How Semantic Searching Unlocks AI

Artificial Intelligence, InnovationQ Tags: AI, Artificial Intelligence, Intellectual Property, machine learning, Semantic Search, semantics

Although we are a long way (perhaps) from the Artificial Intelligence that Will Smith fights against in I, Robot, wielding some of its power is as close as your laptop. Just as people talk to “Sonny” in the film, you can use natural language to find answers to complex, multi-layered questions.

The Humanization of Searching
In the areas of patent searching and intellectual property discovery and management, it can take years to develop investigation expertise. Professional searchers know how to craft search strings using strategically selected keywords and Boolean operators. And then there is the metadata to consider, especially with patent-related content: tags, codes, dates, assignees, etc. Analyzing the desired outcome for the search and then synthesizing a query to produce the desired results is an advanced skill that allows people to “speak” to a search engine. Owners of these skills are often professional patent searchers, intellectual property lawyers, or members of specialized business units.

If you are an entrepreneur with a small company, an inventor with a handful of ideas, or an engineer within a large company hoping to push your innovative ideas through the patenting process, you need to do some searching to find out whether your idea is patentable and marketable. Sitting down to launch the perfect keyword or Boolean search is daunting and, frankly, unnatural as you dissect your own thoughts. What are the Boolean operators doing? How do you now if you are asking the right things? How do you know if you have missed anything important? Are you even looking in the right places?

But, thanks to advancements in database tools and search capabilities, IP analytics and management solutions are not just for patent attorneys and IP lawyers anymore. Natural Language Processing (NLP) technologies have enabled the development of semantic searches that “seek information based on the intent of the searcher and the context of the
search terms, rather than relying on simply matching the exact words used in the query”*. The semantic query suits a range of users, and with the right database and tools, can take ideation to precise answers in a matter of minutes.

Semantic search queries take the uncomfortable, robot-like query structuring process to a method of human input. You can really write the to the machine and tell it what you are looking for in your own words. You can even copy text from a document and enter that as a search query. The semantic engine derives the context, associates meanings, and identifies the documents that have the concepts most relevant to your needs. The output matches the idea that you input; much better for the discovery of innovations. It’s new stuff – there are different ways of expressing it, so you need a semantic search to recognize when a different expression represents a similar idea.

If you are already an expert searcher or have the benefit of a team of searchers at your disposal – think of how your power can be expanded with semantic searching. You can have keywords, plus Boolean, plus semantic searching. The opportunities for discovery are great.

And what about AI?
Artificial Intelligence comes from machine learning algorithms. In tools such as InnovationQ and InnovationQ Plus, machine learning algorithms combined with semantic-based algorithms help the searcher find the most relevant documents through natural language. Specific learning algorithms identify slight changes in meanings and conceptual relationships over time. The system knows how to navigate the intricacies of language depending on the context. It recognizes many meanings for one term or multiple words that can have the same meaning, and then accordingly matches the document in which they appear with the identified concept in the query. The algorithms here have a technical focus, adapted for patents and technical documents. In addition, the neural network learning algorithms provide the capacity to consume large blocks of text, so you don’t have to limit your query to a list of keywords or small number of characters – copy and paste pages into the query field if that is what you need to do.

Find hidden data faster when you need to make critical business decisionsBeginning with a semantic search supports specialized analytics capabilities. These search tools are not strictly automated. They identify key concepts, which come from your natural language queries and subsequent semantic analysis. From your results set, you can select a document to feed into another search. Filter the set by assignee, date, inventor, and more. You can perform visualizations to look at the scope of the results from multiple perspectives. Then, drill down or expand your search. The analytics features step you further into the discovery process that already had a head start with AI.

AI is generated when the system applies machine learning based on semantic input. The intelligence you gain, however, is far from artificial. It is applicable. Leverage it to make critical business decisions. Use it to prove your idea is not only new and useful, but also marketable. Show that it has its place and can move to monetization. Find other organizations that are doing similar work and partner with them or license your technology.

Why is semantic searching different? Because it is smarter. For real.

*To read about the power of semantic searching for intellectual property management and using InnovationQ Plus, access the free download for’s e-book, Increase Intelligence Around IP with Semantic Search.