The agency wants natural language processing bots that any employee can customize and deploy without significant tech skills.
Natural language processing technologies have come a long way in recent years but the Office of Personnel Management wants to know if machines are smart enough to decipher the least natural language of all: government policies.
As computers get better at navigating the subtleties of human language, OPM officials are looking at how artificial intelligence could be used to “gain insights into statutory and regulatory text to support policy analysis,” according to a request for information posted Tuesday on FedBizOpps. Through bots and other technologies, officials are hoping these algorithms can read through large amounts of legalese and produce “a general understanding of the statutory and regulatory text as a whole.”
Ideally, the bots would be so simple to use that employees without notable tech skills would be able to build, customize and deploy them with ease.
Officials are also interested in gathering realistic information about upfront costs for deploying these technologies and have asked for respondents to include a rough order of magnitude determination with their submissions that includes “project costs—both initial and cycle costs—schedule, effort and risks.”
Potential solutions should include capabilities for topic modeling, text categorization, text clustering, information extraction, named entity resolution, relationship extraction, sentiment analysis, and summarization.
Responses are due to OPM by noon April 19.