1 Concepts and Experiments on Psychoanalysis Driven Computing

2025-04-28 0 0 3.26MB 14 页 10玖币

侵权投诉

Concepts and Experiments on Psychoanalysis

Driven Computing

Minas Gadalla

Department of Computer Engineering & Informatics

University of Patras, Greece

email: gkantalla@ceid.upatras.gr

Sotiris Nikoletseas

University of Patras and CTI, Greece

email: nikole@cti.gr

Jos´

e Roberto de A. Amazonas

Department of Computer Architecture

Technical University of Catalonia - UPC

Barcelona - Spain

email: jose.roberto.amazonas@upc.edu Jos´

e D. P. Rolim

University of Geneva, Switzerland

email: jose.rolim@unige.ch

Abstract—This research investigates the effective incorporation of the

human factor and user perception in text-based interactive media. In

such contexts, the reliability of user texts is often compromised by

behavioural and emotional dimensions. To this end, several attempts

have been made in the state of the art, to introduce psychological

approaches in such systems, including computational psycholinguistics,

personality traits and cognitive psychology methods.

In contrast, our method is fundamentally different since we employ

a psychoanalysis-based approach; in particular, we use the notion of

Lacanian discourse types, to capture and deeply understand real (possi-

bly elusive) characteristics, qualities and contents of texts, and evaluate

their reliability. As far as we know, this is the ﬁrst time computational

methods are systematically combined with psychoanalysis. We believe

such psychoanalytic framework is fundamentally more effective than

standard methods, since it addresses deeper, quite primitive elements

of human personality, behaviour and expression which usually escape

methods functioning at “higher”, conscious layers. In fact, this research

is a ﬁrst attempt to form a new paradigm of psychoanalysis-driven

interactive technologies, with broader impact and diverse applications.

To exemplify this generic approach, we apply it to the case-study of

fake news detection; we ﬁrst demonstrate certain limitations of the well-

known Myers–Briggs Type Indicator (MBTI) personality type method,

and then propose and evaluate our new method of analysing user texts

and detecting fake news based on the Lacanian discourses psychoana-

lytic approach.

1 INTRODUCTION

User-related and generated data (for simplicity referred to

as user data) constitute a core component for social applica-

tions based on interactive media technologies. However, in

various contexts where user perception is more susceptible

to emotional or some other form of bias, user data reliability

is often compromised.

A scenario that accurately exempliﬁes the impact of the

lack of reliability of user data can be illustrated in disaster

management. In a hypothetical event of a car accident on

a frequented highway, where lots of other cars are passing

by, we can imagine loads of social media posts describing

details about it, before any truly reliable information (e.g.

the police arriving) is conveyed. In such a case, if the high

volume/uncertain quality user-based information could be

instantly ﬁltered, so that a reliable source which e.g. accu-

rately reports the degree of passenger injury severity could

be identiﬁed, the timely arrival of an ambulance could be

of life-saving impact. Of course, the above is a non-trivial

task as it implies mechanisms which are yet non-existent;

it is certain, though, that an interdisciplinary approach is

necessary to capture the diverse aspects of human social

way of expressing perception in terms of data and model it

in a formal way, in order to conceptualise such mechanisms

and infer the reliability of the information and knowledge

that can be acquired from such data.

Because of this lack of reliability associated with user

data, interactive media technologies-based applications that

depend on the characterisation and prediction of the in-

formation and knowledge acquired from such data may

be severely impaired and are rarely adopted by the actors

involved in real life scenarios.

A relevant, recently established research topic is on

detecting fake news. In particular, “fake news detection” is

deﬁned as the task of categorising news along a continuum

of veracity, with an associated measure of certainty; veracity

arXiv:2210.00850v1 [cs.CL] 29 Sep 2022

is compromised by the occurrence of intentional deceptions

[1]. The state-of-the-art methods for detecting the spread of

fake news can be coarsely classiﬁed into two categories. The

linguistic approaches are based on “language leakages” that

take place when someone tries to conceal a lie [2]. Here, cer-

tain verbal aspects are monitored, such as frequencies and

patterns of pronouns, conjunctions, and negative emotion

word usage; a task that is found very difﬁcult to achieve.

On the other hand, the network approaches are based on

corresponding properties and behaviour of how the news is

spread. Here, linked data and social network behaviour are

studied [3].

We conjecture that psychoanalysis theories may be used

to provide the tools for a third methodology to be devel-

oped. One may choose to disseminate fake news for several

reasons; e.g., due to being irrational or because there is

something to gain. Independently of the different motives,

certain text qualities characteristic of fake news can be

captured by a psychoanalytic examination of the texts.

