Incorporating Communicative Patterns into Ebdi Agents

Yanet Sánchez

, Teresa Coma

, Antonio Aguelo

and Eva Cerezo

Departamento de Informática e Ingeniería de Sistemas, Universidad de Zaragoza, Maria de Luna 1, Zaragoza, Spain

Departamento Ciencias de la Educación, Universidad de Zaragoza, Violante de Hungría 23, Zaragoza, Spain

Departamento de Psicología y Sociología, Universidad de Zaragoza, Violante de Hungría 23, Zaragoza, Spain

Keywords: BDI Agent, Emotions, ECA, Communicative Patterns.

Abstract: This work proposes the use of the well-known Satir communicative patterns in Emotional Belief-Desires-

Intentions frameworks (EBDIs) aimed to support the management of Embodied Conversational Agents

(ECAs). It shows how to include Satir’s model into the ABC-EBDI framework. The framework is based on

the ABC psychological model and considers, not only the behavioural and emotional consequences of events,

but also the underlying beliefs. This has made possible the connection with the Satir model that specifies

facial and body expressions, voice intonation and linguistic structures related to five universal communication

patterns. The consideration of the communication styles makes it possible to link the expressive capabilities

of the agents with the BDI cognitive processing and to manage them an integrated way.

1 INTRODUCTION

Nowadays, one of the main objectives of AI is the

modelling of intelligent agents that reproduce

realistic human behavior. To develop these agents,

structures are needed that are capable of being

flexible and able to cope with the multiple and

simultaneous demands of the internal and external

environment. Among the most well-known, the BDI

(Rao and Georgeff, 1995) cognitive framework

stands out. The BDI framework is very popular due

to its simplicity and robustness for implementing

intelligent agents. It is based on three fundamental

mental attitudes: beliefs (that represent information

about the environment and oneself), desires (that

represent the motivational state of the agent) and

intentions (that represent the selected action plans that

the agent is committed to achieve and that give the

deliberative character to the model). Nowadays, the

focus is being put in the modeling of emotions and

their influence on the cognitive process, pursuing a

more credible and humanlike behavior. Emotional

BDIs (EBDIs) go in that direction. In the last years

EBDIs have emerged that consider not only emotions

https://orcid.org/0000-0002-3532-4086

https://orcid.org/0000-0002-7439-8459

https://orcid.org/0000-0002-8118-1012

https://orcid.org/0000-0003-4424-0770

but other aspects: some works integrate emotions

with mood (Hernández et al., 2004), with personality

(Puica and Florea, 2013) or combine the three of them

(De Rosis et al., 2003), (Neto and da Silva, 2012),

(Alfonso et al., 2014) (Sánchez-López et al, 2019). A

complete state of the art on EBDIs can be found in

(Sánchez-López and Cerezo, 2019).

Some EBDIs (Bevacqua et al., 2010), (Ochs et al.,

2010) and (Becker-Asano and Wachsmuth, 2008),

have been applied to the management of ECAs

(Embodied Conversational Agents). ECAs are

characters usually with human-line appearances

endured with the use of natural language and non-

verbal behaviors (Cassell, 2000). ECAs can be found

in many applications: in medical domains (Bickmore

et al., 2016), in virtual storytelling (Gris et al., 2016)

and role-playing (Emonts et al., 2012), as

interviewers (Nunamaker Jr. et al., 2011), etc. In

(Bevacqua et al., 2010) an interactive storyteller is

presented. In this case, according to its emotional or

mental state, the agent may vary the quality of its

behaviors: it may use more or less extended gestures,

the arms can move at different speeds and with

different acceleration. In (Ochs et al., 2010), the

Sánchez, Y., Coma, T., Aguelo, A. and Cerezo, E.

Incorporating Communicative Patterns into Ebdi Agents.

