What is cognitive architecture in HCI?
Cognitive architecture is a theory about the structures that create a mind in natural or artificial systems. It focuses on how these structures work with each other and use the knowledge and skills that are incorporated into the architecture to create and manage intelligent behavior in various complex environments.
The Elementary Perceiver and Memorizer (EPAM), created in 1960 by Ed Feigenbaum was one of the first possible cognitive architecture models. He intended to use the EPAM cognitive architecture model to glean insights into the inner workings of the human brain.
Generically cognitive architectures include creating artificial intelligence and modeling natural intelligence at appropriate levels of abstraction. A grand unified architecture is integrated across higher-level thought processes as well as aspects that are essential for successful intelligent behavior in human-like environments. These include emotions, motor control, and perception. Functionally elegant architectures bring an expanse of capabilities from interactions with a tiny set of mechanisms. These can be considered to be a set of cognitive Newton’s laws.
A cognitive architecture is more than just a theory of cognition. It has been defined to be an embodiment of “a scientific hypothesis aboutthose aspects of human cognition that are relatively constant over time and relatively independent of task.”
Essentially, that means that it is an attempt to describe those aspects of the human cognitive system that are pretty much universal, both across as well as within individuals. A cognitive architecture alone usually is not able to describe human performance on any specific task, it needs to be provided information about how to carry out that particular task. This information is usually based on a thorough task analysis of the target activity that is being modeled.
A cognitive architecture is also a piece of executable software. It is code written by a programmer or several programmers (usually the latter). This is a major way in which cognitive architectures are different from the majority of theories in cognitive psychology.
The things that research on cognitive architectures brings to HCI isn’t quite obvious right off the bat. It could seem like rather highly theoretical cognitive science, especially when researchers are debating the ramifications of low-level features of the architecture. But cognitive architectures may in fact be some of the most HCI-relevant cognitive science work there is. There are several roles that cognitive architectures can fill in HCI research and practice.
What kind of outputs do cognitive architectures produce?
The vast majority of cognitive architectures don’t just produce a prediction about performance, they actually output actual performance. They generate a timestamped sequence of actions that can be compared with actual human performance on a task.
These timestamps indicate that cognitive architectures create models that are quantitative. These models can do more than just predicting that one task is faster than the other, they can predict exactly how much faster the task is than the other task. This has several implications in the domain of engineering.
The knowledge that has to be supplied to the architecture usually needs to be supplied in the language of the architecture. Structuring knowledge in this form is quite like programming, so architecture-based modelers generally tend to have strong programming skills.
Sometimes, the cognitive architecture interacts directly with the actual software that the human users are utilizing for the purpose of performing the task. In other situations, some sort of connecting software needs to be built. The model of a task usually has three components: the architecture, task knowledge, and a dynamic task environment that the model interacts with. The output of this system is a a timespamped behavior stream.
Cognitive architectures usually tend to be research tools in academic laboratories, but there are now quite a few consulting and technical companies that make use of cognitive architectures in their work and even offer models developed with cognitive architectures to their clients.
What is the purpose of cognitive architecture?
Cognitive architecture seeks to employ the research that is carried out in the domain of cognitive psychology to build a complete computer model of cognition. Cognitive architecture acts as a blueprint for creating and implementing intelligent agents.
It concentrates on merging cognitive science and artificial intelligence and seeks to create artificial computational system processes which behave like natural cognitive systems.
The ultimate purpose of cognitive architecture is to model the human brain and eventually empower us to build artificial intelligence that is on par with humans (Artificial General Intelligence).
Which are the successful cognitive architecture models?
In 1995, Russell S and Norvig P said that there are four ways through which artificial intelligence could be realized: systems that think like humans, systems that think rationally, systems that act like humans, and systems that act rationally.
All of these ways have been explored by the cognitive architectures currently available.
Here are three of the successful cognitive architecture models:
1. Soar
Soar’s primary goal was to create generalized agents that could perform multiple tasks and would even act as the foundation for the emulation of human cognitive capacity. Drawing from ACT-R and LIDA (which shall be highlighted subsequently), it was developed at Carnegie Mellon University by John Laird, Allen Newell and Paul Rosenbloom.
2. Active Control of Thought-Rational (ACT-R)
Active Control of Thought-Rational (ACT-R) seeks to understand how the brain organizes itself into singular processing modules, thereby minimizing cognitive functions to the most basic operations that can enable cognition. ACT-R was developed by John Robert Anderson, also at Carnegie Mellon University.
3. Learning Intelligent Distribution Agent (LIDA)
The Learning Intelligent Distribution Agent (LIDA) model was developed as an integrated model. It sought to model human cognition all the way from perception & action to high-level reasoning. The LIDA model was developed at the University of Memphis by Stan Franklin and his colleagues.
