Psych 07-Representation of Knowledge

Non‐symbolic (a.k.a. sub‐symbolic) representation

(Bottom‐up approach: Knowledge is represented by weights on connections in a network) Advantages: Noise tolerant (no formal rules to follow) Especially applied to perception, learning, pattern recognition, motor skills E.g., Neural Networks

Symbolic representation

(Top‐down approach: Information described with language, labels, ...)

What are propositional representations? What are semantic networks? What are they used for?

-

What are the differences between perception‐based (a.k.a. mental or perceptual imagery) and meaning‐based representations (a.k.a. knowledge representation)?

-

When remembering pictures, do we remember the physical picture (details) or its meaning (check Mandler and Ritchey [1977] and Gernsbacher [1985] experiment)?

-

meaning‐based representations (a.k.a. knowledge representation)

-Knowledge Representation or abstraction of information into memory -Meaning based representations: Mental representations that retain the gist of events and discard most of the detail. After time has passed, our memories lose some of the information of the original experience. Our loss of information is selective: We tend to retain the information that is most meaningful or useful.

Semantic Networks Quillian (1966)

-People store information in a network structure. -Semantic networks usually represent relations between concepts in a hierarchical organization of categories, such as that a canary is a bird and a bird is an animal, by linking nodes for the two categories with isa links. -The organization of the network and its hierarchies can be studied by measuring reaction times when judging the truth of assertions about information contained in the network. -Both the strength of the connections between facts and concepts (determined by frequency of experience) and the distance between them in the semantic network have effects on retrieval time.

What are schemas? What are they used for?

-Schemas represent categorical knowledge according to a hierarchical organized slot structure, in which slots specify values of various attributes that members of a category possess. Unless contradicted, a concept inherits the features of its superset. Used specially to represent beliefs and expectations about particular aspects of the world (e.g., Brewer and Treyens (1981): participants waiting in the office of the experimenter; at test, asked what did they remember from the room. Many participants remembered things that would be expected to be found in an office, but were not actually present in the room).

How do we process meaningful memories for visual information? How good is our visual memory for pictures and sentences?

-Shepard (1967): Compared memory for sentences (verbal condition) vs. memory for pictures (visual condition). -Subjects had higher rate of errors in the sentence condition (although with very good performance, 11.8%) than in the picture condition (1.5%). Perhaps the most impressive demonstration of visual memory is the experiment by Standing (1973), who showed that participants had only a 17% error rate after studying 10,000 pictures. But...

Mandler and Ritchey (1977)

Although people show very good memory for pictures, what they seem to be remembering is some interpretation of the picture rather than the exact picture itself. -Distinction between the meaning of a picture and the physical picture. Type change: meaningful change versus -Token change: irrelevant change, non‐meaningful Participants remember type changes more likely than token changes.

perception‐based (a.k.a. mental or perceptual imagery)

Perception-based representations: MENTAL IMAGERY Mental representations that preserve much of the structure of the original perceptual experience.

Propositional Networks

Similar to semantic networks, but the relation between concepts are formally defined. Formal representation of information in a network structure, based on logic propositions (propositional calculus) Proposition: smallest unit of meaning that can have a truth value.

Gernsbacher (1985):

Subjects were asked to recall pictures where the right‐left orientation was reversed, and asked to recall the pictures after 10 s and after 10 min. Memory for detail (verbal and visual) is available initially but is forgotten rapidly, whereas memory for meaning is retained.

How do we process meaningful memories for verbal information? Does being warned have any impact on how we process the information (check Wanner [1968] experiment)?

Wanner (1968): Study of warned vs. unwarned participants After processing a linguistic message, people usually remember just its meaning and not its exact wording (style), unless they are warned that stylist changes are an important part of the test, then, they will remember both meaningful and stylistic changes.

What information is more probably to be retained in memory?

We tend to retain the information that is most meaningful or useful