Despite the remarkable achievements of Large Language Models (LLMs) in various Natural Language Processing tasks, their competence in abstract language understanding remains a relatively under-explored territory. Figurative language interpretation serves as ideal testbed for assessing this as it requires models to navigate beyond the literal meaning and delve into underlying semantics of the figurative expressions. In this paper, we seek to examine the performance of GPT-3.5 in zero-shot setting through word-level metaphor detection. Specifically, we frame the task as annotation of word-level metaphors in proverbs. To this end, we employ a dataset of English proverbs and evaluated its performance by applying different prompting strategies. Our results show that the model shows a satisfactory performance at identifying word-level metaphors, particularly when it is prompted with a hypothetical context preceding the proverb. This observation underscores the pivotal role of well-designed prompts for zero-shot settings through which these models can be leveraged as annotators for subjective NLP tasks.

This paper presents the development of a novel multimodal multilingual dataset in Russian and English, with a particular emphasis on the exploration of laughter detection techniques. Data was collected from YouTube stand-up comedy videos with manually annotated subtitles, and our research covers data preparation and laughter labeling. We explore two laughter detection approaches presented in the literature: peak detection using preprocessed voiceless audio with an energy-based algorithm and machine learning approach with pretrained models to identify laughter presence and duration. While the machine learning approach currently outperforms peak detection in accuracy and generalization, the latter shows promise and warrants further study. Additionally, we explore unimodal and multimodal humor detection on the new dataset, showing the effectiveness of neural models in capturing humor in both languages, even with textual data. Multimodal experiments indicate that even basic models benefit from visual data, improving detection results. However, further research is needed to enhance laughter detection labeling quality and fully understand the impact of different modalities in a multimodal and multilingual context.

The analysis of humor using computational tools has gained popularity in the past few years, and a lot of resources have been built for this purpose. However, most of these resources focus on standalone jokes or on occasional humorous sentences during presentations. In this paper I present a new dataset, SCRIPTS, built using stand-up comedy shows transcripts: the humor that this dataset collects is inserted in a larger narrative, composed of daily events made humorous by the ability of the comedian. This different perspective on the humor problem can allow us to think and study humor in a different way and possibly to open the path to new lines of research.

Abstract Native language identification (NLI)—the task of automatically identifying the native language (L1) of persons based on their writings in the second language (L2)—is based on the hypothesis that characteristics of L1 will surface and interfere in the production of texts in L2 to the extent that L1 is identifiable. We present an in-depth investigation of features that model a variety of linguistic phenomena potentially involved in native language interference in the context of the NLI task: the languages’ structuring of information through punctuation usage, emotion expression in language, and similarities of form with the L1 vocabulary through the use of anglicized words, cognates, and other misspellings. The results of experiments with different combinations of features in a variety of settings allow us to quantify the native language interference value of these linguistic phenomena and show how robust they are in cross-corpus experiments and with respect to proficiency in L2. These experiments provide a deeper insight into the NLI task, showing how native language interference explains the gap between baseline, corpus-independent features, and the state of the art that relies on features/representations that cover (indiscriminately) a variety of linguistic phenomena.

The present paper addresses the study of cross-linguistic and cross-modal iconicity within a deep learning framework. An LSTM-based Recurrent Neural Network is trained to associate the phonetic representation of a concrete word, encoded as a sequence of feature vectors, to the visual representation of its referent, expressed as an HCNN-transformed image. The processing network is then tested, without further training, in a language that does not appear in the training set and belongs to a different language family. The performance of the model is evaluated through a comparison with a randomized baseline; we show that such an imaginative network is capable of extracting language-independent generalizations in the mapping from linguistic sounds to visual features, providing empirical support for the hypothesis of a universal sound-symbolic substrate underlying all languages.

In recent years emotion detection in text has become more popular due to its potential applications in fields such as psychology, marketing, political science, and artificial intelligence, among others. While opinion mining is a well-established task with many standard data sets and well-defined methodologies, emotion mining has received less attention due to its complexity. In particular, the annotated gold standard resources available are not enough. In order to address this shortage, we present a multilingual emotion data set based on different events that took place in April 2019. We collected tweets from the Twitter platform. Then one of seven emotions, six Ekman’s basic emotions plus the “neutral or other emotions”, was labeled on each tweet by 3 Amazon MTurkers. A total of 8,409 in Spanish and 7,303 in English were labeled. In addition, each tweet was also labeled as offensive or no offensive. We report some linguistic statistics about the data set in order to observe the difference between English and Spanish speakers when they express emotions related to the same events. Moreover, in order to validate the effectiveness of the data set, we also propose a machine learning approach for automatically detecting emotions in tweets for both languages, English and Spanish.

