Objects of Choice

Mind ◽  
2020 ◽  
Author(s):  
Wolfgang Schwarz
Keyword(s):  
Expected Utility ◽  
Rational Agents ◽  
Special Domain ◽  
Adequate Representation

Abstract Rational agents are supposed to maximize expected utility. But what are the options from which they choose? I outline some constraints on an adequate representation of an agent’s options. The options should, for example, contain no information of which the agent is unsure. But they should be sufficiently rich to distinguish all available acts from one another. These demands often come into conflict, so that there seems to be no adequate representation of the options at all. After reviewing existing proposals for how to construe decision-theoretic options and finding them all wanting, I suggest that our model of rational agents should include a special domain of ‘virtual’ option propositions to serve as formal objects of deliberation and choice.

scholarly journals Computational Goals, Values and Decision-Making

2020 ◽  
Vol 26 (5) ◽  
pp. 2487-2495 ◽  
Author(s):  
Louise A. Dennis
Keyword(s):  
Artificial Intelligence ◽  
Decision Making ◽  
Utility Function ◽  
Expected Utility ◽  
Engineering Ethics ◽  
Rational Agent ◽  
Autonomous Agent ◽  
Science And Engineering ◽  
Rational Agents ◽  
Do So

Abstract Considering the popular framing of an artificial intelligence as a rational agent that always seeks to maximise its expected utility, referred to as its goal, one of the features attributed to such rational agents is that they will never select an action which will change their goal. Therefore, if such an agent is to be friendly towards humanity, one argument goes, we must understand how to specify this friendliness in terms of a utility function. Wolfhart Totschnig (Fully Autonomous AI, Science and Engineering Ethics, 2020), argues in contrast that a fully autonomous agent will have the ability to change its utility function and will do so guided by its values. This commentary examines computational accounts of goals, values and decision-making. It rejects the idea that a rational agent will never select an action that changes its goal but also argues that an artificial intelligence is unlikely to be purely rational in terms of always acting to maximise a utility function. It nevertheless also challenges the idea that an agent which does not change its goal cannot be considered fully autonomous. It does agree that values are an important component of decision-making and explores a number of reasons why.

scholarly journals The Off-Switch Game

2017 ◽  
Author(s):  
Dylan Hadfield-Menell ◽  
Anca Dragan ◽  
Pieter Abbeel ◽  
Stuart Russell
Keyword(s):  
Expected Utility ◽  
Potential Risk ◽  
Simple Game ◽  
Rational Agent ◽  
Rational Agents ◽  
Reward Function ◽  
Special Case

It is clear that one of the primary tools we can use to mitigate the potential risk from a misbehaving AI system is the ability to turn the system off. As the capabilities of AI systems improve, it is important to ensure that such systems do not adopt subgoals that prevent a human from switching the system off. This is a challenge because many formulations of rational agents create strong incentives for self-preservation. This is not caused by a built-in instinct, but because a rational agent will maximize expected utility and cannot achieve whatever objective it has been given if it is dead. Our goal is to study the incentives an agent has to allow itself to be switched off. We analyze a simple game between a human H and a robot R, where H can press R’s off switch but R can disable the off switch. A traditional agent takes its reward function for granted: we show that such agents have an incentive to disable the off switch, except in the special case where H is perfectly rational. Our key insight is that for R to want to preserve its off switch, it needs to be uncertain about the utility associated with the outcome, and to treat H’s actions as important observations about that utility. (R also has no incentive to switch itself off in this setting.) We conclude that giving machines an appropriate level of uncertainty about their objectives leads to safer designs, and we argue that this setting is a useful generalization of the classical AI paradigm of rational agents.

Dimensionality and Reliability Assessment of the Pain Patient Profile Questionnaire

2008 ◽  
Vol 24 (1) ◽  
pp. 22-26 ◽  
Author(s):  
Brian E. McGuire ◽  
Michael J. Hogan ◽  
Todd G. Morrison
Keyword(s):  
Self Report ◽  
Pain Patient ◽  
Fit Indices ◽  
Depression And Anxiety ◽  
Oblique Rotation ◽  
Pain Experience ◽  
Patient Profile ◽  
Adequate Representation ◽  
Pain Patients ◽  
Report Measure

Abstract. Objective: To factor analyze the Pain Patient Profile questionnaire (P3; Tollison & Langley, 1995 ), a self-report measure of emotional distress in respondents with chronic pain. Method: An unweighted least squares factor analysis with oblique rotation was conducted on the P3 scores of 160 pain patients to look for evidence of three distinct factors (i.e., Depression, Anxiety, and Somatization). Results: Fit indices suggested that three distinct factors, accounting for 32.1%, 7.0%, and 5.5% of the shared variance, provided an adequate representation of the data. However, inspection of item groupings revealed that this structure did not map onto the Depression, Anxiety, and Somatization division purportedly represented by the P3. Further, when the analysis was re-run, eliminating items that failed to meet salience criteria, a two-factor solution emerged, with Factor 1 representing a mixture of Depression and Anxiety items and Factor 2 denoting Somatization. Each of these factors correlated significantly with a subsample's assessment of pain intensity. Conclusion: Results were not congruent with the P3's suggested tripartite model of pain experience and indicate that modifications to the scale may be required.

Sign in / Sign up

Export Citation Format

Share Document