Browse Prior Art Database

Methodology for Cognitive Security Applied to Automated Agent Capability

IP.com Disclosure Number: IPCOM000248604D
Publication Date: 2016-Dec-21
Document File: 5 page(s) / 76K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed are a methodology, framework, and apparatus to model access to cognitive resources based on cognitive roles. This disclosure provides a framework and methodology for artificially separating different agents’ capabilities into cognitive levels, which translate to access levels for information.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 30% of the total text.

1

Methodology for Cognitive Security Applied to Automated Agent Capability

Cognitive systems help users solve complex cognitive problems. Such systems provide answers, procedures, approaches, and knowledge in order to help human users complete cognitive tasks. The computer-human interaction is essential to the new cognitive era. Systems implemented today provide users limitless access to all cognitive capabilities.

Traditional access control models define user access based on different patterns. For example, Attribute-Based Access Control (ABAC) limits access to resources based on attributes. Discretionary Access Control (DAC) is based on ownership. Identity Based Access Control (IBAC) is based on user identities. Role-Based Access Control (RBAC) associates access with a role to be performed by a user.

Previous methodologies cannot be directly applied to many new cognitive resources, having been conceived and designed for other types of resources, systems, and technologies, many of which differ considerably from cognitive resources, systems, and technologies. In addition, the cognitive resource of a neural network structure lacks the ability to secure resulting “output” neurons. Further, a decision tree as a cognitive resource lacks the ability to secure particular branches of the decision tree.

The novel contribution is a methodology, framework, and apparatus to model access to cognitive resources based on cognitive roles. This disclosure provides a framework and methodology for artificially separating different agents’ capabilities into cognitive levels . In this context, defining cognitive levels is to separate cognitive capabilities into different groups, levels, or categories. These discrete cognitive groups, levels or categories are available to different roles or users, depending on the level of access attached to them.

A Cognitive Security component (Figure 1) is responsible for using the information about the access level assigned to the user or the role, and with that, interacting with every single cognitive resource to determine the best way to implement the cognitive access level associated with the user.

This methodology and framework enable companies to assign different levels of expertise to individual users based on different criteria. It allows companies to provide cognitive based permission levels and customized security treatment to the users so that a user gains access to information based on an associated cognitive level (i.e., level of skills or expertise).

For example, the agent gives an inexperienced information technology (IT) user detailed step-by-step instructions, but gives an experienced IT more advanced instructions and tasks, even to achieve a similar task. In another example, a cognitive agent may detail the full explanation and instructions for how to apply a bandage to a person, including videos, graphics, using basic common language. The same cognitive agent may directly tell a nurse where to apply the banda...