The overarching aim of this research work is to develop

a radically new theoretical framework for interpreting user-

generated data in the context of social interactions. In par-

ticular, by combining elements from two very divergent

disciplines – Computer Science and Psychoanalysis – this

work will develop the theoretical methods and tools for

gaining a deep, holistic understanding of the behavioural

context of individuals, groups and crowds from the data

they generate. This work focus on improving and providing

a fundamentally new perspective in terms of the corre-

sponding technologies.

In this context, this research has the ambitious goal

of laying the foundations for a new paradigm of

psychoanalysis-driven technologies.

The speciﬁc contributions of this paper, within the

broader aims stated previously, may be summarised as

follows:

•evaluation of the fake news detection accuracy

method based on the personality traits concept

demonstrating its limitations;

•proposal of a new method to classify user data based

on the Lacanian discourses psychoanalytical concept;

•evaluation of the fake news detection accuracy based

on the Lacanian discourses approach;

•deﬁnition of a framework and roadmap for the future

development of psychoanalysis-driven computing.

After this Introduction, Section 2 describes and com-

pares related works published in the recent years, Section

3 describes the personality traits concept and introduces the

novel psychoanalysis-driven approach based on Lacanian

discourses, Section 4 evaluates the potential of the adopted

Psychological and Psychoanalytic approaches to identify

reliability related characteristics of enunciations, Section 5

presents the computational approach followed, and Section

6 summarises the conclusions and proposes a roadmap for

future work.

2 RELATED WORK

Relevant, yet different, approaches of combining psycho-

logical and social dimensions with computational meth-

ods include computational psycholinguistics, personality

traits, behavioural analysis, emotional states and cognitive

psychology methods. Compared to all those approaches,

our approach is fundamentally different since we adopt a

psychoanalytic perspective; in particular, we employ the

powerful notion of Lacanian discourse types. To the best

of our knowledge, this is the ﬁrst attempt of systemati-

cally bringing together psychoanalysis and computing. We

believe that such a psychoanalytic approach is eventually

more effective compared to the previously mentioned meth-

ods, since it addresses deeper, fundamental elements of

human personality, behaviour and expression which usually

escape methods operating at a “higher” conscious layer.

Having stressed this general novelty of our research

methodology, we below discuss relevant, recent research

related to the particular case study (fake news detection)

which we use in order to exemplify our method.

A fundamental approach of combining psychology with

computational linguistics (based on abstract formulations of

phrases via a collection of ﬁnite-state transition networks)

is described in [4]; in particular, the author envisions so-

phisticated natural language technologies as a key factor for

improving the (rather poor) performance of current conver-

sational systems used by modern technology. The abundant

availability of massive data along with effective AI methods

(including deep learning) is expected to further facilitate this

vision. We note that, although the notion of conversation is

directly relevant to the notion of discourse, the approach

taken in that paper is more limited (psychological) than the

directly psychoanalytic attempt we pursue in our research.

For the more concrete aspect of detecting misinforma-

tion in online social networks, [5] suggests the application

of cognitive psychology concepts. An efﬁcient algorithm

for detection of spread of misinformation in Twitter is

proposed, based on text and network-wide qualities such

as the consistency of message, the coherency of message,

the credibility of source and the general acceptability of

message in the network. Again, no psychoanalytic elements

are considered when evaluating the qualitative properties

of the text. Also, the use of objective, global information

is employed, in contrast to our approach which focuses

on each text separately (however, our method can also be

extended to include global information about texts).

Psychological factors (in particular, emotions as ex-

pressed in Reddit conversations) are addressed in [6], to

propose a model for passively detecting mental disorders.

The suggested model is based entirely on emotional states

and the transitions between these states identiﬁed in Reddit

posts, in contrast to content-based representations (e.g., n-

grams, language model embeddings etc.) in the relevant

state of the art. The scope is to overcome the domain and

topic bias of content-based representations, towards more

general applicability. Our approach aims to avoid a content-

speciﬁc bias, focusing on underlying qualities of texts cap-

tured by the discourse type identiﬁcation. In fact, discourse

types are an even more “primitive” aspect of texts, more so

than emotions; so the generality of our method might be

broader.

In another psychology-based research for fake news

detection, a behavioural analysis approach is taken [7]. In

particular, the authors use supervised learning algorithms

to proﬁle fake-news spreaders, based on the combination

of Big Five personality traits and stylometric features. The

method is evaluated on a tweeter dataset in both English

and Spanish languages. In a similar spirit, [8] aims to

understand the motivations for sharing fake news and the

corresponding personality traits among social media users

in India; in particular, the ﬁndings suggest that the passing

time, information sharing and socialisation gratiﬁcations

lead to instant sharing news. Also, people who exhibit ex-

traversion, neuroticism and openness share news on social

media platforms instantly; in contrast, agreeableness and

conscientiousness personality traits lead to authenticating

news before sharing. We note that that work focuses on

fake news spreading, not fake news identiﬁcation itself.

