Attribution Task Research Articles

Distilled and fused style embedding for authorship attribution with weighted multi-kernel ensemble model

In this age of Artificial Intelligence, the dispute over ownership of the web contents are increasing. In forensic studies and plagiarism cases, determining the author of a document undoubtedly plays a crucial role. The task of Authorship Attribution (AA) involves learning the style of the writers based on available documents to predict the ownership of unknown documents. Hence, capturing the style of writing is primary besides challenging. In this work, key features are systematically extracted, multiple deep-learning models are trained on various aspects of the text, and two novelties: Distilled and Fused Style Embedding (DFSE) and a multi-kernel ensemble model with kernels weighted based on Z-scores are proposed. The performance of the proposed model is superior to the baseline model and many contemporary models for ingenre and multi-topic datasets.

Indonesian news article authorship attribution multilabel multiclass classification using IndoBERT

<p class="p1">Recent developments in technology have made it easier to produce digital con- tent, especially textual articles. But, it has a negative impact in the form of a rising public skepticism of digital data due to plagiarism. Indonesia, one of the world’s most populous countries, is not resistant to this problem. To resolve it, the authorship attribution (AA) task must be executed. However, there has been little investigation on AA for Indonesian articles. As a result, this research applies the AA task to an Indonesian digital news articles dataset. Continuing the previous research, dataset modification was carried out to increase data com- plexity by adding a new class, namely the author’s gender, and also by balancing the distribution of data versus labels to minimize potential overfitting, and model hyper-parameter configurations were carried out to enhance the results gained. This research successfully applied the IndoBERT model to the Indonesian AA task, yielding results in the form of precision = 0.92, recall = 0.90, and F1-score = 0.91. These results indicate that the Indonesian AA task has a lot of potential for development since it identifies writing patterns that may benefit the forensic field, detect plagiarism, and analyze Indonesian texts.</p>

Does ChatGPT have a typical or atypical theory of mind?

In recent years, the capabilities of Large Language Models (LLMs), such as ChatGPT, to imitate human behavioral patterns have been attracting growing interest from experimental psychology. Although ChatGPT can successfully generate accurate theoretical and inferential information in several fields, its ability to exhibit a Theory of Mind (ToM) is a topic of debate and interest in literature. Impairments in ToM are considered responsible for social difficulties in many clinical conditions, such as Autism Spectrum Disorder (ASD). Some studies showed that ChatGPT can successfully pass classical ToM tasks, however, the response style used by LLMs to solve advanced ToM tasks, comparing their abilities with those of typical development (TD) individuals and clinical populations, has not been explored. In this preliminary study, we administered the Advanced ToM Test and the Emotion Attribution Task to ChatGPT 3.5 and ChatGPT-4 and compared their responses with those of an ASD and TD group. Our results showed that the two LLMs had higher accuracy in understanding mental states, although ChatGPT-3.5 failed with more complex mental states. In understanding emotional states, ChatGPT-3.5 performed significantly worse than TDs but did not differ from ASDs, showing difficulty with negative emotions. ChatGPT-4 achieved higher accuracy, but difficulties with recognizing sadness and anger persisted. The style adopted by both LLMs appeared verbose, and repetitive, tending to violate Grice's maxims. This conversational style seems similar to that adopted by high-functioning ASDs. Clinical implications and potential applications are discussed.

Automatic authorship attribution in Albanian texts.

