Deception Detection Research Articles

Development of Non-contact Deception Detection using Facial Expression and Analysis of Significant Time Period

본 연구는 기존의 접촉식 거짓말 탐지의 단점을 해결하기 위하여 비접촉식 거짓말 탐지 기법을 개발하고자 얼굴 표정 변화를 촬영하고 분석하였다. 거짓말 탐지를 위한 영상 데이터 얻고자 모의범죄 실험을 진행하였으며, 긴장정점검사(POT)을 이용하여 데이터를 수집하였다. 거짓말을 할 때의 표정 분석을 위해 합성곱 신경망(CNN) 계열의 stacked hourglass 네트워크 4층을 쌓은 모델을 이용하여 얼굴에서 특징점을 추출하였다. 영상에서 프레임 단위로 추출한 66개의 특징점을 시간 순으로 배치하여 장단기메모리(LSTM) 모델을 이용하여 거짓말 영상과 진실 영상을 분류하는 모델을 개발하였으며, 모델 학습을 진행하고 이를 토대로 평가 데이터를 통한 분류 정확도는 85.19%이었다. 학습이 완료된 모델의 결과를 분석하기 위하여 LSTM 활성도 분석을 진행하였다. 개발한 방법을 이용하여 사람의 얼굴에서의 표정 변화를 이용한 거짓말 탐지의 가능성을 확인할 수 있었으며, 비접촉식 거짓말 탐지에 중요한 시점을 도출하여 추후 비접촉식 거짓말 탐지의 가능성을 확인할 수 있었다.

A deceptive reviews detection model: Separated training of multi-feature learning and classification

The increasing online reviews play an essential role in the e-commerce platform, which profoundly affects the purchase decisions of consumers. However, rampant dishonest sellers manipulate other buyers or robots to post deceptive reviews for profit. Recently, the detection of deceptive reviews has attracted general research attention, which mainly comprises two directions, traditional methods based on statistics and intelligent methods based on neural networks. These methods use a single feature or multiple features for classifier design. To make full use of different features for better feature representation of detecting deceptive reviews, this paper proposes a new feature fusion strategy and verifies its performance by comparing it with other feature fusion strategies. First, we utilize three independent models for feature extraction: the TextCNN, the Bidirectional Gated Recurrent Unit (GRU), and the Self-Attention are used to learn local semantic features, temporal semantic features, and weighted semantic features of reviews, respectively. Secondly, after obtaining different feature representations from the fully connected layers of these three models, we concatenate them together to form the final documental representation. Finally, we use a full connection layer and the sigmoid function to further learn and complete deceptive review detection. Experiments on three balanced and unbalanced in-domain small datasets (hotel, restaurant, doctor) and mixed-domain datasets show that our model is superior to baselines. Experiments on large-scale data with various imbalanced proportions verify the effectiveness of our method. We also analyze the results of different datasets from the perspective of part of speech to improve the model's interpretability.

A Corpus-Based Study of Linguistic Deception in Spanish

In the last decade, fields such as psychology and natural language processing have devoted considerable attention to the automatization of the process of deception detection, developing and employing a wide array of automated and computer-assisted methods for this purpose. Similarly, another emerging research area is focusing on computer-assisted deception detection using linguistics, with promising results. Accordingly, in the present article, the reader is firstly provided with an overall review of the state of the art of corpus-based research exploring linguistic cues to deception as well as an overview on several approaches to the study of deception and on previous research into its linguistic detection. In an effort to promote corpus-based research in this context, this study explores linguistic cues to deception in the Spanish written language with the aid of an automatic text classification tool, by means of an ad hoc corpus containing ground truth data. Interestingly, the key findings reveal that, although there is a set of linguistic cues which contributes to the global statistical classification model, there are some discursive differences across the subcorpora, yielding better classification results on the analysis conducted on the subcorpus containing emotionally loaded language.