DOI: 10.5220/0010909300003116

In Proceedings of the 14th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2022) - Volume 1, pages 391-400

ISBN: 978-989-758-547-0; ISSN: 2184-433X

391

application is a 3D talking head that shows facial

expressions according to the intensity of the elicited

emotions, in order to empathize with the user. In

(Becker-Asano and Wachsmuth, 2008) a virtual

human named Max, who plays the role of museum

guide, is developed. Max is capable of showing facial

expressions and body gestures, according to its

cognitive reasoning capabilities and emotional state.

In these and all other applications the goal is to

achieve credible humanlike behaviors. To do so the

consideration not only of cognitive but affective

capabilities is a must, but also a proper use of agent’s

embodiment and expressive channels. And the

question of how to connect the agent affective-

cognitive process with its expressive channels (facial

expressions, gaze, body postures and speech) is still

an open question.

In this paper, we propose to stablish this

connection through the use of a well-known

psychological communication model: the Satir model

(Andreas & Satir, 1991). The model specifies five

universal communication patterns, each of them with

their particular expressions in terms of facial

expressions, body gestures, voice intonation and

linguistic structures. Satir also specifies the type of

thoughts that emerge in each pattern. In fact, these

patterns are not determined by the personality but

directly related to the thoughts that emerge during

interaction. These thoughts can be linked to agent’s

beliefs, making it possible to consider communication

patterns into EBDI frameworks.

The aim of this paper is to propose the

consideration of communicative patterns in EBDI

agents and to show its inclusion in an existing EBDI

framework. The inclusion is based on:

1- Widening the concept of the behavioural

consequences of events considering not only

agent’s intentions but their expression.

2- Establishing the link between the intentions

and their expression through the processing

of agent’s beliefs.

3- Managing the expression of the intentions in

a coordinated way through the use of Satir

communicative patterns.

Satir’s patterns have been incorporated into the

ABC-EBDI framework (Sanchez et al., 2020), one of

the more advanced EBDIs that considers emotions,

mood and personality, and their influence in all the

cognitive processing stages. The framework is the

result of applying a psychotherapeutic model, the

Ellis’s ABC model (Ellis, 1994) intensively used in

the therapeutic ambit, to the EBDI scheme. The

selection of the ABC model is because it allows

modeling not only emotions and actions in adverse

situations but the underlying human beliefs that

conditions human’s thoughts.

The remainder of the paper is organized as

follows. In section 2 the Satir communication model

is presented. In section 3, the integration of the Satir

communication model into the ABC-EBDI is

presented. In section 4, the ECA application being

developed is shown and, finally, conclusions are

presented.

2 SATIR COMMUNICATION

PATTERNS

The Satir model (Andreas & Satir, 1991) is a very

well-known psychological model extensively used in

family/couple therapy but also applied in different

fields (Health, Education, Engineering, etc.). The

model proposes a graphic representation of the

behavior patterns that manifest themselves in

situations of personal conflict, in a way that is easily

observable and still applied (Peters & Das, 2021).

Observing people while communicating, Satir

discovered five universal patterns of communication:

Placating, Blaming, Computing, Distracting, and

Leveling (see Figure 1). These styles are ways of

communicating and every person has those he/she

uses more frequently. But it is dynamic: in the real

word individuals move between the communication

styles rapidly and frequently: a person can be leveling

at one moment, and then blaming in a rapid sequence

depending on how she/he is coping with her/his

internal process.

Satir also noticed that there were three parts of

communication in every transaction, they are: 1. self,

2. other, and 3. context. She noticed that in the first

four, Placating, Blaming, Computing and Distracting,

the three components of self, other, and context are

out of balance. One or two of the components are

considered of higher value than the other or are

excluded. In the fifth communication style, Leveling,

all three -self, other, and context- have equal value

and are in equal balance.

