New AI Product from Duke Professors Tackles Information Discovery and Synthesis

“AI is completely changing the way we consume information,” says Jon Reifschneider. “At some point in the future, the vast majority of information flowing to us will do so with the assistance of AI. So, the question then becomes: what role is it playing?”

Reifschneider has been working with artificial intelligence (AI) and machine learning (ML) since the early 2010s, when he used AI systems to help create bespoke weather forecasts for large enterprise clients – helping understand the predicted impact of weather on operations for, say, an electric utility worried about weather-related power outages.

But he has always had a passion for teaching, and in 2019 Reifschneider moved to the Pratt School of Engineering at Duke University, where he now runs the Artificial Intelligence for Product Innovation Master’s Program as well as the Center for Research & Engineering of AI Technology in Education (CREATE).

Now, Reifschneider is blending his interests in education and building AI-powered software. Along with colleague Adjunct Associate Professor Pramod Singh, Reifschneider and team are building AI capabilities to assist in learning and information discovery – and taking them to the market.

Their new AI-powered research platform Inquisite is now available to the public, and the team is excited for more users to try it. Anyone can sign up for a free tier that offers a taste of the functionality, and users registering with an official Duke email address receive a 50% discount for the paid tiers.

Helping Students with AI

“In Fall of 2020, I started getting really interested in how we can use AI to improve the learning experience by helping students get unstuck when they have questions,” says Reifschneider.

This was a couple years before ChatGPT was released publicly – if you can remember those times – and the natural language processing capabilities Reifschneider was working with were more rudimentary. Basically, the system he built allowed students to submit a query related to a course they were taking and the AI would help determine what part of that course material would be most beneficial for the student to review. 

As Large Language Models (LLMs) became increasingly more useful, Reifschneider evolved the early system into a much more powerful search technology that could pinpoint answers to detailed questions from large bodies of text such as multiple textbooks. Such systems offer the potential to assist students in answering questions without having to wait for the next office hour session with their TA, Reifschneider says.

“More and more instructors and students are at least trying these types of tools out, if not integrating them into their workflows now,” says Reifschneider of the adoption of AI in the education realm. “And I think that’s a trend that still is really early.”

Expanding into a New Field

Expanding from their initial focus in the education realm, the team started thinking about another more open-ended potential application of the technology: research.

“We realized that in some sense, research is kind of a similar problem but on a much larger scale,” says Reifschneider. “You have complex questions that you need to dig into, but unlike in the context of a course, the information to answer those questions may come from a wide variety of different sources – research papers, books, websites, and data. Searching through this vast ocean of information and distilling insights can be a time-consuming and very manual process.”

To help with this challenge, they built a system they now call Inquisite that helps researchers and knowledge workers search and synthesize information across many sources, including scholarly papers, technical and scientific reports, and trusted sources on the web.

Here’s how it works: A query is translated into a numerical representation called an embedding. This is cross referenced with a database of a few hundred million English-language research papers, reports, and websites that have been similarly embedded – this is called a semantic search. The system then uses proprietary algorithms the team developed to assign relevancy and trust scores to each source, sorts and filters them, then cleans up the results for display.

“Research often involves sifting through overwhelming amounts of information to find relevant and trustworthy insights,” says Singh. “With Inquisite, we’ve designed a system that streamlines this process by leveraging advanced techniques like semantic search and AI agents. Our goal is to help researchers focus on analysis and innovation, rather than spending time on manual searches and information sorting.”

The team has also added a text editor directly in the tool to streamline the process for researchers, along with an AI assistant that helps draft text based on the identified sources and find new sources when needed. This addition, as well as the more in-depth semantic searching and results scoring the system does is what Reifschneider says distinguishes Inquisite from other research AI tools like Perplexity and Elicit.

“Our goal with Inquisite is not to build a better version of Google, but rather to develop a tool that acts much more like a highly capable research assistant – helping you find and synthesize the best sources of information,” envisions Reifschneider.

The Winding Path of Commercialization

While friends and colleagues have been beta testing Inquisite, the team has also been preparing the technologies for wider deployment.

With support from Daniel Dardani, Director of Physical Sciences and Digital Innovations Licensing and Corporate Alliances at the Office for Translation & Commercialization (OTC), multiple potential paths for spinning out the technology were considered.

“We went back and forth for a while about setting it up – a nonprofit or a for profit company – and ended up with something in the middle, which is a called a public benefit corporation,” says Reifschneider. “Our goal has always been to build something useful, that can have a real, positive impact in the world by accelerating the process of learning and innovation.”

“It sounds cliché but impact matters just as much, if not more, than income when it comes to seeing Duke technology operate in society,” says Dardani. “That is why we work with our innovators on ways that makes sense for them to best bring their technologies to the public.” 

The Inquisite team has also been working with Duke New Ventures at OTC, partnering with seasoned tech industry executive and Mentor-in-Residence Carlos Pignataro, as they embark on the next leg of their commercialization journey.

“I am thrilled to be working with this team,” says Pignataro. “It’s exciting to partner and help them navigate customer discovery, refine their product-market fit, ask questions, and watch their product come to life – for researchers in academia and in business.”

Now that Inquisite is publicly available, the team hopes to learn even more from its users to guide them into the future. Users registering with an official Duke email address receive a 50% discount on paid tiers, but anyone can register for a free tier to try out the search tool.

“We would love for people to try it out,” says Reifschneider. “We love to get feedback from our users, who can reach us directly or leave feedback via the Inquisite website.”