Deception and Lie Detection in the Courtroom: The Effect of Defendants Wearing Medical Face Masks

During the COVID-19 pandemic, defendants and witnesses (as well as the prosecution and defense counsel) may wear medical face masks to prevent the spread of the virus. Alternatively, courtrooms proceedings may take place virtually. In this article, we discuss how these deviations from normal procedures may affect jurors' lie detection ability and decision-making. Although research addressing this specific question does not exist, we are able to formulate an informed view based on the extensive deception literature. Since nonverbal signs of deception in the face and body are virtually absent, we conclude that medical face mask-wearing or virtual courtroom proceedings will not hamper jurors' lie detection abilities. If jurors can hear the speech well, they may become better at detecting deception if they pay more attention to speech content, which may occur as a result of mask-wearing in the courtroom.

Deception Detection in Politics: Can Voters Tell When Politicians are Lying?

In this study, we investigate voters’ unaided perceptions of whether politicians are lying. We conduct an experiment in which participants attempt to uncover politicians’ dishonesty by watching videos of their speeches. We find that verbal cues (specifically, the amount of detail in the speech) and general demeanor cues explain the success (failure) of veracity judgments far better than paraverbal and nonverbal cues. We also find evidence of a truth bias—people are more likely to judge statements to be true than false—despite the political setting, where voters might have been more skeptical. However, gender plays a deterministic role for veracity judgments in political context; female politicians are more likely to be judged as honest.

Misinformation vs. Situational Awareness: The Art of Deception and the Need for Cross-Domain Detection.

The world has been afflicted by the rise of misinformation. The sheer volume of news produced daily necessitates the development of automated methods for separating fact from fiction. To tackle this issue, the computer science community has produced a plethora of approaches, documented in a number of surveys. However, these surveys primarily rely on one-dimensional solutions, i.e., deception detection approaches that focus on a specific aspect of misinformation, such as a particular topic, language, or source. Misinformation is considered a major obstacle for situational awareness, including cyber, both from a company and a societal point of view. This paper explores the evolving field of misinformation detection and analytics on information published in news articles, with an emphasis on methodologies that handle multiple dimensions of the fake news detection conundrum. We analyze and compare existing research on cross-dimensional methodologies. Our evaluation process is based on a set of criteria, including a predefined set of performance metrics, data pre-processing features, and domains of implementation. Furthermore, we assess the adaptability of each methodology in detecting misinformation in real-world news and thoroughly analyze our findings. Specifically, survey insights demonstrate that when a detection approach focuses on several dimensions (e.g., languages and topics, languages and sources, etc.), its performance improves, and it becomes more flexible in detecting false information across different contexts. Finally, we propose a set of research directions that could aid in furthering the development of more advanced and accurate models in this field.

Did You Just Lie to Me?Deception Detection in Face to Face versus Computer Mediated Communication

ABSTRACT Individuals’ deception detection ability during either a face-to-face (FtF) interaction or through computer-mediated communication (CMC) was explored under more naturalistic conditions where they were not forewarned that deception may be involved. Participants discussed a social issue either in a room together or by instant messaging from separate rooms. Prior to discussion, some participants were asked to deceive their partner regarding their actual opinion on the issue. Results showed that mode of communication did not influence participants’ deception detection accuracy rate, nor their truth bias. Regardless of mode of communication, deceptive participants experienced the same level of physiological arousal as the non-deceivers. In contrast, deceivers reported experiencing higher levels of anxiety but only in the FtF condition. Findings highlight how for different communication modalities, a multitude of interactive factors may influence deception detection.

Distrust, False Cues, and Below-Chance Deception Detection Accuracy: Commentary on Stel et al. (2020) and Further Reflections on (Un)Conscious Lie Detection From the Perspective of Truth-Default Theory.

OPINION article Front. Psychol., 21 July 2021Sec. Cognitive Science https://doi.org/10.3389/fpsyg.2021.642359

Causal indicators for assessing the truthfulness of child speech in forensic interviews

When interviewing a child who may have witnessed a crime, the interviewer must ask carefully directed questions in order to elicit a truthful statement from the child. The presented work uses Granger causal analysis to examine and represent child–interviewer interaction dynamics over such an interview. Our work demonstrates that Granger Causal analysis of psycholinguistic and acoustic signals from speech yields significant predictors of whether a child is telling the truth, as well as whether a child will disclose witnessing a transgression later in the interview. By incorporating cross-modal Granger causal features extracted from audio and transcripts of forensic interviews, we are able to substantially outperform conventional deception detection methods and a number of simulated baselines. Our results suggest that a child’s use of concreteness and imageability in their language are strong psycholinguistic indicators of truth-telling and that the coordination of child and interviewer speech signals is much more informative than the specific language used throughout the interview.