The first four styles are adopted in stress situations

involving the self-esteem. In those situations, most of

the people adopt one of these four communication

styles to hide their feelings:

• Placating: apologetic, eager to please (hides

fear). The person gives the other person

higher value than they do to themselves:

self is devaluated over the other and the

context. Expressions such as “Please do not

ICAART 2022 - 14th International Conference on Agents and Artiﬁcial Intelligence

392

Figure 1: Satir’s patterns: Superreasonable, Irrelevant, Leveling, Placating and Blaming. Adapted from (Andreas & Satir,

1991).

shout on me”, “I will do anything you ask”

are used. Placating searches for other’s love

and acceptance.

• Blaming: fault-finding, critical (hides pain).

What the person wants is more important

than what the other person wants and the

situation they are in. Expressions such as “If

it weren’t for you, I would not be so angry.”

“It’s your fault I hit you. If you had not

provoked me, you would not have gotten

hit.” Blamers give the other person

responsibility for their feelings and life.

They tend towards arguments, threats and

physical violence towards other.

• Superreasonable: computing, abstract

(afraid of feelings). The person shows no

emotions or affect. They tend to have tense

body posture and their responses are

intellectual, authoritative, and reasonable

and come as a lecture to the other person.

Examples of expressions are: “I am calm,

cool, and collected”, “A true man never

expresses his feelings or anger, hurt,

disappointment”, etc.

• Irrelevant: irrelevant, talkative (afraid of

reality). The words do not make sense, the

person talks about something else, changes

the subject, and makes inappropriate jokes.

The individual does not connect to the

context and reality or the other people. He

or she excludes the self, the other and the

context. Their basic message is: I do not

matter, you do not matter, and the situation

does not matter. When asked a question

they often do not answer it directly and may

respond with an irrelevant comment.

The fifth pattern, Leveling, is very different and

distinct from the four other communication styles.

Leveling considers the self, other people and the

context in communications and recognizes each part

as having equal value. The person is aware of his or

her physiological and bodily responses and shares

thoughts and feelings.

For each pattern the model defines several

communication characteristics that allow the

individual’s behaviour to be personalized. Each

pattern defines the body posture, facial expression,

voice and linguistic structure as shown in Table 1 and

Table 2.

Satir also goes into the detail of the type of

thoughts that emerge in each pattern:

• Placating: “I feel like nothing, I'm dead

without him. I'm worth nothing”.

• Blaming: “I am lonely and unsuccessful”.

• Irrelevant: “Nobody understands me, I do

not belong anywhere”.

• Superreasonable: “I feel easily at people’s

mercy”.

Thoughts related to the Leveling pattern are

objective and positive, always oriented towards

achieving the objectives, to flexibility and openness

to change. The individual enjoys the freedom to be

himself/herself and to accept and to love others.

These thoughts are the way of linking both

models, Satir and ABC-EBDI framework, as it is

explained next section.

Incorporating Communicative Patterns into Ebdi Agents

393

Table 1: Body gestures and facial expressions of Satir

communication patterns.

Pattern Body posture Facial

expression

Blaming Pointing with a

finger. Tense,

distorted, flat,

compressed

breath

Tense facial

muscles, pursed

lips, expand

nasal passages,

annoyed look

Placating Body in low

position, as

kneeling begging,

and saying yes to

everything.

Slumped,

swaying, head

turned upwards,

hand held forward

pleadingly

Sad facial

gesture, look to

the floor

Superreasonable Straight body

posture, feeling of

tranquility and

control.

Unanimated, stiff,

non-reactive

Does not finch,

does not show

emotions, direct

gaze

Irrelevant The body goes in

different

directions

(uncoordinated

movements),

joining the knees

in an exaggerated

manner, bending

the shoulders.

Distracted,

lazy,

entertaining,

distracted gaze

Leveling Freedom of

movement

Neutral

expression,

look into the

eyes

3 INTEGRATION OF THE SATIR

MODEL INTO THE ABC-EBDI

To explain the integration, first an overview of the

ABC-EBDI framework is presented.

