barcodefontsoft.com

Modeling Social Emotions and Social Attributions in .NET Integrated barcode 128 in .NET Modeling Social Emotions and Social Attributions

Modeling Social Emotions and Social Attributions generate, create code128 none in .net projects Visual Studio .NET Introduction directly execu Code 128C for .NET ted by some agent) or an abstract action. An abstract action may be decomposed hierarchically in multiple ways and each alternative consists of a sequence of primitive or abstract sub-actions.

The desirability of action effects (i.e., effects having positive/negative signi cance to an agent) is represented by utility values (Blythe, 1999) and the likelihood of preconditions and effects is represented by probability values.

A non-decision node (or And-node) is an abstract action that can be decomposed only in one way. A decision node (or Or-node), on the other hand, can be decomposed in more than one way. In a decision node, an agent needs to make a decision and select among different options.

If a decision node A can be decomposed in different ways a 1 , a 2 , . . .

a n , we will refer to a 1 , a 2 , . . .

a n as alternatives of each other. Clearly, a primitive action is a non-decision node, whereas an abstract action can be either a non-decision node or a decision node. Consequences or outcomes (we use the terms as exchangeable in this chapter) of actions are represented as a set of primitive action effects.

The consequence set of an action A is de ned recursively from leaf nodes (i.e., primitive actions) in plan structure to an action Aas follows.

Consequences of a primitive action are those effects with non-zero utility, and all the consequences of a primitive action are certain. For an abstract action, if the abstract action is a non-decision node, then the consequence set of the abstract action is the union of the consequences of its sub-actions. If the abstract action is a decision node, we need to differentiate two kinds of consequences.

If a consequence p of a decision node occurs among all the alternatives, we call p a certain consequence of the decision node; otherwise p is an uncertain consequence of the node. In addition, each action step is associated with a performer (i.e.

, the agent that performs the action) and an agent who has authority over its execution. The performer cannot execute the action until authorization is given by the authority. This represents the hierarchical organizational structure of social agents.

4.1.2 Attribution Variables Weiner and Shaver de ne the attribution process in terms of a set of key variables:2 Causality refers to the connection between actions and the effects they produce.

In our approach, causal knowledge is encoded via hierarchical task representation. Interdependencies between actions are represented as a set of causal links and threat relations. Each causal link speci es that an effect of an action achieves a particular goal that is a precondition of another action.

Threat relations specify that an effect of an action threatens a causal link by making the goal unachievable before it is needed.. Note that these models differ in terminology. Here we adopt the terminology of Shaver. Jonathan Gratch, Wenji Mao, and Stacy Marsella Foreseeability refers to an agent s foreknowledge about actions and consequences. We use know and bring-about to represent foreseeability. If an agent knows that an action brings about certain consequence before its execution, then the agent foresees that the action brings about the consequence.

Intention is generally conceived as a commitment to work toward a certain act or outcome. Intending an act (i.e.

, act intention) is distinguished from intending an outcome of an act (i.e., outcome intention) in that the former concerns actions whereas the latter concerns consequences of actions.

Most theories argue that outcome intention rather than act intention is the key factor in determining accountability and intended outcome usually deserves more elevated accountability judgments (Weiner, 1986, 2001). We use intend with do to represent act intention and intend with achieve for outcome intention. Because our work is applied to rich social context, comparing with (Bratman, 1987; Grosz & Kraus, 1996), we include indirect intentions in our work.

For example, an agent intends an action or a consequence, but may not be the actor himself/herself (i.e., by intending another agent to act or achieve the consequence), or an agent intends to act but is coerced to do so.

Similar difference exists in coercion. An agent may be coerced to act (i.e.

, act coercion) yet not be coerced to achieve any outcome of the action (i.e., outcome coercion), depending on whether the agent has choices in achieving different outcomes among alternatives.

It is important to differentiate act coercion and outcome coercion, because it is the latter that actually in uences our judgment of behavior, and is used to determine the responsible agent. We use coerce with do to represent act coercion and coerce with achieve for outcome coercion. In the case of outcome coercion, the responsible agent for a speci c outcome is the performer or the authority of an action, but the action may not be the primitive one that directly leads to the outcome.

4.1.3 Representational Primitives In modeling Shaver and Weiner s attribution theory, we need to map attribution variables into representational features of an agent s causal interpretation.

Here we de ne a number of speci c primitive features that support this mapping. Let x and y be different agents. Let A and B be actions and p a proposition.

The following primitives are adopted in the system: (1) and-node(A): A is a non-decision node in plan structure. (2) or-node(A): A is a decision node in plan structure. (3) alternative(A, B): A and B are alternatives of performing the same higher-level action.

(4) effect(A): Effect set of a primitive action A.. Modeling Socia .net framework Code-128 l Emotions and Social Attributions (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) consequence(A): Certain consequence set of A. performer(A): Performing agent of A.

authority(A): Authorizing agent of A. know(x, p): x knows p. intend(x, p): x intends p.

coerce(y, x, p): y coerces x to achieve the proposition p. want(x, p): x wants p. by(A, p): By acting A to achieve p.

bring-about(A, p): A brings about p. do(x, A): x does A. achieve(x, p): x achieves p.

responsible( p): Responsible agent for p. superior(y, x): y is a superior of x..

4.1.4 Axioms W e identify the interrelations of attribution variables, expressed as axioms.

The axioms are used either explicitly as commonsense inference rules for deriving key attribution values, or implicitly to keep the consistency between different inference rules. Let x and y be different agents. Let A be an action and p a proposition.

The following axioms hold from a rational agent s perspective (To simplify the logical expressions, we omit the universal quanti ers in this chapter, and substitute A for do( , A) and p for achieve( , p) here). (1) y(coerce(y, x, A)) intend(x, A) (2) intend(x, A) ( y(coerce(y, x, A)) p( p consequence( A) intend(x, p)) (3) intend(x, p) A( p consequence(A) intend(x, A)) (4) intend(x, by(A, p)) know(x, bring-about( A, p)) The rst axiom shows that act coercion entails act intention. It means that if an agent is coerced to perform an action A by another agent, then the coerced agent intends A.

3 The second and the third axioms show the relations between act intention and outcome intention. The second one means that if an agent intends an action A and the agent is not coerced to do so (i.e.

A is a voluntary act), then the same agent must intend at least one consequence of A. The third means that if an agent intends a consequence p, the same agent must intend at least one action that has p as a consequence.4 Note that in both axioms, intending an action or a consequence includes.

The notion of Code 128 Code Set B for .NET intention in this axiom is not identical to the typical implication of intention in literatures, as here it is applied to coercive situations. This axiom is not true in general cases, as the agent may not know that an action brings about p.

Here we apply it within the restrictive context of after-action evaluation, where actions have been executed and the consequence has occurred..
Copyright © barcodefontsoft.com . All rights reserved.