Automatic authorship identification is a challenging task that has been the focus of extensive research in natural language processing. Regardless of the progress made in attributing authorship, the need for corpora in under-resourced languages impedes advancing and examining present methods. To address this gap, we investigate the problem of authorship attribution in Albanian. We introduce a newly compiled corpus of Albanian newsroom columns and literary works and analyze machine-learning methods for detecting authorship. We create a set of hand-crafted features targeting various categories (lexical, morphological, and structural) relevant to Albanian and experiment with multiple classifiers using two different multiclass classification strategies. Furthermore, we compare our results to those obtained using deep learning models. Our investigation focuses on identifying the best combination of features and classification methods. The results reveal that lexical features are the most effective set of linguistic features, significantly improving the performance of various algorithms in the authorship attribution task. Among the machine learning algorithms evaluated, XGBoost demonstrated the best overall performance, achieving an F1 score of 0.982 on literary works and 0.905 on newsroom columns. Additionally, deep learning models such as fastText and BERT-multilingual showed promising results, highlighting their potential applicability in specific scenarios in Albanian writings. These findings contribute to the understanding of effective methods for authorship attribution in low-resource languages and provide a robust framework for future research in this area. The careful analysis of the different scenarios and the conclusions drawn from the results provide valuable insights into the potential and limitations of the methods and highlight the challenges in detecting authorship in Albanian. Promising results are reported, with implications for improving the methods used in Albanian authorship attribution. This study provides a valuable resource for future research and a reference for researchers in this domain.

Variability of theory of mind versus pragmatic ability in typical and atypical development

IntroductionNumerous studies have linked deficits in Theory of Mind (ToM) with language problems. We aimed to explore the similarities and differences between children's patterns of performance on a ToM task that requires minimal linguistic skill and a pragmatic inference task that relies on both ToM and language. We assessed variability in pragmatic inference skills and ToM across populations of children (8–14 years) displaying varying cognitive profiles. We further compared the sensitivity of ToM versus pragmatic ability to core language skills, memory and executive functioning (EF). MethodToM was tested using the Social Attribution Task (SAT-MC-II). Pragmatic ability was assessed in an implicature comprehension task. Receptive vocabulary, grammar comprehension, short-term and working memory (STM and WM) capacity and EF were measured using Hungarian adaptations of standard tasks and tests developed by the authors’ lab. In addition to typically developing (TD) children (n = 33), we included children with neurodevelopmental disorders where ToM and/or language abilities are vulnerable: autism spectrum disorder (ASD, n = 26), attention deficit hyperactivity disorder (ADHD, n = 25) and developmental language disorder (DLD, n = 18). ResultsResults revealed a significant but only moderate positive correlation between pragmatic inference and ToM indicating that the two abilities are related but distinct. The ASD group showed impairments in both ToM and pragmatic inference ability but no significant deficit was observed in ADHD or DLD relative to TD children in either skill. However, while SAT-MC-II results were only affected by verbal WM and vocabulary measures, pragmatic performance was associated with STM, verbal WM, EF, grammatical skills and vocabulary. ConclusionOur findings indicate that disentangling the contributions of different cognitive skills to ToM tasks may help clarify the role of ToM in language skills and identify distinct patterns of ToM and pragmatic skills in developmental disorders.

Inference through innovation processes tested in the authorship attribution task

Urn models for innovation capture fundamental empirical laws shared by several real-world processes. The so-called urn model with triggering includes, as particular cases, the urn representation of the two-parameter Poisson-Dirichlet process and the Dirichlet process, seminal in Bayesian non-parametric inference. In this work, we leverage this connection to introduce a general approach for quantifying closeness between symbolic sequences and test it within the framework of the authorship attribution problem. The method demonstrates high accuracy when compared to other related methods in different scenarios, featuring a substantial gain in computational efficiency and theoretical transparency. Beyond the practical convenience, this work demonstrates how the recently established connection between urn models and non-parametric Bayesian inference can pave the way for designing more efficient inference methods. In particular, the hybrid approach that we propose allows us to relax the exchangeability hypothesis, which can be particularly relevant for systems exhibiting complex correlation patterns and non-stationary dynamics.

An Empirical Investigation of the Relationships Among Self-Esteem, Depression and Self-Serving Bias in People with Internet Gaming Disorder.