3.1 ABC-EBDI Overview

The ABC-EBDI framework (Sanchez et al., 2020),

based on Ellis’s ABC model (Ellis, 1994) focuses in

the modelling on how an individual feels, what he/she

thinks and his/her conduct in adverse situations. In the

ABC-EBDI framework, components and processes

have been defined that allow affect (emotions, mood,

personality) and behavioural modelling to be

Table 2: Voice and linguistic structures of Satir

communication patterns.

Pattern Voice Linguistic

structures

Blaming Scream with

a hard voice,

tense, shrill

Universal

quantifiers: all,

everyone, never,

everything, always,

every time. Assumed

causal relationships

(if, then: because)

Use of negative

questions: Why not

do it?

Placating Nasal voice,

with

complacent

tone,

whining,

squeaky,

pressed

Use of restrictions:

if, only, even, at all.

Use of many

subjunctives: could,

would, might,

should, etc.

Superreasonable Monotonous

and dry

Omission of

nominative

arguments. Use the

longest possible

words, even if not

sure of the meanings.

Deletion of reference

indices. Use of

nominalizations. Use

of nouns without

reference indices: it,

one, people etc.

Deletion of the

subject / subject

reference.

Irrelevant Singsong

that clashes

with words,

erratic, fast,

animated

Arbitrary use of all

three language

patterns, missing

references and links,

rapid change of the

other patterns.

Words without

meaning and

irrelevant

Leveling Warm Direct answer

integrated in the logical reasoning process of BDI

agents. Its main cognitive management components

are:

- Beliefs system (B): this comprises what the

agent believes about an event, the

information about himself and the

environment, and the reactive behavior of

the agent. The system starts from three sets:

basic beliefs (B

) (general information of the

ICAART 2022 - 14th International Conference on Agents and Artiﬁcial Intelligence

394

agent about itself and the environment),

context beliefs (B

) (cognitions about an

activating event) and those that represent

operant behaviors (Bop).

- Desires (D): These represent the

motivational state of the agent. The system

starts from two sets of desires: basic desires

): to be happy, not to die and not to suffer,

and context desires (D

- Conduct (C): This is defined as the

behavioral consequences following the ABC

model. It comprises intentions and how they

are performed. The consideration, not only

of the intentions but, of the way of

expressing them opens the door to the

management of the agent’s expressions

(facial expressions, body gestures,

intonation …). Here we propose to handle

these expressions in a coordinated way by

mean of the Satir model.

The general cognitive/affective process is as

follows (see Figure 2): When an event (A) occurs, the

agent perceives (perceive) the environment or its

internal state, and perceptions about A arise. Later on,

beliefs are reviewed (blf_revision). Depending on the

activating event type, three types of beliefs about A

can be obtained: operant behaviors (B

), basic beliefs

) and context beliefs (B

). If beliefs related to

operant behaviors arise, the cognitive process directly

filters (filter) the agent intentions and selects (select),

conduct to finally execute the action (action).

If Context beliefs (B

) arise they are processed

(brf_processing) and classified as irrational/rational

beliefs (B

I/R

After the beliefs revision process,

irrational/rational desires (D

) are obtained

(options). In this stage the event is also evaluated to

know if it is motivationally relevant (according to the

desires) and will therefore elicit emotions: if there is

high motivational relevance the agent will “feel”

emotions and the emotional generation process

(affective management) starts. If desires are irrational

dysfunctional emotions will arise; rational desires

will lead to functional ones. Eight

dysfunctional/functional negative emotions

(anger/annoyance, guilt/remorse, anxiety/concern

and depression/sadness) and three positive emotions

(gratitude, happiness and pride) are modeled. The

Besides, the agent mood (M) is updated according to

the elicited emotions. Afterwards, conducts (C) are

selected (select). The process is influenced by the

irrational/rational character of desires and the

dysfunctional/functional nature of emotions: if

Figure 2: A general overview of the ABC-EBDI

framework.

desires are irrational (D

) and emotions are

dysfunctional, the conducts will be maladaptive (C

)

and, on the contrary, if desires are irrational (D

) and

emotions are functional, the conduct will be adaptive

). As stated before, the agent’s conduct comprises

the intentions to achieve the irrational/rational desires

and how the agent will express them. Finally,

conducts are executed (action).

