March 18, 2019

Egocentric Versus Allocentric Artificial Intelligence

Have you ever noticed that you make different decisions about a person or situation based on whether you had direct interaction with the situation or observed it from afar? This is the fundamental difference between egocentric and allocentric relationships.  No, this is not a blog meant for psychologists, it is actually targeted to IT security professionals looking for a better understanding of the new approaches that Artificial Intelligence and Machine Learning (AIML) bring to cyber security.

Egocentric AI

Put most simply, egocentric relationships are one to one. Specifically, what is directly observed by the source. A Stanford study titled “Understanding Egocentric Activities” explains the value delivered by egocentric models:

“We demonstrate the ability to learn a hierarchical model of an activity by exploiting the consistent appearance of objects, hands, and actions that results from the egocentric context.  We show that joint modeling of activities, actions, and objects leads to superior performance in comparison to the case where they are considered independently.”

The paper goes on to discuss in detail how an egocentric model can independently observe and predict behaviors in a far superior method than the classic big data approach of analyzing huge data sets to make better decisions: 

“We show that object and action models can be learned with very little supervision, by exploiting the joint properties of objects, hands, and actions.  We propose a hierarchical inference architecture in which bottom-up propagation of evidence for objects and actions is used to predict the activity category, followed by top-down refinement of object and action descriptions based on the activity model. We demonstrate that our approach can produce superior results in comparison to standard bag-of-words type representations for activity categorization.”

Allocentric AI

Egocentric modeling alone is not enough to get a true view of the subject being observed because of the potential bias of the source. Adding an allocentric view will create an additional dimension from which to make better formed decisions. Put most simply, allocentric relationships are between two entities (i.e. a person and their mobile phone) but observed by a 3rdparty.

For clarification: In an egocentric relationship, the source can observe the target directly and directly develop information about the interaction, while in an allocentric relationship the source observes the target interacting with a 3rdparty and can then infer information about the interaction.

If you are interested in a deeper understanding of the differences and relationships of egocentric and allocentric then check out this Harvard study titled “Development of Spatial Ability Tests” which reported: 

“In our lab, we examine individual differences in egocentric (imagining taking a different perspective in space) and allocentric (mentally manipulating objects from a stationary point of view) spatial abilities and develop assessments of these abilities. Our research also seeks to discover the relation of these two types of spatial ability to locomotion and spatial navigation.”

So, in the context of AIML-based identity authentication it is the algorithm observing and modeling specific cyber behavior patterns of an individual based on both egocentric and allocentric perspectives in order to build a machine learning model that can predict when someone else is trying to use credentials not meant for them.

Continuous Cognitive Authentication

Today we need both an ego and allocentric view to make decisions, especially if we want to establish an immutable identity for cyber authentication. Acceptto understands that you are extremely busy and don’t have time to waste on remembering passwords, retrieving and typing in PINs or scanning various body parts in order to differentiate your identity from someone else. Your immutable identity is a combination of your physical behaviors, attributes and Digital DNA. We believe passwords are no longer relevant and that what you need is a way to immutably authenticate someone in order to be truly secure and compliant.

We call it Continuous Cognitive Authentication. You can eliminate preventable harm with our Biobehavioral AIML technology that enables frictionless authentication, prevents credentials stuffing instantaneously, ensures your true immutable identity continuously, and dramatically reduces risk, likelihood of fraud and cost of helpdesk operations without the guesswork or latency.

Acceptto is a transformative multi-factor authentication technology that delivers continuous identity protection and peace of mind in an age where passwords are ineffective and identity authentication is mission critical.

See for yourself what Acceptto can do to ensure your employees, partners and customers can authenticate without passwords and still ensure security and privacy, especially for your PCI compliance requirements. Request more information today.

 

cyber security identity authentication egocentric vs allocentric