Over the last few months, we’ve been asking ROSS users how they search for cases within jurisdictions and date ranges. We listened carefully, gained a lot of new insights and decided to overhaul our approach to jurisdiction and date filters. The new design is easier to use, gives you more control over how you use filters and provides better information on which filters are in use.
Before you begin your search the new design presents a jurisdiction and date filter drop-down menu above the query bar. When you select the dropdown, ROSS displays state and federal jurisdictions with corresponding checkboxes. You check the boxes to limit your query by jurisdiction and date range. The jurisdictions and date range you select will be continuously displayed above the query bar throughout your search.
ROSS also uses named entity recognition technology to analyze your AI search and set jurisdiction and date filters automatically. You’ll know when ROSS is setting filters because the system will underline any word it recognizes in your natural language query as a jurisdiction or date.
ROSS will add the jurisdictions and dates it detects in your query to your jurisdictional filters even if you haven’t checked them manually but will never subtract a checked jurisdiction. For example, if you check Delaware and then enter a query requesting cases in the Second Circuit, ROSS won’t drop Delaware even though it’s not within the Second Circuit.
We’ve re-designed the filters so that automatic and manual filters work better together and are more intuitive to use alone and in combination. We think you’ll love the new filters. If you don’t, let us know.
ROSS is built on state of the art natural language processing (NLP). NLP understands the context of legal queries rather than simply matching keywords or boolean strings. NLP generally delivers superior search results. But since ROSS NLP analyzes context, responsive cases may contain synonyms or stemming of root words that do not precisely match the terms in the query. Lawyers have told us that for some searches they prefer to limit responsive passages to those containing a specific phrase, word or legal term-of-art. As always, we listened carefully to their feedback and added a new feature that uses quotation marks to return only cases with exact query matches.
ROSS users can now use double quotation marks within a query to retrieve exact matches to the query language. For example, if a query seeks cases relevant to the standard for patentability of “software based business methods”, ROSS will only return cases containing the exact root words within the quotation marks. Use of quotation marks might be helpful in that situation because ROSS will not otherwise limit its relevance analysis to cases using the exact query language, e.g. it may also provide cases relevant to patentability and business methods that are indirectly related to software. When you want to hone in on a precise phrase, quotation marks may be the right search strategy.
Some finer points to keep in mind when using the exact match feature:
Users told us that they would like the ability to expressly exclude specific words or phrases from the ROSS natural language analysis powered by AI. So we’re introducing an exclusion filter to our search bar. Simply select the exclusion filter and type in the words or phrases you want to exclude. ROSS won’t return or rank any cases that contain those words. In tests lawyers have reported that the exclusion function gives them an additional level of control and results in more precise queries.
For example, there are many Lanham Act cases that discuss false advertising claims in the context of comparative pharmaceutical advertising. Since pharmaceutical advertising is governed by criteria that are not relevant to other kinds of advertising, pharmaceutical cases can sometimes clutter false advertising search results. Using the exclusion filter to eliminate cases using the word pharmaceutical will sharpen some search results.
The exclusion filter won’t affect our forthcoming Boolean capabilities. More on ROSS and boolean in a later post.
Join the leading minds in legal tech and get the latest news and updates on the law, legal technology, artificial intelligence, and much more.
