[HCI] 2. Human 2

Thinking

• Reasoning(추론) - 앞 뒤를 생각해서 이건 이렇게 될 것 같아! 처럼 추측하는 행위
  • deduction, induction, abduction
• Problem solving - reasoning과는 달리 답이 없는 상태에서 답을 찾아내려고 하는 방식

Deductive Reasoning

• Deduction(연역): If A, then B
  • Derive logically necessary conclusion from given premises(전제).
    • e.g. If it is Friday then she will go to work
    • It is Friday
    • Therefore she will go to work.
  • Logical conclusion not necessarily true:
    • e.g. If it is raining then the ground is dry
    • It is raining
    • Therefore the ground is dry
• 꼭 옳다는 것을 의미하는 것은 아님(잘못될 수도 있음)
  • 전제가 틀려있을 수도 있다.

Deduction (cont.)

• When truth and logical validity clash …

  • 조건이 불충분할 수 있다.
  • e.g. Some people are babies
  • Some babies cry
  • Inference(추론) - Some people cry
  • some 때문에 전체적으로 맞는 경우가 아니라 부분적으로 맞는 경우가 된다.(불충분)

  Correct?

• People bring world knowledge to bear

Inductive Reasoning(귀납적 추론)

• Induction: Generalizing from previous cases to learn about new ones
  • Generalize from cases seen to cases unseen
    • case들을 모아 일반화시키는 방식
    • e.g. All elephants we have seen have trunks, therefore all elephants have trunks.
• Unreliable:
  • Can only prove false, not true – checking all elephants (?)
  • 하나만 잘못된 것을 찾으면 바로 틀린 논제가 되어버림
    • 그래서 이를 증명하려면 모든 경우를 확인해야 함.
  • … but useful!
    • 하지만 효율적임!
• Humans not good at using negative evidence
  • e.g. Wason’s cards.

Wason’s cards

If a card has a vowel(모음) on one side, it has an even number on the other

Is this true?

How many cards do you need to turn over to find out?

…. and which cards?

• E를 뒤집으면 반드시 짝수가 있어야 하지만 4를 뒤집었을 땐 모음이 없어도 된다.(조건을 잘 읽으면…)
  • 그런데 사람들은 4뒤에는 모음이 있어야 된다고 착각
• 이것이 negative evidence!
• 조건문이 앞에 것이 뒤에 것을 보장하지만 뒤에 것이 앞에 것을 보장하지 않는다.

Abductive Reasoning(가설 유도적 추론)

• Abduction: Reasoning from a fact back to the action or state that caused it
  • Reasoning from event to cause
    • e.g. Sam drives fast when drunk.
    • If I see Sam driving fast, assume drunk.
• Unreliable:
  • Can lead to false explanations(가설이기 때문에)

정말 모르겠을 때 추론하는 방법

Reasoning

• Deductive reasoning is valid
  • Deductive inference guarantees if the premises are true so it the conclusion
• Inductive and abductive reasoning are unreliable
  • Cannot guarantee true conclusions even with true premises(전제)
  • Inductive reasoning: conclusion merely likely
  • Abductive reasoning: taking your best shot
• Deductive reasoning is limited to application of rules to examples
• Inductive and abductive reasoning are important for generating hypotheses

Reasoning (cont.)

• Humans rely on inductive and abductive inference, as well as deduction
• Humans regulate inherent unreliability of these reasoning modes
  • Using contextual knowledge
  • Using belief revision
• Computers rely primarily on deductive inference
  • Although some AI techniques use induction and abduction

Reasoning to Problem-Solving

• Where reasoning involves inference about a familiar problem domain
• Problem solving refers to the ability to design solutions to problems in unfamiliar problem domains
• Humans use at least three types of problem-solving techniques
  • Heuristics (해 보고 판단하는 것)
  • Analogy(유추) and metaphor(비유)
  • Learning