People are generally characterized by a self-serving bias which describes the tendency to ascribe positive outcomes or success to internal or personal causes (self-enhancement motivation) and ascribe negative outcomes or failure to external or situational causes (self-protection motivation). It has been found that the individuals with internet gaming disorder (IGD) who have low self-esteem and high depression exhibit an attenuated self-serving bias. However, the relationships among self-esteem, depression and self-serving bias are not clearly identified. A sample of 138 IGD participants completed self-esteem and depression scales and a causal attribution task (Study 1) to examine the relationships among self-esteem, depression and self-serving bias (both self-enhancement and self-protection). In follow-up Study 2, 28 IGD participants were recruited to undertake self-affirmation intervention which can affirm one's sense of global self-view and bolster self-esteem to explore whether self-affirmation would trigger a reduction of depression and a raise of self-serving bias. The results of path analysis in Study 1 showed that the self-serving bias was predicted by self-esteem and depression, and the depression played a mediating role between self-esteem and self-serving bias. The results of Study 2 showed that the IGD participants reported higher self-esteem, lower depression and engaged in more self-protection motivation after affirming-self manipulation as compared with affirming-other manipulation. These findings suggest that self-esteem predicts self-serving bias through depression and self-affirmation could trigger an increase of self-esteem, further decrease depression and improve self-serving bias for the individuals with IGD. The present article clearly identified the relationships among these factors and provided a new approach to promote positive self-concept in individuals with IGD. Future research is warranted to explore the lasting benefits of self-affirmation on domains of education, relationships and gaming withdrawal for the individuals with IGD among different populations.

Evaluating gender-based effectiveness within multi-attribute task battery studies

The prevalent notion that females outperform males in multitasking is challenged by scrutinizing the cognitive abilities of each gender across various demand scenarios (Buser & Peter, 2012; Lui et al., 2020; Szameitat et al., 2015). Investigating cognitive performance at two levels of workload defined as high- and low-demand situations, this research delves into task-specific variations, cognitive strategies, and adaptability to multitasking challenges. Sixty participants in total, with thirty participants in each gender group, engaged in tasks encompassing resource management, tracking, monitoring, and communication using the U.S. Army Aeromedical Research Laboratory's Multi- Attribute Task Battery simulation software. Statistical variance analysis demonstrated notable task discrepancies across different difficulty levels (DL). However, the gender-based analysis indicated non-significant differences in given tasks, except communication tasks. Findings indicate that men tend to perform well in tasks requiring communication and target tracking, benefiting from auditory cues, while women excel in fuel management, displaying strong decision-making abilities. In assessments of visual and spatial perception, females outperformed males at low difficulty levels, whereas males showed greater proficiency at high difficulty levels. This paper advocates for a balanced understanding and approach of cognitive abilities in diverse scenarios.

Significance of Single-Interval Discrete Attributes: Case Study on Two-Level Discretisation

Supervised discretisation is widely considered as far more advantageous than unsupervised transformation of attributes, because it helps to preserve the informative content of a variable, which is useful in classification. After discretisation, based on employed criteria, some attributes can be found irrelevant, and all their values can be represented in a discrete domain by a single interval. In consequence, such attributes are removed from considerations, and no knowledge is mined from them. The paper presents research focused on extended transformations of attribute values, thus combining supervised with unsupervised discretisation strategies. For all variables with single intervals returned from supervised algorithms, the ranges of values were transformed by unsupervised methods with varying numbers of bins. Resulting variants of the data were subjected to selected data mining techniques, and the performance of a group of classifiers was evaluated and compared. The experiments were performed on a stylometric task of authorship attribution.

Crossing Linguistic Barriers: Authorship Attribution in Sinhala Texts

Authorship attribution involves determining the original author of an anonymous text from a pool of potential authors. The author attribution task has applications in several domains, such as plagiarism detection, digital text forensics, and information retrieval. While these applications extend beyond any single language, existing research has predominantly centered on English, posing challenges for application in languages such as Sinhala due to linguistic disparities and a lack of language processing tools. We present the first comprehensive study on cross-topic authorship attribution for Sinhala texts and propose a solution that can effectively perform the authorship attribution task even if the topics within the test and training samples differ. Our solution consists of three main parts: (i) extraction of topic-independent stylometric features, (ii) generation of a small candidate author set with the help of similarity search, and (iii) identification of the true author. Several experimental studies were carried out to demonstrate that the proposed solution can effectively handle real-world scenarios involving a large number of candidate authors and a limited number of text samples for each candidate author.

