2. Introduction Interested in the functional benefits of emotion for a cognitive agent Appraisal theories of emotion PEACTIDM theory of cognitive control Use emotion as a reward signal to a reinforcement learning agent Demonstrates a functional benefit of emotion Provides a theory of the origin of intrinsic reward 2
3. Outline Background Integration of emotion and cognition Integration of emotion and reinforcement learning Implementation in Soar Learning task Results 3
4. Appraisal Theories of Emotion A situation is evaluated along a number of appraisal dimensions, many of which relate the situation to current goals Novelty, goal relevance, goal conduciveness, expectedness, causal agency, etc. Appraisals influence emotion Emotion can then be coped with (via internal or external actions) Situation Goals Appraisals Coping Emotion 4
7. Unification of PEACTIDM and Appraisal Theories 7 Perceive Raw Perceptual Information Environmental Change Encode Motor Suddenness Unpredictability Goal Relevance Intrinsic Pleasantness Stimulus Relevance Motor Commands Prediction Outcome Probability Attend Decode Causal Agent/Motive Discrepancy Conduciveness Control/Power Stimulus chosen for processing Action Comprehend Intend Current Situation Assessment
8. Distinction between emotion, mood, and feeling(Marinier & Laird 2007) Emotion: Result of appraisals Is about the current situation Mood: “Average” over recent emotions Provides historical context Feeling: Emotion “+” Mood What agent actually perceives 8
9. Emotion, mood, and feeling Cognition Active Appraisals Perceived Feeling Emotion Feeling Combination Function Pull Mood Decay 9
14. Learning task: Encoding 14 North Passable: false On path: false Progress: true East Passable: false On path: true Progress: true West Passable: false On path: false Progress: true South Passable: true On path: true Progress: true
15. Learning task: Encoding & Appraisal 15 North Intrinsic Pleasantness: Low Goal Relevance: Low Unpredictability: High East Intrinsic Pleasantness: Low Goal Relevance: High Unpredictability: High West Intrinsic Pleasantness: Low Goal Relevance: Low Unpredictability: High South Intrinsic Pleasantness: Neutral Goal Relevance: High Unpredictability: Low
16. Learning task: Attending, Comprehending & Appraisal 16 South Intrinsic Pleasantness: Neutral Goal Relevance: High Unpredictability: Low Conduciveness: High Control: High …
22. Discussion Agent learns both internal (tasking) and external (movement) actions Emotion allows for more frequent rewards, and thus learns faster than standard RL Mood “fills in the gaps” allowing for even faster learning and less variability 22
23. Conclusion & Future Work Demonstrated computational model that integrates emotion and cognitive control Confirmed emotion can drive reinforcement learning We have already successfully demonstrated similar learning in a more complex domain Would like to explore multi-agent scenarios 23
24. 24 HIGH INTENSITY alert tense excited nervous elated stressed happy upset NEGATIVE VALENCE POSITIVE VALENCE sad contented depressed serene lethargic relaxed fatigued calm LOW INTENSITY Circumplex models Emotions can be described in terms of intensity and valence, as in a circumplex model: Adapted from Feldman Barrett & Russell (1998)
25. Computing Feeling from Emotion and Mood 25 Assumption: Appraisal dimensions are independent Limited Range: Inputs and outputs are in [0,1] or [-1,1] Distinguishability: Very different inputs should lead to very different outputs Non-linear: Linearity would violate limited range and distinguishability
26. Computing Feeling Intensity 26 Motivation: Intensity gives a summary of how important (i.e., how good or bad) the situation is Limited range: Should map onto [0,1] No dominant appraisal: No single value should drown out all the others Can’t just multiply values, because if any are 0, then intensity is 0 Realization principle: Expected events should be less intense than unexpected events
Notes de l'éditeur
Be careful about how say agent generates appraisal values
Say prediction is our extension
A cognitive architecture is a set of task-independent mechanisms that interact to give rise to behavior.
In this environment, the agent’s sensing is limited: it can only see the cells immediately adjacent to it in the four cardinal directions. The agent has a sensor that tells it its Manhattan distance to the goal. However, the agent has no knowledge as to the effects of its actions, and thus cannot evaluate possible actions relative to the goal until it has actually performed them. Even then, it cannot always blindly move closer to the goal because given the shape of the maze, it must sometimes increase its Manhattan distance to the goal in order to make progress in the maze.
Mention relaxation and direction
15 episodes50 trialsCutoff at 10kdcsmedian
1st and 3rd quartiles shownReach optimality at the same time, but mood is less variable
This is an extension of previous workThese constraints define a set of equations. This is one possible equation which improves previous work that seems to work well for our current models.
This is an extension of previous workUnifies intensity for all feelings in one equation (others use different equations for each “kind” of feeling)Again these constraints define a set of possible functions, of which this is one that seems to work well for us