Problem-Solving Techniques

• Heuristics: using informed trial and error based on rules of thumb
  • 사람이 대충 정해서 하겠다.(사바사)
  • E.g. Chess openings (develop your pieces, gain control of the center of the board)
• Analogy and metaphor: adapting solutions from one problem domain to another
  • E.g Object-orientation (treating software modules as physical objects)
• Learning: improving performance by acquiring skills over time with repeated exposure to a problem
  • E.g. Multiply 1496 by 20

How Problem-Solving works

• Process of finding solution to unfamiliar task using knowledge
  • Storage in LTM, then application of knowledge
• Observation:
  • People are more heuristic than algorithmic - 사람들은 algorithmic보단 heuristic을 더 좋아함.
    • Try a few quick shots rather than plan because resources are simply not available
  • People often choose suboptimal strategies for low priority problems
    • 우선순위가 낮다고 생각될 수록 대충함
  • People learn better strategies with practice
    • 연습하면 더 좋아짐
• Several theories.
  • Gestalt
  • Problem Space Theory
  • Analogy
  • Skill Acquisition

Gestalt Theory

• means “unified whole”
• Problem solving both productive and reproductive
• Productive draws on insight and restructuring of problem
• Attractive but not enough evidence to explain `insight’ etc.
• Move away from behaviourism and led towards information processing theories
  • 행동주의에서 벗어나 정보처리 이론으로 이어집니다.
• Principles:
  • Similarity
  • Continuation
  • Closure
  • Proximity (근처에 있는 것일수록)
  • Figure and Ground(foreground, background에 따라서)
• 전혀 상관없어 보이는 도형일지라도 모이면 무슨 의미인지 알 수 있다.

Problem Space Theory

• Problem space comprises problem states
  • 문제를 단계적으로 해석
  • state들을 내가 원하는 쪽으로 몰고 감.
• Problem-solving involves generating states using legal operators
• Heuristics may be employed in order to select operators
  • e.g. means-ends analysis
  • 수단을 통해 목적을 달성하도록 하는 것
• Operates within human information processing system
  • e.g. STM limits etc.
  • short term memory가 관여
• Largely applied to problem-solving in well-defined areas
  • e.g. puzzles rather than knowledge intensive areas

​

Other Problem Solving Issues

• Analogy
  • Analogical mapping:
    • Novel problems in new domain?
    • Use knowledge of similar problem from similar domain
  • Analogical mapping difficult if domains are semantically different
    • 의미적으로 다르면 비유적인 mapping 어려움
• Skill Acquisition
  • Skilled activity characterized by chunking (묶는 행위)
  • Lot of information is chunked to optimize STM
  • Conceptual rather than superficial grouping of problems
    • 문제의 표면적인 그룹화가 아닌 개념적 그룹화
  • Information is structured more effectively

Mental Models

• How do we understand how the world works?
  • 우리는 세상이 어떻게 돌아가는지 어떻게 이해하는가
  • More than simple memory (단순한 메모리 공간 이상이다.)
  • Includes expectations of behaviour
    • 행동의 기대를 포함
• I throw a ball at my friend, and she catches it
  • How did I know how hard to throw it?
  • How did she know where to stand to catch it?
• We have an understanding of the mechanics of the world
  • How objects interact, the relationships between them
  • 물체가 어떻게 상호작용하고 있는지 그 메카닉 적인 부분을 잘 이해하고 있음.

What is a mental model?