In recent years, the increasing interest in the development of automatic approaches for unmasking deception in online sources led to promising results. Nonetheless, among the others, two major issues remain still unsolved: the stability of classifiers performances across different domains and languages. Tackling these issues is challenging since labelled corpora involving multiple domains and compiled in more than one language are few in the scientific literature. For filling this gap, in this paper we introduce DecOp (Deceptive Opinions), a new language resource developed for automatic deception detection in cross-domain and cross-language scenarios. DecOp is composed of 5000 examples of both truthful and deceitful first-person opinions balanced both across five different domains and two languages and, to the best of our knowledge, is the largest corpus allowing cross-domain and cross-language comparisons in deceit detection tasks. In this paper, we describe the collection procedure of the DecOp corpus and his main characteristics. Moreover, the human performance on the DecOp test-set and preliminary experiments by means of machine learning models based on Transformer architecture are shown.

For a long time, philosophers, linguists and scientists have been keen on finding an answer to the mind-bending question “what does abstract language look like?”, which has also sprung from the phenomenon of mental imagery and how this emerges in the mind. One way of approaching the matter of word representations is by exploring the common semantic elements that link words to each other. Visual languages like sign languages have been found to reveal enlightening patterns across signs of similar meanings, pointing towards the possibility of identifying clusters of iconic meanings. With this insight, merged with an understanding of verb predicates achieved from VerbNet, this study presents a novel verb classification system based on visual shapes, using graphic animation to visually represent 20 classes of abstract verbs. Considerable agreement between participants who judged the graphic animations based on representativeness suggests a positive way forward for this proposal, which may be developed as a language learning aid in educational contexts or as a multimodal language comprehension tool for digital text.

Textual emotion detection has a high impact on business, society, politics or education with applications such as, detecting depression or personality traits, suicide prevention or identifying cases of cyber-bulling. Given this context, the objective of our research is to contribute to the improvement of emotion recognition task through an automatic technique focused on reducing both the time and cost needed to develop emotion corpora. Our proposal is to exploit a bootstrapping approach based on intensional learning for automatic annotations with two main steps: 1) an initial similarity-based categorization where a set of seed sentences is created and extended by distributional semantic similarity (word vectors or word embeddings); 2) train a supervised classifier on the initially categorized set. The technique proposed allows us an efficient annotation of a large amount of emotion data with standards of reliability according to the evaluation results.

In this paper, we describe experiments designed to explore and evaluate the impact of punctuation marks on the task of native language identification. Punctuation is specific to each language, and is part of the indicators that overtly represent the manner in which each language organizes and conveys information. Our experiments are organized in various set-ups: the usual multi-class classification for individual languages, also considering classification by language groups, across different proficiency levels, topics and even cross-corpus. The results support our hypothesis that punctuation marks are persistent and robust indicators of the native language of the author, which do not diminish in influence even when a high proficiency level in a non-native language is achieved.

Several NLP studies address the problem of figurative language, but among non-literal phenomena, they have neglected exaggeration.This paper presents a first computational approach to this figure of speech.We explore the possibility to automatically detect exaggerated sentences.First, we introduce HYPO, a corpus containing overstatements (or hyperboles) collected on the web and validated via crowdsourcing.Then, we evaluate a number of models trained on HYPO, and bring evidence that the task of hyperbole identification can be successfully performed based on a small set of semantic features.