Overshadowing and salience attribution in relation to cannabis use

Aberrant attentional salience has been implicated in the cannabis-psychosis association. Here, history and frequency of cannabis use were examined against changes in overshadowing (OS), a cue competition paradigm that involves salience processing. Additionally, we examined the association between OS and alternative measures of aberrant salience, as well as schizotypy, in a non-clinical adult sample.280 participants completed an online geometry learning-based OS task, while a subset (N = 149) also completed the Salience Attribution Task (SAT) measure of aberrant salience. All completed the Schizotypal Personality Questionnaire (SPQ), Aberrant Salience Inventory (ASI), and the modified Cannabis Experience Questionnaire (CEQmv). Differences across OS and SAT performance stages and between cannabis use groups were assessed using mixed ANOVAs. Multiple regression and correlational analyses assessed the relationships between OS and SAT task metrics and SPQ and ASI subscale scores.Current cannabis users had significantly lower OS scores during the testing phase relative to those who do not use cannabis, at medium effect sizes. Schizotypy or ASI scores did not mediate this relationship. In the SAT, current cannabis users presented significantly higher implicit aberrant salience relative to non-users. Scores in the first training phase of the OS task significantly predicted higher explicit aberrant and adaptive salience scores in the SAT.These data indicate an association between regular cannabis use and abnormalities in cue competition effects in a healthy adult sample. Comparisons of OS and SAT cast new light on putative overlapping mechanisms underlying performance across different measures of salience.

Importance of Characteristic Features and Their Form for Data Exploration.

The nature of the input features is one of the key factors indicating what kind of tools, methods, or approaches can be used in a knowledge discovery process. Depending on the characteristics of the available attributes, some techniques could lead to unsatisfactory performance or even may not proceed at all without additional preprocessing steps. The types of variables and their domains affect performance. Any changes to their form can influence it as well, or even enable some learners. On the other hand, the relevance of features for a task constitutes another element with a noticeable impact on data exploration. The importance of attributes can be estimated through the application of mechanisms belonging to the feature selection and reduction area, such as rankings. In the described research framework, the data form was conditioned on relevance by the proposed procedure of gradual discretisation controlled by a ranking of attributes. Supervised and unsupervised discretisation methods were employed to the datasets from the stylometric domain and the task of binary authorship attribution. For the selected classifiers, extensive tests were performed and they indicated many cases of enhanced prediction for partially discretised datasets.

Oxytocin modulates neural activity during early perceptual salience attribution

Leading hypotheses of oxytocin’s (OT) role in human cognition posit that it enhances salience attribution. However, whether OT exerts its effects predominantly in social (vs non-social) contexts remains debatable, and the time-course of intranasal OT’s effects’ on salience attribution processing is still unknown. We used the social Salience Attribution Task modified (sSAT) in a double-blind, placebo-controlled intranasal OT (inOT) administration, between-subjects design, with 54 male participants, to test existing theories of OT’s role in cognition. Namely, we aimed to test whether inOT would differently affect salience attribution processing of social stimuli (expressing fearfulness) and non-social stimuli (fruits) made relevant via monetary reinforcement, and its neural processing time-course. During electroencephalography (EEG) recording, participants made speeded responses to emotional social (fearful faces) and non-emotional non-social (fruits) stimuli - which were matched for task-relevant motivational salience through their (color-dependent) probability of monetary reinforcement. InOT affected early (rather than late, P3b and LPP) EEG components, increasing N170 amplitude (p = .041) and P2b latency (p .001; albeit not of P1), regardless of stimuli’s (emotional) socialness or reinforcement probability. Fear-related socialness affected salience attribution processing EEG (p .05) across time (N170, P2b and P3b), being later modulated by reinforcement probability (LPP). Our data suggest that OT’s effects on neural activity during early perception, may exist irrespective of fear-related social- or reward-contexts. This partially supports the tri-phasic model of OT (which posits OT enhances salience attribution in an early perception stage regardless of socialness), and not the social salience nor the general approach-withdrawal hypotheses of OT, for early salience processing event-related potentials.