• A cognitive structure that encodes how an aspect of the world operates
• Includes information about which object classes interact with which other classes
• Also includes information about how objects interact, and how interactions change properties
• Malleable(온순한) structures which are strengthened by successful application
  • successful application으로 강화된 온순한 구조

Mental model example

The circle is above and to the right of the square

• We can ask questions about this situation
  • Which object is to the left?
  • Which object is at the bottom?
  • Would the circle balance on the square?
• By creating a mental model, these questions are simple to answer
  • Add in knowledge we already have
  • You can simply ‘see’ the answer

How we use mental models

• We use mental models to generate predictions
  • Predict where the ball will land
• Expertise is generally associated with mental models closer to reality
  • An expert cricket(크리켓;운동) player understands ball physics better
  • But mental models are always unconscious (‘knowing’ doesn’t help)
• Mental models are not based on accurate physics
  • Can lead to incorrect predictions
  • Based on experience, etc.
  • Change as they are used
  • 실제로 그렇지 않더라도 경험에 의해서 그렇게 생각이 되어진다.

Expanding mental models

• HCI researchers have expanded the idea
  • Includes a user’s understanding of the workings of a computing device
  • Expectations of menu structures, where files are stored, etc.
• In HCI understanding user’s mental models is important
  • Understand how learning a system works
  • Reduce stress/workload by supporting user’s mental models
  • Tailor systems to fit how users perceive the system

Mental models and user-interfaces(UI)

• Any interface requires a mental model to use
  • What to expect when typing, clicking, etc
  • How to interpret the consequences of actions (e.g. link new window with clicking of the icon)
• Mental models are important when interface has little feedback
  • Difficult to recognize results
  • Few cues to evaluate the new state of the system
  • Problems with impoverished(궁핍한) devices

Errors and Mental Models

• Types of Error

  • Slips

    • Right intention, but failed to do it right

    • Causes: poor physical skill, inattention(부주의) etc.
      • 어떻게 하는지 알고 있는지 잘못누름

    • Change to aspect of skilled behaviour can cause slip

    • 잘못누른 것이기 때문에 설명으로 해결되지 않음

• Mistakes

  • Wrong intention
  • Cause: incorrect understanding (잘못된 이해)
    • humans create mental models to explain behaviour.
    • if wrong (different from actual system) errors can occur
  • 처음부터 이해를 잘못한 것.

Emotion

• Various theories of how emotion works
  • James-Lange: emotion is our interpretation of a physiological response to a stimuli
    • 자극에 대해서 신체적 반응이 나타내는 것
  • Cannon: emotion is a psychological response to a stimuli
    • 심리적으로 나타나는 것
  • Schachter-Singer: emotion is the result of our evaluation of our physiological responses, in the light of the whole situation we are in
    • 개가 짖으면 어떤 반응이 보여질 텐데 나중에 가서 ‘아 이게 공포였구나’를 느끼는 것이 감정이다.(?)
• Emotion clearly involves both cognitive and physical responses to stimuli

Emotion (cont.)

• The biological response to physical stimuli is called affect
• Affect influences how we respond to situations
  • positive -> creative problem-solving
  • negative -> narrow thinking
  • “Negative affect can make it harder to do even easy tasks; positive affect can make it easier to do difficult tasks” - Donald Norman
• Implications for interface design
  • Stress will increase the difficulty of problem-solving
  • Relaxed users will be more forgiving(용서) of shortcomings(결점) in design
  • Aesthetically pleasing and rewarding interfaces will increase positive affect
    • interface의 효율 뿐 아니라 보기 좋은 것 -> positive affect를 주기 때문에

positive affect는 중요하구나!

Individual Differences

• Long term
  • sex, physical and intellectual abilities
• Short term
  • effect of stress or fatigue(피로도; 지속적인)
• Changing
  • age

Ask yourself: will design decision exclude section of user population?

Psychology vs. Design of Interactive System

• Some direct applications
  • e.g. blue acuity is poor
    • blue should not be used for important detail
• 눈에 띄면서 자세한 걸 원하면 그린
• 눈에 확실히 띄게 하려면 레드
• 블루는 자세한 걸 나타날 때 맹점을 보임
• However, correct application generally requires understanding of context in psychology, and an understanding of specific experimental conditions
• A lot of knowledge has been distilled in
  • guidelines
  • cognitive models
  • experimental and analytic evaluation techniques