Use of global trait cues helps to explain older adults’ decrements in detecting children’s lies

PurposePrevious research has established that lie‐detection accuracy decreases with age; however, various mechanisms for this effect have yet to be explored, particularly when examining the detection of children’s lies. The present study investigated if younger and older adults detect children’s lies using different cues (verbal content, verbal auditory, non‐verbal, global traits) to explore if cue usage may help to explain this age‐related decline.MethodA total of 100 younger (18–30 years) and 100 older adults (66–89 years) watched child interview videos (half were truth‐tellers; half were lie‐tellers coached to conceal a transgression). Participants provided veracity judgements (truth vs. lie) and described the cues that they relied on to make their judgements.ResultsOlder adults used marginally significantly fewer verbal content and significantly more global trait cues compared to younger adults. The use of global trait cues partially mediated the age‐related decline in detection accuracy.ConclusionThese results present a partial mechanism for the age‐related decline in deception detection. This can inform psychological theory on how ageing affects perceptions of child witnesses and deception detection abilities.

Detecting Deception from Gaze and Speech Using a Multimodal Attention LSTM-Based Framework

The automatic detection of deceptive behaviors has recently attracted the attention of the research community due to the variety of areas where it can play a crucial role, such as security or criminology. This work is focused on the development of an automatic deception detection system based on gaze and speech features. The first contribution of our research on this topic is the use of attention Long Short-Term Memory (LSTM) networks for single-modal systems with frame-level features as input. In the second contribution, we propose a multimodal system that combines the gaze and speech modalities into the LSTM architecture using two different combination strategies: Late Fusion and Attention-Pooling Fusion. The proposed models are evaluated over the Bag-of-Lies dataset, a multimodal database recorded in real conditions. On the one hand, results show that attentional LSTM networks are able to adequately model the gaze and speech feature sequences, outperforming a reference Support Vector Machine (SVM)-based system with compact features. On the other hand, both combination strategies produce better results than the single-modal systems and the multimodal reference system, suggesting that gaze and speech modalities carry complementary information for the task of deception detection that can be effectively exploited by using LSTMs.

Deception detection in text and its relation to the cultural dimension of individualism/collectivism

AbstractAutomatic deception detection is a crucial task that has many applications both in direct physical and in computer-mediated human communication. Our focus is on automatic deception detection in text across cultures. In this context, we view culture through the prism of the individualism/collectivism dimension, and we approximate culture by using country as a proxy. Having as a starting point recent conclusions drawn from the social psychology discipline, we explore if differences in the usage of specific linguistic features of deception across cultures can be confirmed and attributed to cultural norms in respect to the individualism/collectivism divide. In addition, we investigate if a universal feature set for cross-cultural text deception detection tasks exists. We evaluate the predictive power of different feature sets and approaches. We create culture/language-aware classifiers by experimenting with a wide range of n-gram features from several levels of linguistic analysis, namely phonology, morphology and syntax, other linguistic cues like word and phoneme counts, pronouns use, etc., and token embeddings. We conducted our experiments over eleven data sets from five languages (English, Dutch, Russian, Spanish, and Romanian), from six countries (United States of America, Belgium, India, Russia, Mexico, and Romania), and we applied two classification methods, namely logistic regression and fine-tuned BERT models. The results showed that the undertaken task is fairly complex and demanding. Furthermore, there are indications that some linguistic cues of deception have cultural origins and are consistent in the context of diverse domains and data set settings for the same language. This is more evident for the usage of pronouns and the expression of sentiment in deceptive language. The results of this work show that the automatic deception detection across cultures and languages cannot be handled in unified manners and that such approaches should be augmented with knowledge about cultural differences and the domains of interest.

