Raytheon Intelligence & Space, a Raytheon Technologies business, is leading the IARPA Better Extraction from Text Toward Enhanced Retrieval, or BETTER, program, the Arlington, VA-based company announced May 25. The end-to-end system, Cross-Lingual Extraction for Versatile and Effective Retrieval, CLEVER, will improve the information discovery cycle and customize English-speaker search results based on user needs and the semantic context.
Given an English-only query supplied by the user, the CLEVER system will retrieve relevant documents from vast amounts of foreign language text, and extract mission-relevant information such as events in the form of who did what to whom when and where.
“Analysts are bombarded with massive amounts of information daily; they do not have time to read lengthy content while facing rapidly changing information needs,” said Bonan Min, a Raytheon BBN scientist. “Our system seamlessly learns from queries and then provides better and more meaningful results, without making analysts do more work.”
Leveraging Raytheon BBN’s expertise in machine learning and large language models, the team developed cross-lingual Information extraction and retrieval technologies. These technologies address on-demand, personalized information needs via human-in-the-loop active learning.
“Context is something intuitive. Humans consider context naturally. For non-native speakers understanding context across languages is critical to extract the appropriate semantic information. With machine learning and artificial intelligence, we’re able to deliver the semantic information and search results that today’s intelligence analysts need,” explained Min.
Source: Raytheon
Your competitors read IC News each day. Shouldn’t you? Learn more about our subscription options, and keep up with every move in the IC contracting space.
By 
