Loop AI Labs: Artificial Intelligence to Interpret Unstructured Data

With the enormous growth of unstructured data like e-mails, customer requests or transcriptions, enterprises often lack the tools to analyze and extract the relevant value from the information. Though common structuring techniques such as manual tagging with metadata or text mining are being heavily used, companies are still faced with numerous difficulties while extracting information, particularly when the text data is in more than one language. “Bridging the gap between unstructured data and insights, we offer services that model organizations’ data and build an Application Programming Interface (API), which enables them to understand and exploit the unstructured text in any language,” begins Bart Peintner, CTO, Loop AI Labs. Founded in 2012, the company offers the Loop Cognitive Computing Platform, a machine intelligence technology that automatically interprets natural language and extracts structured representations of the concepts and relationships contained in small or large amount of texts. The company’s APIs can be deployed to work on the cloud or using a physical appliance on the site of the company’s own data center.


Any kind of text data can be injected in the APIs, and a structured representation can be extracted from it automatically


Developers can organize any kind of unstructured text by integrating the model produced by the Loop Cognitive Computing Platform into their applications, using a collection of simple APIs. Using these APIs, Loop helps developers leverage any kind of unstructured text without data science or artificial intelligence experience, making concept extraction a process of simple API calls. The Person Digital GenomeAPI enables applications to instantly derive a deep profile of a user, including interests, life context, and personality insights. The deep profile helps businesses understand their clients' preferences and create the next generation of personalized applications.“The text data can be built in the form of a model to understand the conceptual content, even from the smallest text data
Gianmauro Calafiore, CEO
within a domain,” illustrates Peintner. The Semantic API estimates the degree of semantic relatedness between a base document and a set of target documents based on their digital genomes. It returns a list of target documents ordered by their degree of semantic relatedness to the base document. The Thing Genome API analyzes a company’s internal documents to encode hundreds of traits about the important aspects behind the business—products, records, emails, customer support logs or transcriptions, conversations, and more. “Any kind of text data can be injected in the APIs, and a structured representation can be extracted from it automatically,” informs Peintner.

One of Loop’s clients, a customer support company, was facing issues in responding to the large number of technical support requests that required domain-specific expertise to understand. Loop AI Labs trained a model based on many customer support interactions and used it to list the first sentence from each request and its relatedness to the characteristics of the analysts. In this way, the company was able skip the first level of the problem, giving experts the chance to understand their queries more clearly and to respond more effectively, providing a better customer experience. “Our technology provides a customized solution to companies’ specific needs,” illustrates Patrick Ehlen, Chief Scientist, Loop AI Labs.

Loop AI Labs is keeping up with technical advancements in the academic research community and continues to refine the technology for their business. “We are an artificial intelligence research organization focused on deep learning, trying to make a system that can understand human language,” concludes Ehlen.

Company
Loop AI Labs

Headquarters
San Francisco, CA

Management
Gianmauro Calafiore, CEO and Bart Peintner, CTO, Patrick Ehlen, Chief Scientist

Description
Working to radically change how machines can autonomously learn and understand the human world, mirroring the same learning process that humans use