Incorporating Communicative Patterns into Ebdi Agents

395

During the cognitive-affective process, when

Context beliefs arise, they are identified as one of

Ellis irrational or rational beliefs and further

classified following the classification:

1- Irrational beliefs classification:

• Demandingness (DEM): There are three

types:

- Demands for Comfort (DEM

comfort): those related to comfort

and equity/rights.

- Self-related demands of

achievement and competence

(DEM achievement/ competence):

those related to oneself, one’s

achievements and personal

competence.

- Demands for control (DEM

control): those related to rigid and

dominant conduct.

• Awfulizing (AWF): A catastrophic

evaluation, the worst that can happen.

• Global evaluation-self or other-downing

(GE/SD): An excessively global negative

evaluation about himself, others or the

world in general.

• Low frustration tolerance (LFT): If an

individual tolerates a situation according

to the nature of his or her formulated

desires.

2- Rational beliefs:

• Preferences: These refer to preferentially

desiring something and they are

expressed in the form of "I want" and "I

wish".

• Non-awfulizing (non-AWF): This refers

to a flexible negative evaluation done by

an individual for not having satisfied his

or her preferences.

• Unconditional acceptance (non-GE/SD):

This is the antidote to the irrational

belief/global/critical evaluation (GE/SD).

• High frustration tolerance (non-LFT):

This refers to the tolerance capacity of an

individual when preferences are not met.

In Table 3, examples of thoughts related to each

category are presented. In Table 4 the Ellis’ General

Irrational and Rational Beliefs (first column) and their

categorization (second column) are shown.

Table 3: Example of thoughts related to Irrational/Rational

beliefs categorization.

Irrational/Rational

Category

Examples of thoughts

DEM comfort

I am worried that with this

hanging over me, my

family might not love me

anymore

DEM

achievement/competence

I cannot afford to be weak

in this situation

DEM control I have cancer because of

the kind of life that

circumstances force me to

live".

AWF I cannot do anything; this

is only the beginning of the

end

GE/SD Maybe, because of this,

they will stop loving me

and I couldn't stand it.

LFT I don't feel able to face this

by myself. I will put

myself in their hands.”

Preferences I wish I had no cancer

non-AWF I want to concentrate on

what I have to do now

non-GE/SD Life does not always turn

out as one would wish

non-LFT I am the one who decides

what is the best for me and

I know all the possibilities

The agent conduct is determined by the

irrational/rational cognitive-affective processing that

results in a maladaptive or adaptive behaviour. More

detailed description of the framework can be found in

(Sanchez et al., 2020).

ICAART 2022 - 14th International Conference on Agents and Artiﬁcial Intelligence

396

Table 4: Mapping between Irrational/Rational beliefs (first column) and Satir’s patterns (third column) by means of the Beliefs

categorization (second column). Example thoughts related to Satir’s patterns are also shown.

General Irrational beliefs (IB) Category Satir’s Pattern

IB-I - It is an extreme need, for the adult

human being, to be loved and approved by

every significant person in his environment.

DEM comfort

Global evaluation/self or other-downing

(GE/SD)

Placating - I'm dead without

him/ No one will accept me.

IB-II - To consider myself as a valid person,

I must be very competent, sufficient and

able to achieve anything that I propose.

DEM achievement and competence

Global evaluation/self or other-downing

(GE/SD)

Blaming - I am unsuccessful/I

am unable to do anything

right.

IB-III - People who do not act as "should"

are vile, evil and infamous and should be

punished for their evil.

