Chapter 11 linear notes

Page 1

Problem Solving

what are the diff types? how are they solved? why are reps important?

Problem can be a puzzle (activity), life hindrance or part of a practical task

Problem solving: need to achieve a goal but not sure how to achieve said goal

Protocol analysis: think aloud while problem solving - analysed - WM info verbalized (comes from LTM) - finds diff strategies and indiv diffs - can be problematic if affects nature of task

PA research: medical diagnosis - clinical terms more than previous knowledge - real life application

Simple Problem Solving

e.g. a puzzle - focus on rep not b/g knowledge

Gestalt - goal structuring and reps - x ray problem - break task into subgoals

Sets - misleading reps can inhibit psolving e.g. habitual strategies hinder better ones

Functional fixity - use object in a novel way - candles containers/steam trains irl

Reps - tower of Hanoi - diff rules/constraints - state space - moves and corresponding states

Isomorphs - two problems share same state space - reveals diffs e.g. diff rules (move/change) result in diff difficulty levels

Internal reps: processing and rep boundaries re encoding and storage of info, external: implicit rules, constraints from environment so no need to internalise - ease of psolving

Info processing: psolvers search for solution f/wards(actions) or b/wards (start w/goal) tower of Hanoi involves holding goals and results in ltd capacity WM - one action at a time - return to start and re-explore - systematic - progressive deepening - look ahead and evaluate possible steps till achieved - avoids going in circles

Info processing and insight: restructuring reps - generate actions/operators from LTM cued by prob rep - apply operator to prob -> new prob state and further operators - if misleading initial rep - no useful operators - mental blank - can help by adding info to rep or elaborating e.g. change encoding and make goal less restrictive

Analogical Problem Solving

use old solution to solve new problem

analogous story e.g. castle and soliders for x ray task

Anolli found hints help to utilise and retrieve these solutions - analogical paradox - real life more pressure to encode the story than lab bc artificial, superficial - hint leads to deeper rep

Structure mapping - structural alignment between base and target reps - connecting the two situations - computer model structural mapping engine

Basic analogies - diff base analogies re electrical flow - flow of fluid through pipe - helps battery probs

Complex Problem Solving

complex involves drawing on extensive b/g knowledge - expert = mental capacity or knowledge?

adversary - opponent e.g. chess, non-adversary - non-opponent e.g. code-breaking

Chess - superior info processing - think aloud - no quant diff between experts and skilled - recall reconstruction of board - experts better - reconstruction w/o mem test showed experts more meaningful placement - more knowledge and better organisation and accessibility of knowledge - skill depends on acquisition and usage of domain knowledge

Experts work f/wards - protocol anal - novices work b/wards - experts use domain knowledge to generate good rep of prob

Experts have better reps - think aloud when categorize prob - deeper structure properties compared to surface novices - group properties of problems together to aid future problem solving - more efficient

Practice makes perf - power of practice law - various evidence - indiv and seq task components executed more efficiently and qual better reps of task structure - declarative knowledge from WM used in larger chunks and shift from alg to mem based processing

Ericsson - start young as it takes time to become expert level - practice not always perf - same level skill waiters - best bc most efficient encoding strategies - indiv diffs within as well as between experts

Modal model - chess model of expertise - pattern recog hype - dependent on pattern ability from LTM -> also applies to non-adversary - links between expertise and knowledge - modal model - expertise depends on acquisition and organisation of domain rel knowledge and skill in LTM - criticised bc too descriptive - lack of explan power - makes experts just 'know more' and we know this is only part of it

Prospects for problem solving research

memory determines skill - questioned e.g. not mem but playing strength - best moves chosen

Generic and specific - if lack of domain knowledge experts use generic strategies? design experiment re reasoning - confirmed generic strategies which questions domain specificity of expertise and supports domain general skills

Indiv diffs in learners - novices not just lack but how they acquire skills - good/bad learners e.g. map - good = efficient encoding of spatial info and recognising gaps in knoweldge - comp progam - good learners use metacog processes and strategies

novices identified by how much explanation - protocol anal - good learners spontaneous self explanation - prompts to self explain aid learning - qual over length of explanations

Skill acquisition - b/wards anal can have neg effect on skill acquisition - emphasis on goal overloads system and means less resources for retrieving relevant info - goal task - b/ward, non-goal - forward - less emphasis on goal means better learning and skill acquisition

Alt expl - goals specificity vs systematicity - non specific goals aid learning bc more hyp testing - distinction btwn learning and PS important

Ignore irrel info - two sources we choose first to reduce amount of irrel info processed - against skill acquisition theory

Experts indiv diffs - homogeneous assumption 'typical' expert - there are indiv diffs - e.g. specialist vs expert? knowledge of commands but diff usage, also some domains not possible to know everything - diff kinds of expertise