Moral Decision-Making by Analogy: Generalizations versus Exemplars

AAAI Conferences

Moral reasoning is important to accurately model as AI systems become ever more integrated into our lives. Moral reasoning is rapid and unconscious; analogical reasoning, which can be unconscious, is a promising approach to model moral reasoning. This paper explores the use of analogical generalizations to improve moral reasoning. Analogical reasoning has already been used to successfully model moral reasoning in the MoralDM model, but it exhaustively matches across all known cases, which is computationally intractable and cognitively implausible for human-scale knowledge bases. We investigate the performance of an extension of MoralDM to use the MAC/FAC model of analogical retrieval over three conditions, across a set of highly confusable moral scenarios.


Blass

AAAI Conferences

Moral reasoning is important to accurately model as AI systems become ever more integrated into our lives. Moral reasoning is rapid and unconscious; analogical reasoning, which can be unconscious, is a promising approach to model moral reasoning. This paper explores the use of analogical generalizations to improve moral reasoning. Analogical reasoning has already been used to successfully model moral reasoning in the MoralDM model, but it exhaustively matches across all known cases, which is computationally intractable and cognitively implausible for human-scale knowledge bases. We investigate the performance of an extension of MoralDM to use the MAC/FAC model of analogical retrieval over three conditions, across a set of highly confusable moral scenarios.


Blass

AAAI Conferences

Autonomous systems must consider the moral ramifications of their actions. Moral norms vary among people and depend on common sense, posing a challenge for encoding them explicitly in a system. I propose to develop a model of repeated analogical chaining and analogical reasoning to enable autonomous agents to interactively learn to apply common sense and model an individual's moral norms.


Analogical Abduction and Prediction: Their Impact on Deception

AAAI Conferences

To deceive involves corrupting the predictions or explanations of others. A deeper understanding of how this works thus requires modeling how human abduction and prediction operate. This paper proposes that most human abduction and prediction are carried out via analogy, over experience and generalizations constructed from experience. I take experience to include cultural products, such as stories. How analogical reasoning and learning can be used to make predictions and explanations is outlined, along with both the advantages of this approach and the technical questions that it raises. Concrete examples involving deception and counter-deception are used to explore these ideas further.


Forbus

AAAI Conferences

We believe that the flexibility and robustness of common sense reasoning comes from analogical reasoning, learning, and generalization operating over massive amounts of experience. Million-fact knowledge bases are a good starting point, but are likely to be orders of magnitude smaller, in terms of ground facts, than will be needed to achieve human-like common sense reasoning. This paper describes the FIRE reasoning engine which we have built to experiment with this approach. We discuss its knowledge base organization, including coarse-coding via mentions and a persistent TMS to achieve efficient retrieval while respecting the logical environment formed by contexts and their relationships in the KB. We describe its stratified reasoning organization, which supports both reflexive reasoning (Ask, Query) and deliberative reasoning (Solve, HTN planner). Analogical reasoning, learning, and generalization are supported as part of reflexive reasoning. To show the utility of these ideas, we describe how they are used in the Companion cognitive architecture, which has been used in a variety of reasoning and learning experiments.