Roland von Borstel, Director of Discovery Solutions and Danny Chan, Senior Discovery Engineer, are headed to Relativity Fest 2018 to present the session "Activating Active Learning” on Monday, October 1 at 11 a.m. CT. We sat down with them to learn more about their session, the benefits of using Active Learning, and what they’re looking forward to at the conference.
Q1: Your upcoming session at Relativity Fest is on “Activating Active Learning”. Could you give us an overview of what this technology does and why it’s so beneficial?
Roland: Active Learning is Artificial Intelligence (AI) technology that makes it easier to identify relevant documents in a review pool by utilizing coding decisions, from which it is constantly learning.
Danny: For people new to Active Learning, I compare it to Spotify’s or Netflix’s intelligence that recommends the songs or movies you’d be most interested in. The technology pushes the most “interesting” documents to the reviewer based on his or her coding decisions. Documents are classified by relevancy, and the system generally serves up only those that are applicable.
Q2: Active Learning is part of Ricoh eDiscovery’s Intelligent Review offering. What has changed since incorporating this technology into your review strategy?
Roland: Yes, Active Learning is one of the tools we now use for Intelligent Review, and in most cases has enabled our Review Solutions Team to finish their review much faster than before. One thing to note is that our Intelligent Review solution has never been a linear review of any document that comes into the database. Intelligent Review is a unique process that uses analytics tools, like Active Learning, to filter out potentially irrelevant documents and identify the next best document for review to our specialized reviewers. The entire review is accelerated as a result, which means significant cost and time savings for our clients.
Danny: Historically, a typical linear “eyes-on” review would have reviewers going through each document one at a time flagging the relevant ones. Through our expertise and knowledge of the latest, best-in-class analytics tools, Intelligent Review is designed and customized to reduce the need for "eyes-on" review. By incorporating Active Learning into the Intelligent Review process, we can send reviewers the most responsive documents in even less time. We’re currently at about a 97 percent relevance rate during review with Active Learning, which our clients have been finding very beneficial.
Q3: Could you share a real-world example of a recent review that realized these benefits?
Roland: We recently completed a large, complex review of half a million documents. We identified that we could use roughly 43 percent (about 145,000) of those documents for Active Learning, during which we observed a relevancy return rate of 95 percent. Of those 145,000 documents, we only needed to review 1/3 of them based on the prioritized review process, which led to significant overall time and cost savings for the client.
Q4: Active Learning technology is not entirely new, but continues to experience some market resistance. What would you say is different about Relativity’s version of this technology?
Roland: It's true that Relativity did not invent Active Learning technology, but they have worked hard to perfect it. Through seamless integration into the already existing, and highly successful, Relativity framework, end-users can operate Active Learning without the requirement of significant training. Relativity has also made it easy for administrators to setup and monitor their cases. It certainly takes time and experience to optimize workflows around Active Learning, but even first-time users are seeing incredibly positive results.
Danny: We recently received feedback from a first-time Active Learning reviewer who was impressed by how substantive the responsive documents were right from the beginning of the project. These highly meaningful results are due to Relativity's proprietary algorithm that learns quickly based on a reviewer's initial coding to identify relevant documents from the start. In our experience, it's the fastest, most user-friendly Active Learning platform on the market.
Q5: What do you hope attendees will take away from your session?
Roland: I hope they will see that Active Learning is an effective tool that can be used to significantly reduce cost and time in their review process. But, they should also understand that Active Learning itself won’t solve everything, and that it’s the people behind the technology that make it incredibly useful and powerful.
Danny: I hope they come away with an understanding of how incorporating Active Learning into their review process can result in more timely and cost-effective reviews, while avoiding the pitfalls that can occur in traditional, linear, "eyes-on" reviews.
Q6: What are you looking forward to most at Relativity Fest?
Roland: I’m really looking forward to connecting with our Relativity users and clients. Relativity Fest is always a great way to meet as a community and learn what is coming up next. The keynote is always fascinating too.
Danny: I’m looking forward to reconnecting with my contacts and spreading the word about Active Learning. I’m also looking forward to Andrew Sieja’s keynote. It’s always very insightful into what’s coming down the road for Relativity.
Don't miss Roland and Danny's session "Activating Active Learning" at Relativity Fest 2018. For more information on the conference, visit: relativityfest.com.
You may also be interested in...
Five Reasons Active Learning is Better than TAR
We have had the privilege of piloting Relativity Active Learning advanced analytics in Canada and believe it is going to transform document review. These are our initial findings.
Our Flawed Love of Keyword Searches
Lawyers often think they’re experts at crafting keyword search terms to reduce the volume of documents for manual review. But as humans, they don't know what they don't know.