How do we measure success? By seeing our work applied.
InferLink's R&D activities, products, and solutions are motivated by the principle that technology is born when hard problems meet creative talent.
See what we can do for you.
From Entity Resolution and Social Media Analysis to full scale custom development, InferLink's team delivers. Take a look at our R&D capabilities below and let us know how we may help you.
Data sources often use slightly different terms to refer to the same thing. Our Entity Resolution solution can make highly accurate judgments to determine when two sources are referring to same real-world entity, including people, companies and products.
Our social media research capabilities include inferring meaningful profiles of individual social media users and groups, detecting and surfacing topics of conversations, and connecting the dots across relationships, even when those relationships are not obvious.
We develop novel statistical models and scalable computational algorithms for matching, clustering, topic analysis and other inference problems, based on many years of experience in machine learning and AI.
Our text mining research involves structuring highly variable text such as social chatter or online postings, extracting and cleaning entity mentions from text, and uncovering information about group memberships based on linguistic clues.
Our deep experience in data collection includes the ability to harvest data from Web sites, extract data from standard text (such as news), cull information from structured sources (such as databases or even Wikipedia), and make sense of extremely variable text such as social media chatter or online postings.
We develop technology for intelligently searching through large collections of semi-structured or unstructured data, using AI technologies to find the most relevant information.
Our technical team works closely with enterprise clients, applying AI technologies to solve challenging information extraction and integration problems.
Our products and solutions are built to work, whatever the challenge or complexity. We successfully adapt and apply these offerings directly to the customized needs of a wide range of organizations.
EntityBase aggregates, merges, and matches data about people, organizations, products, and other entities. Probabilistic matching technology seamlessly integrates data from multiple sources, even when those sources use different names, terminologies or data formats to describe the same entity. Machine learning automatically tunes matching algorithms for different types of data.
PSI4 is a link-analysis and social media comprehension platform that enables researchers and analysts to conduct investigations on the Internet. Analysts can interactively harvest data from public Web sites, analyze the content, and identify interesting patterns and links.
RSX extracts information from Web sites, using InferLink's next-generation, fully-automatic web scraping technology. The system can autonomously crawl and analyze a web site, in order to extract all the semi-structured data on the site, clean the data and deliver it in a convenient format. RSX also makes it simple to monitor a site and create a scheduled webfeed with any updated data from the site -- without any programming.
SocialPortrait provides actionable social-media insights, allowing companies to insert themselves into social conversations, market to finely tuned audiences and find channels to communicate with the right potential customers at just the right time.
ThreatRank provides corporate IT teams with relevant, real-time intelligence on computer security issues. This AI application scours the Internet for news, evaluations and mitigation strategies. ThreatRank's algorithms and state-of-the-art machine learning techniques prioritize, categorize, and polish raw data into actionable intelligence based on an organization’s unique needs.
ActiveSearch search is a semantic search system. Unlike most commercially-available systems, which are based on keyword matching, ActiveSearch employs a massive knowledge-graph of concepts to understand the documents it reads, so that users can find precisely the information they want.
Connectd is customizable, intelligent behavorial health platform for patients, providers, and caregivers. Combining active and passive information gathering techniques, Connectd uses machine learning to automatically adapt to user needs, interests, and goals. Connectd is rapidly portable to a wide variety of applications and domains, from veterans with PTSD to foster families to large health systems focused on reducing physician burnout.
Pedro Szekely, Craig A. Knoblock, Jason Slepicka, Andrew Philpot, Amandeep Singh, et al.
Brian Amanatullah, Greg Barish, Matthew Michelson and Steven Minton
Steven Minton, Sofus A. Macskassy, Peter M. LaMonica, Kane See, Craig A. Knoblock, Greg Barish, Matthew Michelson, Raymond A. Liuzzi
Greg Barish, Patricia Lester MD, William Saltzman, and Eric Elbogen
Matthew Michelson, Sofus A. Macskassy,
Steven N. Minton, and Lise Getoor
Matthew Michelson and Sofus A. Macskassy
Matthew Michelson and Craig A. Knoblock
Steven N. Minton, Claude Nanjo, Craig A. Knoblock, Martin Michalowski, and Matthew Michelson
Craig A. Knoblock, José Luis Ambite, Kavita Ganesan, Maria Muslea, Steven Minton, Greg Barish, Evan Gamble, Claude Nanjo, Kane See, Cyrus Shahabi, Ching-Chien Chen
Greg Barish, Craig A. Knoblock
Zachary G. Ives, Craig A. Knoblock, Steven Minton, Marie Jacob, Partha Pratim Talukdar, Rattapoom Tuchinda, José Luis Ambite, Maria Muslea, Cenk Gazen