Authorship Attribution on Short Texts in the Slovenian Language

The study investigates the task of authorship attribution on short texts in Slovenian using the BERT language model. Authorship attribution is the task of attributing a written text to its author, frequently using stylometry or computational techniques. We create five custom datasets for different numbers of included text authors and fine-tune two BERT models, SloBERTa and BERT Multilingual (mBERT), to evaluate their performance in closed-class and open-class problems with varying numbers of authors. Our models achieved an F1 score of approximately 0.95 when using the dataset with the comments of the top five users by the number of written comments. Training on datasets that include comments written by an increasing number of people results in models with a gradually decreasing F1 score. Including out-of-class comments in the evaluation decreases the F1 score by approximately 0.05. The study demonstrates the feasibility of using BERT models for authorship attribution in short texts in the Slovenian language.

The Effect of Paternal and Maternal Helicopter Parenting on the Career Preparation Behavior of High-School Students: Sequential Mediation Effects of Academic Achievement Attribution and Career Decision-Making Self-Efficacy

This study examined the sequential mediation effects of academic achievement attribution and career decision-making self-efficacy on the effect of paternal and maternal helicopter parenting on high-school students' career preparation behavior. A total of 285 (119 male and 166 female) Korean high-school students in the second grade participated in the study. Research variables were measured using the Career Preparation Behavior Scale (Kim, 1997), Helicopter Parenting Scale (LeMoyne & Buchanan, 2011), Attribution Questionnaire (Weiner, 1979), and Career Decision-Making Self-Efficacy Scale-Short Form (Betz et al., 1996). To examine the sequential mediating effect, data analysis was performed using SPSS 29.0 and PROCESS MACRO (v4.2) Model 6. The results revealed no correlations between helicopter parenting and academic failure attribution. However, higher paternal and maternal helicopter parenting were found to indirectly reduce high-school students' career preparation behavior through lower internal academic success attribution (effort and ability) and higher external academic success attribution (task difficulty and luck), which reduced career decision-making self-efficacy. These findings can be employed to develop more effective intervention programs comprising career guidance for adolescents, which emphasizes the negative effect of helicopter parenting. This study expands the research field, as previous findings on helicopter parenting mostly focus on college students.

Art authentication with vision transformers

In recent years, transformers, initially developed for language, have been successfully applied to visual tasks. Vision transformers have been shown to push the state of the art in a wide range of tasks, including image classification, object detection, and semantic segmentation. While ample research has shown promising results in art attribution and art authentication tasks using convolutional neural networks, this paper examines whether the superiority of vision transformers extends to art authentication, improving, thus, the reliability of computer-based authentication of artworks. Using a carefully compiled dataset of authentic paintings by Vincent van Gogh and two contrast datasets, we compare the art authentication performances of Swin transformers with those of EfficientNet. Using a standard contrast set containing imitations and proxies (works by painters with styles closely related to van Gogh), we find that EfficientNet achieves the best performance overall. With a contrast set that only consists of imitations, we find the Swin transformer to be superior to EfficientNet by achieving an authentication accuracy of over 85%. These results lead us to conclude that vision transformers represent a strong and promising contender in art authentication, particularly in enhancing the computer-based ability to detect artistic imitations.

Normative values and diagnostic optimisation of three social cognition measures for autism and schizophrenia diagnosis in a healthy adolescent and adult sample.

Awareness of the importance of assessing social cognition skills under conditions showing atypical social behaviours has increased over the years. However, the evaluation of the psychometric properties of the measures and the availability of normative values for the clinical context are still limited. This study aims to revise, provide normative values, and evaluate the clinical validity of the Italian version of three social cognition measures: Advanced Theory of Mind (A-ToM) task, the Emotion Attribution Task (EAT), and the Social Situation Task (SST). Measures were administered to 580 adolescents and adult healthy controls (age range 14-50). We performed differential item functioning and Rasch analysis to revise each task, so normative data of the revised measures were calculated. Moreover, the revised measures were administered to 38 individuals with autism spectrum disorder (ASD) and 35 individuals with schizophrenia spectrum disorders (SSD): ASD and SSD were matched by age, gender, and IQ with a control sample to evaluate clinical validity. ROC analysis showed that the SST is the best measure differentiating between healthy and clinical groups, compared to the A-ToM (AUCASD = 0.70; AUCSSD = 0.65) and EAT (AUCASD = 0.67; AUCSSD = 0.50), which showed poorer performance. For SSD diagnosis, two SST subscales (Violation and Gravity score) indicated the best accuracy (AUCs of 0.88 and 0.84, respectively); for the ASD diagnosis, we propose a combined score between the SST subscale and A-ToM (AUC = 0.86). The results suggest that the proposed measures can be used to support the diagnostic process and clinical practice.