DEM control

Global evaluation/self or other-downing

(GE/SD)

Superreasonable- I feel easily

at people’s mercy.

IB-IV - It is terrible and catastrophic that

things do not work out as one would like.

DEM achievement and competence

Awfulizing (AWF)

Blaming- I am lonely and

unsuccessful.

IB-V - Human disgrace and discomfort are

brought about by external circumstances,

and people have no ability to control their

emotions.

DEM control

Global evaluation/self or other-downing

(GE/SD)

Superreasonable- I feel easily

at people’s mercy.

IB-VI - If something is or can be dangerous,

I must be terribly worried about it and I

must constantly think about the possibility

of it happening.

DEM comfort

Awfulizing (AWF)

Placating- I feel like nothing.

IB-VII - It is easier to avoid the

responsibilities and difficulties of life than to

confront them.

DEM comfort

Low frustration tolerance (LFT)

Irrelevant- Nobody

understands me, I do not

belong anywhere.

IB-VIII - I must depend on others and need

someone stronger to trust.

DEM comfort

Low frustration tolerance (LFT)

Placating - I feel like nothing.

IB-IX - What happened to me will always

continue affecting me.

DEM comfort

Global evaluation/self or other-downing

(GE/SD)

Placating- I'm worth nothing/I

will never be fine.

IB-X - We must be very concerned about

the problems and disturbances of others.

DEM comfort

Global evaluation/self or other-downing

(GE/SD)

Placating – I am alone/No one

love me.

IB-XI - There is a perfect solution to every

problem and if we do not find it, it would be

catastrophic.

DEM achievement and competence

Secondary - Awfulizing (AWF)

Blaming – It's my fault/ There

is no solution, it's all over.

General rational beliefs (RB) Category Satir’s pattern

RB-I - I accept my own limitations and

mistakes, and my behavior is not

conditioned by the continuous search for

recognition and approval of others.

Preferences

Unconditional acceptance (non-GE/SD)

Leveling – I am strong and

confident in myself.

RB-II - Each person has the right to act

according to his or her criteria, without me

having to expect them to behave according

to what I expect, need, or consider right.

Preferences

High frustration tolerance (non-LFT)

RBIII - Life and its circumstances happen in

an independent way to my needs and it is me

who must adapt to manage the possibilities

and difficulties that arise.

Preferences

Non-awfulizing (non-AWF)

Incorporating Communicative Patterns into Ebdi Agents

397

Figure 3: Moving from a console-based output (left) to an ECA-based application (right).

ICAART 2022 - 14th International Conference on Agents and Artiﬁcial Intelligence

398

3.2 Integration

The Satir model goes into the detail of the type of

thoughts that emerge in each pattern and illustrates

that individuals can respond to stressful experiences

in both adaptive and maladaptive ways (Thompson et

al., 2010). Thoughts related to the four patterns linked

to stress and tension are based on irrational beliefs

(related to dysfunctional emotions and maladaptive

behaviours) and in the case of the Levelling pattern,

they are based on rational beliefs and therefore, linked

to functional emotions and adaptive behaviours.

In the ABC-EBDI when Context beliefs arise,

they are identified as one of Ellis irrational or rational

beliefs and further classified following the

classification explained before. As Satir describes the

thoughts linked to each of the patterns, two of the

authors (psychologists) have proposed a link between

the general irrational beliefs and the first four

communication patterns through the categorization of

the general irrational beliefs shown in the second

column in Table 4. As it can be seen, all rational

beliefs are linked to the Leveling pattern.

In the first implementation of the ABC-EBDI just

text output is considered. A use case consisting of a

bad news scenario in healthcare domain was

simulated (Sanchez et al., 2020). The simulation

reproduces the dialog between the doctor (user) and

the patient (agent). In each step, during the dialog the

system inputs are the doctor entries and the outputs

are the following:

1. What the agent thinks: Characterization of

the irrational/rational beliefs about the event

(cognitive information).

