Transform Your Data Into Actionable Insights: The Pulse of AI Interviews NTENT Chief Technology Officer, Dr. Ricardo Baeza-Yates
In an in-depth interview with podcast host, Jason Stoughton from C-Suite Radio’s “The Pulse of AI: The Business Leader’s Guide to AI and Robotics,” NTENT CTO, Dr. Ricardo Baeza-Yates, discusses the evolution of search, and offers insight to C-Suite leaders on how to best use current and future search technology within their businesses.
Dr. Baeza-Yates describes the way search has changed since the 1960s and how the Web introduced a new set of elements that were never there before, such as links. This created a structure between documents and a distributed database containing text from all over the world with an influx of users, generating a high volume of queries that had to be answered.
By 2010 and still today, the core challenge remains including more semantics, taking into account user context, language and intent, in order to provide users what they really want. This becomes progressively difficult as more people use the Web for more than just search, but for navigational and transactional queries as well. When asked by Jason Stoughton how to push the envelope further, Dr. Baeza Yates suggests a combination of better Natural Language Processing (NLP), quality data analytics that convey user context and feedback, and semantic knowledge that’s independent of language in general, but can be incorporated with its lexicon. Beyond this, Dr. Baeza-Yates predicts the next phase of search will be implicit search, where smart assistants will chronically monitor user behavior behind the scenes and imply suggestions related to that behavior in a non-intrusive way.
How can search help CEOs?
In response to Mr. Stoughton’s question regarding how CEOs can use search now to improve their enterprise, Dr. Baeza-Yates encouraged them to make search ubiquitous, allowing access to all company data for searchable information. However, this becomes incredibly difficult to manage with respect to privacy and security factors. He suggests using data mining to extract repeating patterns from company web logs, CRMs and helpdesk logs that could potentially solve some areas of a company’s problem. Data mining differs from search in that, “In search you know what you’re looking for,” said Dr. Baeza-Yates. “In data mining, you are searching for answers, and those answers will be transformed later by an expert to a particular question.”
When asked where Data Science falls into the mix of search technology, Dr. Baeza-Yates replies that it’s very broad. “It’s the science of handling data,” he said, listing search, data mining, NLP and elements of Machine Learning all as parts of Data Science. “We’ll say that Data Science today is everywhere. And the best companies are the ones that can integrate all their data into one knowledge pool, and extract information from it.” He offers an example of how implicit search could take this further, allowing an intelligent agent to monitor employee performance to help increase productivity or solve company problems.
In answer to Mr. Stoughton’s question as to whether search is relevant for all industries, Dr. Baeza-Yates emphasized its importance to industries that handle information such as insurance companies, where the ability to analyze internal data could be of precious value.
Where will search be three to five years from today?
Dr. Baeza-Yates anticipates search to play a big role with respect to voice interaction, implicit search and actionable search. Intelligent voice assistants will usually begin any response to a request with a search for information. Implicit search will continue to improve, with agents studying user behavior, ready to respond in a meaningful way. For example, a modern car that automatically searches and suggests nearby gas stations as you approach an empty tank. Lastly, Dr. Baeza-Yates predicts search will become more actionable, allowing users to apply parameters to their queries at the start. For example, users will be able to request travel plans for a trip to a certain city, on a certain date with given financial or hotel restrictions, and the smart assistant will return a complete itinerary.
How does NTENT’s general ontology work?
NTENT’s general ontology is a structured knowledge base that connects concepts, but also includes an onomasticon that differentiates entities from one language or context to another, that’s then added to the lexicon for a given language. Dr. Baeza-Yates describes building an ontology as a “slow process,” in which bigger isn’t necessarily better. A large ontology consisting of useless data only adds noise. That’s why NTENT uses a data-driven approach that focuses on what matters most to its users, a strategy that limits ambiguity born from words or phrases that have too many meanings.
Click here for the full podcast interview.
About The Pulse of AI Podcast: Veteran technologist, AI enthusiast, author and entrepreneur, Jason Stoughton, takes the Pulse of AI by interviewing the business visionaries, technologists, futurists, C-Suite leaders, venture capitalists and entrepreneurs driving the AI revolution. The podcast is targeted at business leaders and senior executives in the C-Suite who are interested in discovering how Artificial Intelligence and robotics can be applied and leveraged to gain a competitive advantage in today’s marketplace.
About NTENT: NTENT™ sits at the crossroads of semantic search and natural language processing technologies. Our patented, proprietary technology powers our comprehensive platform that transforms structured and unstructured data into relevant and actionable insights. This level of intelligence enables us to predict and deliver relevant information based on user intention. Learn more about NTENT at http://www.ntent.com.