A New Approach for Deception Detection in Open Domain Text

The gigantic growth of platforms that give individuals the ability to write a review that is visible to everyone and the huge number of documents shared on the internet have triggered the researchers to try to detect if these platforms are trying to mislead and deceive people. There is a crucial need to find ways to automatically identify fake reviews and detect deceptive people or groups. The main aim of this research is to detect deception in open domain text by using a machine learning technique. Several sets of features are used to analyse the text including unigram, part of speech, and production rules. The experimental results showed that combined feature sets of (part of speech and production rules) using the support vector machine classifier achieve the best accuracy, and it clearly improves on the accuracy of the results reported in a previous study.

Scientific basis of the System for Analysis of Validity in Evaluation: The SAVE Metaprotocol.

There are multiple situations, often related to the administrative or judicial field, in which it is necessary to use a healthy skepticism, to question the validity of an assertion, appealing to the evidence that can prove or disprove it. (Shermer, 2008). From a child custody issue to facing a harsh criminal conviction, to applying for a disability pension or obtaining an indemnity in an insurance context; in all of them, there is the opportunity to use deception for one's own benefit, harming a third party, through fraud. As we know, opportunity, together with prior motivation or incentive and subsequent justification, constitute the classic fraud triangle proposed by Cressey (1961).
 This questioning of the validity of the case understood as its accuracy or correspondence with what it pretends to be and independently of its various types, is only possible from a method of analysis based on scientific evidence that benefits from using a system ordered by rules for the investigation - which we know as a protocol (Amezcua, 2000) - as well as a multiple approaches (Campbell and Fiske, 1959) that is proportionate to a conception of the detection and demonstration of deception from the approach of complexity (Cardozo, 2011). If, in addition, such a system was sufficiently flexible to be useful in the daily practice of the various fields in which it may be necessary to use it, it could represent a significant advance in this area.
 These, together with those of Behavior Analysis in Ethology, Criminology, and Psychology, are the initial theoretical bases on which the System of Analysis of Validity in Evaluation (SAVE) is designed, establishing four phases in two domains of multiple and orderly but flexible application, to scientifically question the validity of a case and provide it with consistency and even legal value when appropriate. Although SAVE was born in a clinical context (Domínguez-Muñoz et al., 2014) its main area of knowledge is the study of lying and deception, an area in which there is a large bibliography, somewhat dispersed among various disciplines, which must be incorporated as a source of academic knowledge for its use in the applied field (Domínguez-Muñoz et al., 2017).
 

Sitting in Judgment: How Body Posture Influences Deception Detection and Gazing Behavior.

Body postures can affect how we process and attend to information. Here, a novel effect of adopting an open or closed posture on the ability to detect deception was investigated. It was hypothesized that the posture adopted by judges would affect their social acuity, resulting in differences in the detection of nonverbal behavior (i.e., microexpression recognition) and the discrimination of deceptive and truthful statements. In Study 1, adopting an open posture produced higher accuracy for detecting naturalistic lies, but no difference was observed in the recognition of brief facial expressions as compared to adopting a closed posture; trait empathy was found to have an additive effect on posture, with more empathic judges having higher deception detection scores. In Study 2, with the use of an eye-tracker, posture effects on gazing behavior when judging both low-stakes and high-stakes lies were measured. Sitting in an open posture reduced judges’ average dwell times looking at senders, and in particular, the amount and length of time they focused on their hands. The findings suggest that simply shifting posture can impact judges’ attention to visual information and veracity judgments (Mg = 0.40, 95% CI (0.03, 0.78)).

From a Rigid Cover to an Elastic One Via a Blanket too Short. An Adaptive Polygraph Approach

Abstract The Polygraph test or the Psychophysiological Detection of Deception is a short blanket that cannot cover everything without paying in errors; a clever polygraph examiner and a wise usage of polygraph must make a choice whether to cover the feet or the head with this short blanket and conduct the examination accordingly. But a wiser approach should look to turn the short blanket into an elastic cover that can deal differently with different people and different situations. Following two-three decades of blessed efforts to develop strict standardization in the field, the time has come to start steering the polygraph ship back to greater flexibility and creativity, this time relying on scientific thinking and knowledge. Thus, contrary to the existing trend in the field, I call to drive modern Polygraphy towards developing a scientifically based approach that follows the motto of “Different Things to Different People and Different Situations”. In other words, I call for developing an adaptive approach or Adaptive Polygraphy. The Relevant Issue Gravity (RIG) Theory (Ginton, 2009) is presented here as a major theoretical and practical carrier for evolving and shaping the Adaptive Polygraphy. The article analyzes the current situation and draws some lines to follow in developing an Adaptive Polygraphy approach.