2. What the agent feels: Emotions and current

mood (affective information).

3. How the agent behaves: Agent answer and

conduct.

An ECA-based application is being developed to

make profit of the new conduct information: the

communicative pattern activated. In the ECA-based

application the chat-based interface is substituted by

an ECA (Embodied Conversational Agent) interface

(see Figure. 3). The ECA interface is being

implemented in Unity 3D and communicates with the

simulation engine through sockets.

To render the appropriated agent expression, the

value of the emotions elicited and the communication

patterns activated have to be combined to produce a

credible animation. For the moment emotions elicited

during the simulation condition facial expressions

and voice intonation whereas communicative patterns

modulate body gestures. To do so animations for each

of the five Satir’s patterns have been developed as

well facial animations corresponding to each of the

emotions considered (some of them shown in Figure

3).

4 CONCLUSIONS

EBDI frameworks are now capable of managing the

cognitive and affective consequences of events on

agents. Some of them are being used to support the

management of embodied agents able to express

themselves through facial, body gestures and voice.

How to link this expression with the cognitive

affective management of the agent is still an open

question.

In this paper we propose to make this link through

the use of Satir’s model, a well-known psychological

model that establishes five general communicative

patterns. We have shown its integration into an

advanced EBDI framework, the ABD-EBDI known

by its sophisticated management of agent’s beliefs.

An ECA based application is being developed and

will make it possible to compare simulation results

with previous console-based application to assess the

impact of the new expression channels.

ACKNOWLEDGEMENTS

This work partly funded by the Spanish Science,

Innovation and University Ministry (MCIU), the

National Research Agency (AEI) and the EU

(FEDER) through the RTI2018-096986-B-C31

contract, by the Aragonese Government (Group

T60_20R) and Banco Santander and the Universidad

de Zaragoza through a mobility grant for Latin

American students.

REFERENCES

Alfonso, B., Vivancos, E., & Botti, V. J. (2014). An Open

Architecture for Affective Traits in a BDI Agent.

Proceedings of the 6th ECTA. Part of the 6th IJCCI,

320–325. https://doi.org/10.5220/0005153603200325

Andreas, S., & Satir, V. (1991). Virginia Satir, the patterns

of her magic. Real People Press.

Becker-Asano, C., & Wachsmuth, I. (2008). Affect

simulation with primary and secondary emotions.

International Workshop on Intelligent. http://link.

springer.com/chapter/10.1007/978-3-540-85483-8_2

Bevacqua, E., Prepin, K., Niewiadomski, R., de Sevin, E.,

& Pelachaud, C. (2010). Greta: Towards an interactive

Incorporating Communicative Patterns into Ebdi Agents

399

conversational virtual companion. In Artificial

Companions in Society: perspectives on the Present

and Future (pp. 143–156). http://www.academia.edu/

download/44814243/GRETA__Towards_an_Interacti

ve_Conversati20160417-8112-1jimobq.pdf

Bickmore, T. W., Utami, D., Matsuyama, R., & Paasche-

Orlow, M. K. (2016). Improving Access to Online

Health Information With Conversational Agents: A

Randomized Controlled Experiment. Journal of

Medical Internet Research, 18(1). https://doi.org/

10.2196/jmir.5239

Cassell, J. (2000). Embodied conversational agents. MIT

Press.

De Rosis, F., Pelachaud, C., Poggi, I., Carofiglio, V., & De

Carolis, B. (2003). From Greta’s mind to her face:

Modelling the dynamics of affective states in a

conversational embodied agent. International Journal

of Human-Computer Studies, 59(1), 81–118.

https://doi.org/10.1016/S1071-5819(03)00020-X

Ellis, A. (1994). Reason and emotion in psychotherapy

((Rev. ed.)). Birch Lane.