Also, as stated previously, although those works exhibit a

psychology-based ﬂavour too, our own research is based

on psychoanalytic methods, particularly for identifying the

power of characteristic Lacanian discourse types to identify

speciﬁc characteristics in texts in interactive media as, for

example, the fake news identiﬁcation that we selected to

develop the case study of this work.

3 PSYCHOLOGICAL AND PSYCHOANALYTIC AP-

PROACHES TAKEN

In this section, we describe two distinct approaches to po-

tentially infer reliability-related information from user data.

The ﬁrst one is based on the psychological concept of per-

sonality traits, and the second one on the psychoanalytical

concept of Lacanian discourses.

3.1 Personality Type Prediction

Two potential answers to the question ”What makes peo-

ple unique and different from one another?” are motives

or traits [9]. Regarding motives, we can ﬁnd the work

”Motivation and personality: Handbook of thematic content

analysis” [10] and Murray’s list of needs [11]. However,

no computational approach has been implemented or pro-

posed. Personality traits and types, on the other hand, have

numerous theories, taxonomies, and frameworks available,

of which the most well known are the Big Five [12] and the

Myers–Briggs-Type-Indicator (MBTI) [13]. More speciﬁcally,

the trait approach, rather than the motives approach, is still

the most widely used and most accepted conceptual frame-

work for describing human personality and successfully

predicting human behavior, and has been implemented

and applied multiple times in computer-based applications,

such as in social media advertising [14], trait prediction from

facial images [15], and many others. Both these systems are

intended to indicate or predict the subject’s personality, and

thus his or her behavior and preferences, by breaking them

down into many dimensions, or groups. In this research, we

are focusing mainly on the traits approach and particularly

the MBTI system.

The Big Five personality traits comprise the following:

extraversion, conscientiousness, agreeableness, openness to

new experiences, and neuroticism [12]. Meanwhile, the

MBTI personality model describes an individual’s prefer-

ences/behaviour in four dimensions/groups which are also

illustrated in Figure 1:

•Extraversion-Introversion: This dimension measures

how a person gets their energy. Extraverts are ener-

gized by being around people, while introverts are

energized by being alone.

•Sensing - Intuition: This dimension measures how a

person takes in information. Sensors rely on their ﬁve

senses to gather information, while Intuitives rely on

their gut feelings and intuition

•Thinking- Feeling: This dimension measures how

a person makes decisions. Thinkers use logic and

reason to make decisions, while Feelers use their

emotions and values.

•Judging - Perceiving: This dimension measures how

a person lives their life. Judgers like structure and

order, while Perceivers are more ﬂexible and sponta-

neous.

Fig. 1. Personality types groups

This way, each individual is classiﬁed in terms of one of

16 possible four-letter codes, like ESFJ, indicating a person

who may:

•Extraverted (E): is more concerned with the outside

world of people and things than with the inner world

of ideas

•Sensing (S): would rather work with known facts

and solid experience than explore possibilities or

meanings

•Feeling (F): steps into situations to weigh human

values and motives. Prefers to make decisions on the

basis of values.

•Judging (J): prefers a planned, determined orderly

way of life to a ﬂexible spontaneous.

MBTI was developed by Isabel Briggs Myers in the

1940s [13] to implement Jung’s [16] theory of psychological

types and the results from such an approach are obtained

classically from test booklets, answer sheets, and score keys,

and are produced professionally [17] [13]. In Sections 4

and 5 we discuss and use some computational ways of

determining personality types using semantic and linguistic

analysis and present the results from a Fake/Real news

detection application.

3.2 Lacanian Discourses

The Four Discourses theory constitutes an attempt of for-

malisation of the different ways people relate to each other

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

1ConceptsandExperimentsonPsychoanalysisDrivenComputingMinasGadallaDepartmentofComputerEngineering&InformaticsUniversityofPatras,Greeceemail:gkantalla@ceid.upatras.grSotirisNikoletseasUniversityofPatrasandCTI,Greeceemail:nikole@cti.grJos´eRobertodeA.AmazonasDepartmentofComputerArchitectureTechnicalUn...

展开>> 收起<<

1 Concepts and Experiments on Psychoanalysis Driven Computing.pdf

共14页,预览3页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

1 Concepts and Experiments on Psychoanalysis Driven Computing

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: