KM and the environment: Water management uses analytics, big data and collaboration to handle complexity.

Cognitive computing accelerates environmental impact assessment

U.S. and international law requires an environmental impact statement (EIS) to assess the effect that certain actions will have on the quality of the environment. Those actions include construction of buildings and roads, oil exploration, mining, power plants and other major projects. Generating an EIS is a complex process; the elements in an EIS are similar across projects, but not similar enough for a template. Therefore, the development of an EIS is generally expensive and time-consuming.

Coseer has developed a cognitive computing application that uses three basic capabilities of cognitive computing (ingestion, extraction and discovery) to expedite the production of an EIS. “One of the steps is to list the factors that may affect the environment,” says Praful Krishna, CEO of Coseer. “The application can extract the factors from similar projects and tabulate them, which speeds up the process and makes it more objective.”

The system relies on three inputs: the entire corpus of previous impact studies, local area activity based on Internet research (for example, whether fishing is a local activity), and curated knowledge from the Internet, (such as information about similar projects in other regions). “These inputs enable a professional to assess the impact of the project in question within minutes,” adds Krishna.

The core software is Coseer’s DocSeer. The company works with clients to develop a user interface for the EIS and train the system on relevant documents. “The use of cognitive computing allows a more comprehensive review of relevant information and reduces the time and cost required. Without such a system impact assessment, professionals would have to manually go through thousands of documents or rely on subjective judgment and memory,” Krishna says. “We are taking capabilities already proven across industries and pointing them to a specific problem. You don’t need an army to pre-label data or train Coseer systems. We are up and running in 4-12 weeks.” The company hopes to make the system cost-effective so it becomes the norm, whether for small or large projects.