Emonts, M., Row, R., Johnson, W. L., Thomson, E., De

Silva Joyce, H., Gorman, L. G., & Carpenter, R. (2012).

Integration of Social Simulations into a Task-based

Blended Training Curriculum. Land Warfare

Conference.

Gris, I., Rivera, D. A., Rayon, A., Camacho, A., & Novick,

D. (2016). Young Merlin: an embodied conversational

agent in virtual reality. Proceedings of the 18th ACM

International Conference on Multimodal Interaction -

ICMI 2016, 425–426. https://doi.org/10.1145/

2993148.2998534

Hernández, D. J., Déniz, O., Lorenzo, J., & Hernández, M.

(2004). BDIE: a BDI like Architecture with Emotional

Capabilities. In In American Association for Artificial

Intelligence Spring Symposium.

Neto, A. F. B., & da Silva, F. S. C. (2012). A computer

architecture for intelligent agents with personality and

emotions. Human-Computer Interaction: The Agency

Perspective, 263–285.

Nunamaker Jr., J. F., Derrick, D. C., Elkins, A. C., Burgoon,

J. K., & Patton, M. W. (2011). Embodied

Conversational Agent-Based Kiosk for Automated

Interviewing. Journal of Management Information

Systems, 28(1), 17–48. https://doi.org/10.2753/

MIS0742-1222280102

Ochs, M., Sadek, D., & Pelachaud, C. (2010). A formal

model of emotions for an empathic rational dialog

agent. International Journal of Autonomous Agents and

Multi-Agent Systems, 24(3), 410–440. https://doi.org/

10.1007/s10458-010-9156-z

Puica, M.-A., & Florea, A.-M. (2013). Emotional Belief-

Desire-Intention Agent Model: Previous Work And

Proposed Architecture. International Journal of

Advanced Research in Artificial Intelligence(IJARAI),

2(2).

Rao, A. S., & Georgeff, M. P. (1995). BDI Agents : From

Theory to Practice. Proceedings of the First

International Conference on Multiagent Systems

(ICMAS95), 95, 312–319. https://www.aaai.org/Papers/

ICMAS/1995/ICMAS95-042.pdf

Sánchez-López, Y., & Cerezo, E. (2019). Designing

emotional BDI agents: Good practices and open

questions. Knowledge Engineering Review, 34(E26).

https://doi.org/10.1017/S0269888919000122

Sanchez, Y., Coma, T., Aguelo, A., & Cerezo, E. (2020).

Applying a psychotherapeutic theory to the modeling of

affective intelligent agents. IEEE Transactions on

Cognitive and Developmental Systems, 12(2), 285–299.

https://doi.org/10.1109/TCDS.2019.2911643

Thompson, R. J., Mata, J., Jaeggi, S. M., Buschkuehl, M.,

Jonides, J., & Gotlib, I. H. (2010). Maladaptive coping,

adaptive coping, and depressive symptoms: Variations

across age and depressive state. Behaviour Research

and Therapy, 48(6), 459–466. https://doi.org/10.1016/

j.brat.2010.01.007

Yaghoubzadeh, R., Kramer, M., Pitsch, K., & Kopp, S.

(2013, August). Virtual agents as daily assistants for

older adults or cognitively impaired people. In

International Workshop on Intelligent Virtual Agents

(pp. 79-91). Springer, Berlin, Heidelberg.

Y. Sánchez, T. Coma, A. Aguelo, and E. Cerezo, “ABC-

EBDI: An affective framework for BDI agents,” Cogn.

Syst. Res., vol. 58, pp. 195–216, Dec. 2019, doi:

10.1016/J.COGSYS.2019.07.002

Peters, H. C., & Das, B. (2021). SEFT: A Critical Review

and Call to Action. Journal of Professional

Counseling: Practice, Theory & Research, 48(1), 1-14.

ICAART 2022 - 14th International Conference on Agents and Artiﬁcial Intelligence

400