HCI 4) Interaction Flashcards
What makes a good theory of interaction?
- Theories should explain
- Theories should predict
- Theories should help us evaluate
- Theories should guide measurements
- Theories should inform design
Mixed initiative interaction (MII)
Idea of organizing interaction in dialogue where both the computer and human can take the initiative by coupling an automated service.
Needs to infer user’s goals, ask about them and get the right timing.
Principles of Mixed-Initiative Interaction
- Developing significant value-added automation
- Considering uncertainty about a user’s goals
- Considering timing in the status of a user’s attention
- Inferring ideal action in light of costs, benefits, and uncertainties
- Employing dialogue to resolve key uncertainties
- Allowing efficient termination
- Minimizing cost of poor guesses and timing
- Providing mechanisms for efficient agent-user collaboration
- Continuing to learn by observing
- Maintaining working memory of recent interactions
Utility
The utility of an interactive system concerns its match with the tasks users want to do.
i.e. The relation between functionality and users’ needs and wants.
Usability
Usability concerns how easily computer-based tools may be operated when users try to accomplish a task.
Concerns whether users can actually realize a system’s utility.
Acceptability
Acceptability determines how likely a user would choose to use a tool given a choice.
- Practical: cost, reliability, compatibility with other systems
- Social: social norms and roles in settings where they are used
Accessibility
Accessibility is the diversity of users and situations the product can be involved in.
Automation
Allocation of tasks to machines
e.g. autocomplete, error correction, photo enhancement, recommendations, etc.
Types of automation - Acquisition
System sensing and registering of input data
Types of automation -
Analysis
Automation of information analysis
Types of automation -
Decision
Deciding and selecting appropriate actions among decision alternatives
Types of automation -
Action
The machine is partially or fully executing an action choice
Types of automation -
Adaptive
The type and level of automation is allowed to vary depending on context
Levels of automation
0 (human only)
-> 10 (fully autonomous)
User-centric evaluation criteria
- Reduce/increase mental workload
- Affect situational awareness
- Cause complacency from overconfidence or excess trust
- Skill degradation
System-level evaluation criteria
- Automation reliability
- Costs of decision and action outcomes
Personalization
Changing the appearance of interactive systems (non-functional changes)
Cognitive effects due to improved ease of use. Social effects by identity. Emotional effects by feelings.