagen in collaboration with Knowledge Focus, after migrating from the FAST ESP platform.