lecture 1 - Knowledge structures and expertise Flashcards
what is memory
-types
memory is a general term for information storage
declarative memory (episodic and semantic)
-knowing ‘that’
procedural memory
-knowing ‘how’
episodic memory
semantic memory
Episodic memory:
here’s what I did this morning to
make an omelette
-memory of personal experiences and specific events and conetxt, like a mental diary of your life
Semantic memory:
here’s what ingredients one needs to
make an omelette
-stores general knowledge about the world, concepts, facts, and meanings, without personal context. It’s not tied to specific events or personal experiences.
the 3 basic processes of memory
- encoding aquisition
- storage
- retreival (and reconstruction)
why is structure so important for encoding and storage :building knowledge
Structure helps hold onto information:
Unstructured information is highly prone to forgetting (in particular, episodic information is highly prone to forgetting).
The goal of learning is not just to hold onto specific episodes – it’s to store knowledge.
→ Knowledge is organized / structured information.
The knowledge that you already have will then influence processing of new incoming information.
Structure helps
providing structure experiment
-Brandsford and Johnspn 1972
-participants read a text and then receive a memory test
-half of the participants got context (a photo eg balloon photo)
Recall of main ideas from the story
* no picture: 26%
* with picture: 57%
Providing a context – even a novel, unexpected, idiosyncratic context –
helps organize and hold onto information
what do all models of semantic memory assume about knowledge structures
Different types of models of semantic memory, but all models assume that information is organized in a network.
e.g., Collin & Loftus (1975)
-info we have organised in network where the bubbles correspond to inidvidual conecepts and the links correspond to the connections
types of structures (in semantic memory)
- chunks (WM) working memory)
- schemas (larger structures acquired through experience, stored in LTM)
working memory : what counts as an item)
Initial capacity estimate for STM: 7 ± 2 items (Miller, 1956)
16 numbers:
1492177619842001
12 letters:
JPNOHONSKSHI
10 words:
Wales Jane on went a Mark camping to holiday and
what is chunking
Chunking - a memory technique that involves breaking down large pieces of information into smaller bits or “chunks.” knowledge-dependent recoding of information
- a chunk is a meaningful group of information
- chunking relies on accumulated knowledge (i.e., on the contents of LTM)
- chunking is a means of functionally increasing working memory capacity
eg putting the 16 numbers into 4 dates makes it easier to remember
-what is WM’s capacity for amount of chunks?
- can this vary?
– WM capacity is estimated at 4 chunks
– the size of these chunks can vary, the total number of chunks does not
complex structures : schemas in LTM
example of schema
-schemas are clusters / configurations / frameworks (“chunks”) of knowledge for objects, places, events, actions, etc. (schemas for sequences of actions are called scripts)
Schemas are abstract: they hold information about a type of object, place, event, or action rather than about a specific one
-hold fixed info and then variable parts
eg of schema : schema of houses , inuding frame and slots.
fixed part :
-windows
-door
-roof
-wall
variable parts
-size of window
-sixe/type of door
-shape of roof
-color of walls
what are schemas good for
schemas can influence encoding, storage, and retrieval
how do schemas reduce processing load
Schemas create expectations, reduce processing load.
– We don’t have to encode and remember every detail
* FRAMES: central things are expected and encoded quickly
* SLOTS: can be filled with different types of information
– Just activate the right schema at encoding (this takes place automatically) and make note of differences between current instance and the schema
how do schemas allow us to deal with missing information?
- Schemas allow us to deal with missing information by making inferences:
– e.g., if no specific information is given, we’ll assume that a house has, by default, a pointy roof
schema experiment
-office exp
results
-take particpants in the office room, and then take them out and ask them to write down what they remember
-some of the ibjects in offcice were objects you wouldn’d normally see in an office
Better memory for schema-consistent objects (desk, shelves) than inconsistent objects (skull, wine bottle). Also false recall of schema-consistent but absent objects (books, pens, window)
whats the issue with schema supporting the drawing of inferences
schemas support drawing of inferences…
thus also allow for the possibility of predictable errors
what do some experiments about schema inconsistent information show
’ the consistency effec’
But multiple other studies show better recall of schema-inconsistent information
-This information may be encoded with priority: it is unexpected and thus cannot be reconstructed later by activating a schema
-Longer delays: memory becomes schema-consistent over time
providing schemas during encoding…..
helps you retain information
eg giving a title to a pargraph about washing clothes without context
providing schemas at retreival
-home buyer experiment
Anderson & Pichert (1978): -homebuyer/burglar experiment
-Participants are told to read a story either from the perspective of a homebuyer or the perspective of a burglar
First recall (from one perspective)
Participants recalled more items consistent with the queried schema than items consistent with the other schema.
Second recall (after shifting to the other perspective)
Participants now also recalled more items consistent with the new queried schema than during the first recall, as well as fewer
items consistent with the other schema.
→ schemas facilitate retrieval by functioning as retrieval cues
schenas facilitate retreival by….
schemas facilitate retrieval by functioning as retrieval cues
accumulation of knowledge - expertise
-definition
domain-specific mastery, consistent superior performance in a domain
* high levels of knowledge of concepts, facts, procedures
* better organization of this knowledge
how does expertise enhance memory
-example
Important role of expertise: experts are able to create chunks more often (Skilled memory theory; Chase & Ericsson, 1982; Ericsson & Chase, 1982)
-SF increased his digit span to ~80
(200-300 hours of practice)
How? Combining smaller chunks into more complex
chunks (i.e., a complex retrieval structure)
other examples of expertise enhancing memory
Rajan had a digit/letter span of ~15, memorized 30,000 digits of π
(Thompson et al., 1991, 1993)
expert chess players recall chess positions presented very briefly (5 sec) better than novices
shows effect of expertise
(De Groot, 1946; Chase & Simon, 1973
what does exceptional memory require
-principles
exceptional memory = acquired skill (Ericsson et al., 2004)
Principles:
1. Encoding principle: rely on prior knowledge (LTM) to create chunks
2. Retrieval structure principle: use retrieval cues as “pointers” to information in LTM
3. Speed-up principle: speed up with practice
Long-Term Working Memory (LTWM) model proposed by Ericsson and Kintsch in 1995
This theory addresses how expert performance in a specific domain (e.g., chess, music, sports) develops and is maintained. It emphasizes that the key to expert performance is not just the amount of information someone has, but how easily and efficiently they can encode and retrieve relevant domain-specific knowledge from long-term memory (LTM).
