Chapter 2 - The Discipline Matures: Three Milestones Flashcards
Describe Stephen Kosslyn’s experiments
He gave images for subjects to analyze and imprint, giving them a part of the image to focus on, and then distributed parts of the objects depicted.
He discovered that the time needed to recognize the parts increased the further away the part was visually from the place he instructed them to focus on.
He concluded from this trend that image processing involves the scanning of image representations, suggesting the presence of mental images.
Quasi-pictorial imagery
How did ELIZA function?
ELIZA functioned by rephrasing statements made by the people it was conversing with into questions, and when this fails, using set responses to compensate and retry.
ELIZA simply responds to ques by identifying components of a phrase, and through a procedure rephrasing and returning it. The program does not analyze in any capacity the language, nor can it answer questions.
How did SHRDLU function?
SHRDLU was programmed to exist in a virtual world which was displayed on an interface. The program was capable of taking written commands by identifying objects. It also integrated these objects into its memory, and thus was able to have a working memory that made it more realistic.
It was programmed to be able to interact with the world, moving and replacing blocks and physical objects.
How does a computer create an image? Is this how the mind does? What does research suggest?
Computers create images through the use of binary representations on pixel interfaces, with binary being used to program colors and the like, and through that creating an image on an interface.
The mind is suggested to use another technique, as the difference in time to rotate images mentally as suggested by the mental rotation experiment (Shepard and Metzler) shows that the way the information is stored should be different. The mind takes much longer as the degree rotation increases to adjust the image.
The cognitive explosion during human development (ontogeny) runs parallel to what?
The acquisition of language
What are the functional groups (by function) SHRDLU’s distinct systems are split into?
Syntactic analysis, which involves to decoding of sentences grammatically. Parsing a sentence involves identifying units in the sentence by function.
Semantic analysis, which assigns meaning to words in context.
Integrating information, with current information being analyzed with information the system already possesses. Gives additional meaning to phrases
What are the levels of sophistication in visual images (by Marr(s))?
Marr identified three different types of visual sketches, which occur sequentially and are increasingly complex.
A primal sketch includes basic information, such as the intensity of stimulus (light) and basic geometry as detected and perceived by the eyes.
A 2.5D sketch (flippin what?) draws information from the primal sketch to create a model with points that show orientation and indicates a distance from the singular point of vision.
A 3D sketch involves the creation of a virtual image that maintains the shape of the object in motion. This image remains consistent despite changes in the viewer’s particular point of view
What are two cognitive perspectives on the way the mind and its processes can be studied (they’re opposites, Chaaarlie)? Describe in relative detail.
Bottoms up study, which involves examining the processes of the cellular and molecular behaviour of the brain and the nervous system(s). This means researching neurons and their interaction with each other and themselves, ligands/neurotransmitters, and from there expanding to neural oscillations, then systems and the way they interact, etc etc.
Top down study involves creating theses or ideas on how broad systems interact with one another to induce cognition, and from their magnify onto the mechanisms and more microscopic processes involved in their functions.
Remember kiddos, bottoms up, tops down, well that’s the best way to spend your Friday’s
What did Marr conclude from a study conducted by Elizabeth Warrington (cuppa tea gov’na?) on vision?
Due to the difference in visual deficiencies between groups of subjects with different lacerations/damage on the left and right side of their brains, Marr realized that different parts of the brain, physically located in different place, interact to create vision.
What is ELIZA?
ELIZA is a program among many that simulates human conversation in its responses.
Designed by Joseph Weizenbaum
What is missing from Chomsky’s ideas of transformational linguistics? For good measure, what are his ideas?
Chomsky’s transformational linguistics described linguistic understanding and the abstract structure of human language. Talked about how deep structure is an algorithm that humans use to formulate to create the surface structure language used, but he fails to offer an explanation as to how that happens.
It describes what knowledge of language humans have, but he doesn’t show how it is stored or how it is used.
What is the algorithmic level of analysis identified in the study of vision? Who identified it?
It involves the examination of how a cognitive system processes information. Essentially, an exploration of the function that transforms an input into its respective output.
Essentially explains what the computational level identifies. Does so by identifying and listing the complex set of instructions involved in the transduction of information.
Once more, flying saucers invade the stratosphere, and Marrs is the culprit. What has the government been keeping secret?
What is the computational level of analysis identified in the study of vision (what are its tasks, goal)? Who identified it?
Deals with cognition on a general scale, and is tasked directly with:
Identifying functional roles and parts involved with the processing of information (in this case visual)
What the nature of the task being performed is, where the signals come from and the like
Essentially, it is the identification of the stimulus and information put into a system, and what gets put out as a result of it.
Our boy Marr’s bars did
What is the guiding assumption in the computational level of analysis?
Cognitive function is the transformation of one type of information to another.
What is the idea of Marr’s vision? What levels of analysis did he use?
Marr’s vision is the integration of various areas of study into an explanation of the cognitive processes involved with vision.
He identified three levels of analysis, in increasing level of complexity and magnification:
> Computational level
> Algorithmic level
> Implementational level