Indirect and Unconscious Deception Detection: Too Soon to Give Up?

The body of text and references exceed the word-limits which was consulted with and accepted by the Editors. As the system does not allow for inserting here texts and references longer than the initial limit. The whole manuscript is submitted as PDF Word documents.

Deception Detection in Group Video Conversations using Dynamic Interaction Networks

Predicting groups of people who are jointly deceptive is critical in settings such as sales pitches and negotiations. Past work on deception in videos focuses on detecting single deceivers and uses facial or visual features only. We propose the concept of Face-to Face Interaction Networks (FFINs) and Negative Interaction Networks (NINs) to model interactions within a group of people. The use of FFINs and NINs in this paper enables us to leverage network relations in predicting face-to-face deception for the first time. We will use a dataset of 185 videos from a deception-based game called Resistance for deception based prediction using FFINs. We first characterize the behavior of individuals, pairs, and groups of deceptive participants compared to non-deceptive participants. Our analysis reveals that less engaged deceivers are identified early. Moreover, pairs of deceivers tend to avoid mutual interaction and focus their attention on non-deceivers. In contrast, non-deceivers interact with everyone equally. We propose the notion of Negative Interaction Networks (NINs) and create a belief propagation neural net algorithm called BPNN based on dynamic FFINs and NINs to detect deceivers from videos that are just 1 minute long. We show that our method outperforms recent state-of-the-art computer vision, graph embedding, and ensemble methods by at least 20.9% AUROC in identifying deception from videos.

Detection of deception through saccadic movements in the psychological assessment and symptom monitoring of Covid-19

From the theory of the model of the mind's point of view, which indicates that in syllogisms there will be more eye movements from front to back between the terms of the premises that question some of Ford's predictions? (Espino & Santamaría, 1998) During the blinking process, sometimes the superior and inferior eyelids tend to completely or incompletely close. It intervenes notably in the comfort of vision when we perform surrounding vision tasks. Carrying out tasks such as conferences, paper, and electronic versions, it's linked to an increase in dry eye symptoms, therefore, among other factors, of altered blinking. (Rodríguez Montiel, 2015). This article intended to verify the phenomena of lying, deception, and self-deception that are directly related to psychological problems and their treatments (Porcel Medina & Gonzalez Fernandez, 2005). Among its defining features is the adoption of computational metaphor as a source of inspiration for the modeling of the structures and processes of the mind, and methodological functionalism, which legitimizes the study of mental processes and states (the "software") regardless of its physical installation base (the "hardware"). It is a computerized model of the human special abilities for reasoning and problem-solving skills. However, it is another modern language used in artificial intelligence (AI) and the logical extension that is based on first-order predicate calculus that includes an inference engine that uses backward chaining. It is a non-procedural language, which indicates that the instructions do not have to necessarily be executed in the order that they have been entered. (Pino Diez & Gómez Gómez, 2001) It is concluded that the specific causes of this ocular muscular activity constitute a true enigma located in the adjacent area of the midbrain. (García Alcolea, 2009).
 The article will be based according to the title, authors, year of publication, abstract, citations, and bibliographic references since it is an experimental design project.

Affect-Aware Machine Learning Models for Deception Detection

Automated deception detection systems can enhance societal well-being by helping humans detect deceivers and support people in high-stakes situations across health, social work, and legal domains. Existing computational approaches for detecting deception have not leveraged dimensional representations of affect, specifically valence and arousal, expressed during communication. My research presents a novel analysis of the potential for including affect in machine learning models for detecting deception. My work informs and motivates the development of affect-aware machine learning approaches for modeling deception and other social behaviors during human interactions in-the-wild. This research, independently defined and conducted by me, is from work-in-progress towards my undergraduate thesis in the Department of Computer Science at the University of Southern California.