Does believing something to be fiction allow a form of moral licencing or a 'fictive pass' in understanding others' actions?

The human capacity to engage with fictional worlds raises important psychological questions about the mechanisms that make this possible. Of particular interest is whether people respond differently to fictional stories compared to factual ones in terms of how immersed they become and how they view the characters involved and their actions. It has been suggested that fiction provides us with a 'fictive pass' that allows us to evaluate in a more balanced, detached way the morality of a character's behaviour. We use a randomised controlled experimental design to test this. We show that, although knowing whether a substantial film clip is fact or fiction does not affect how engaged with ('transported' by) a troubling story an observer becomes, it does grant them a 'fictive pass' to empathise with a moral transgressor. However, a fictive pass does not override the capacity to judge the causes of a character's moral transgression (at least as indexed by a causal attribution task).

Causanom: Anomaly Detection With Flexible Causal Graphs

Causality-based anomaly detection methods provide at least two significant theoretical benefits over purely statistical methods: 1. Improved robustness to non-anomalous out-of-distribution data, which implies a reduction in false-alarms; 2. A potential for failure localization due to the topological ordering of the causal graph. Recent studies have considered the utilization of causality-based methods for time series anomaly detection, however, these methods require the causal graph to be fixed; resultingly, such methods are not robust to incorrectly estimated causal graphs and are not able to natively model counterfactual scenarios. To address these limitations, we introduce Causanom: a graph-based encoder-decoder neural network for time series anomaly detection. Causanom utilizes a node conditional data-stream representation in conjunction with a weighted graph aggregation function in order to efficiently capture heterogeneous node dynamics whilst allowing for a flexible graphical structure. We show that Causanom can be trained along with auxiliary constraints in order to tune the causal graph and improve performance. Additionally, we show that Causanom can be used to produce counterfactual data, which we leverage to identify violated causal relationships. Using real and synthetic time series data respectively, we show that Causanom performs at least as well as state-of-the-art baselines in the anomaly detection task and outperforms existing methods in a causal attribution task.

Effects of task structure and attribution target on responsibility attributions in human–robot interaction

Responsibility attribution refers to beliefs about the cause of an event or outcome. This study investigates the effects of cooperation and competition on responsibility attributions after success and failure in human–robot interaction. A 2 × 2 × 2 mixed-design experiment was conducted, with task structure (cooperation vs. competition) and attribution target (self vs. robot) as within-subject variables and task outcome (win vs. lose) as a between-subject variable. Forty students from a university campus participated in the experiment. They cooperated with or competed against a humanoid robot in a picture memory game whose outcome either exceeded (success) or did not reach (failure) an expected outcome. Afterward, participants allocated the causal responsibility between themselves and the humanoid robot for a specified performance success or failure. It was found that participants attributed more responsibility for success to themselves than to the humanoid robot, and such self-serving bias was greater in competitive than in cooperative structures. Additionally, participants attributed more responsibility for failure to themselves than to the humanoid robot in cooperative structures but not in competitive structures. Moreover, the competitiveness index of unsuccessful participants moderated the relationship between the task structure and the differential responsibility attribution between the self and the robot. Furthermore, participants evaluated the humanoid robot more positively in cooperative than competitive structures. Finally, participants had higher accuracy in cooperative structures, although they subjectively perceived that they performed better in competitive structures. These findings provide insight to guide people's attributional behavior, task performance, and evaluation of robots in the desired manner from the perspective of task structure design.