Metaphors are intriguing.We know that George Lakoff and Mark Johnson, in their book Metaphors We Live By, pointed out that our daily language is full of metaphors (Lakoff and Johnson 1980).Metaphors are not rare at all as linguistic events and in general as a means of human communication (e.g., in visual art).Thus, more than an exception, they seem to be a necessity for our mind.People use metaphors as strategies to link concepts, to deal with situations, and sometimes to suggest solutions to problems.For example, you can see criminality as a monster or as an illness.What you have in mind to deal with it is probably to fight in the former case, and to cure or to plan prevention in the latter.Nonetheless, metaphors are extremely difficult to model in a computational framework.What is literal?What is metaphorical?Making this distinction has often proved to be a daunting task, even for a human judgment.Metaphors-so bound to our way of thinking and entangled with a huge quantity of knowledge and linguistics subtletiesconstitute an excellent research problem for computational linguistics and artificial intelligence in general.We could probably say that they belong to the AI-complete problems.The difficulty of these computational problems is equivalent to that of solving the central artificial intelligence problem-making computers as intelligent as people.Tony Veale, Ekaterina Shutova, and Beata B. Klebanov are experienced researchers in the field of figurative language processing.Their book is an excellent resource and a good reference for anyone who plans to tackle the subtleties of this complex topic.The book offers a comprehensive approach to the computational treatment of metaphors and of related figurative devices such as simile, analogy, and conceptual blending.The reader is introduced to multiple computational perspectives, from symbolic and statistical approaches to interpretation and paraphrase generation, without omitting contributions from philosophy on what constitutes the significance of a metaphor.The first three chapters introduce the reader to the concept of metaphor, particularly the theoretical foundations profitable for approaching the problem computationally.Particularly useful is the explanation of the related figurative devices: similes (the comparison of one thing with another of a different kind, used to make a description more emphatic or vivid, e.g., John is as brave as a lion); analogy (a comparison between one thing and another, typically for the purpose of explanation, e.g., marriage is slavery); and conceptual blending (a cognitive theory, originally developed by Gilles Fauconnier and Mark Turner [Fauconnier and Turner 2002], which refers to a set of cognitive operations for combining-or blending-words, images, and ideas in a network of "mental spaces" to create meaning, e.g., the painting in George Clooney's attic is a cue to create a conceptual

We present experiments that show the influence of native language on lexical choice when producing text in another language -in this particular case English.We start from the premise that non-native English speakers will choose lexical items that are close to words in their native language.This leads us to an etymologybased representation of documents written by people whose mother tongue is an Indo-European language.Based on this representation we grow a language family tree, that matches closely the Indo-European language tree.

Altering the lyrics of famous songs is a common creative and communicative act, often used for purposes that go beyond simple amusement, such as the creation of companion music for advertisements. In this case, the altered song commonly refers to the advertised product or idea. Here we present a system that can automatically reproduce this process: it starts from a novel text, i.e. the daily news, identifies the key concepts therein contained, expands them and then uses this word cloud to replace some of the words in a song, obeying lexical, metrical and rhyming constraints. The new song is then created by merging these new lyrics, sung by a speech synthesizer, with the original backing music. Our evaluation shows that songs created by the system increase the recall of the news they were created from.

The objective of this research is to develop a technique to automatically annotate emotional corpora. The complexity of automatic annotation of emotional corpora still presents numerous challenges and thus there is a need to develop a technique that allow us to tackle the annotation task. The relevance of this research is demonstrated by the fact that people's emotions and the patterns of these emotions provide a great value for business, individuals, society or politics. Hence, the creation of a robust emotion detection system becomes crucial. Due to the subjectivity of the emotions, the main challenge for the creation of emotional resources is the annotation process. Thus, with this staring point in mind, the objective of our paper is to illustrate an innovative and effective bootstrapping process for automatic annotations of emotional corpora. The evaluations carried out confirm the soundness of the proposed approach and allow us to consider the bootstrapping process as an appropriate approach to create resources such as an emotional corpus that can be employed on supervised machine learning towards the improvement of emotion detection systems.

In this paper, we experiment with a resource consisting of metaphorically annotated proverbs on the task of word-level metaphor recognition.We observe that existing feature sets do not perform well on this data.We design a novel set of features to better capture the peculiar nature of proverbs and we demonstrate that these new features are significantly more effective on the metaphorically dense proverb data.

While the effect of various lexical, syntactic, semantic and stylistic features have been addressed in persuasive language from a computational point of view, the persuasive effect of phonetics has received little attention.By modeling a notion of euphony and analyzing four datasets comprising persuasive and nonpersuasive sentences in different domains (political speeches, movie quotes, slogans and tweets), we explore the impact of sounds on different forms of persuasiveness.We conduct a series of analyses and prediction experiments within and across datasets.Our results highlight the positive role of phonetic devices on persuasion.

The need for creativity is ubiquitous, and mobile devices connected to Web services can help us. Linguistic creativity is widely used in advertisements to surprise us, to get our attention, and to stick concepts in our memory. However, creativity can also be used as a defense. When we walk in the street, we are overwhelmed by messages that try to get our attention with any persuasive device at hand. As messages get ever more aggressive, often our basic cognitive defenses—trying not to perceive those messages—are not sufficient. One advanced defensive technique is based on transforming the perceived message into something different (for instance, making use of irony or hyperbole) from what was originally meant in the message. In this article, we describe an implemented application for smartphones, which creatively modifies the linguistic expression in a virtual copy of a poster encountered on the street. The mobile system is inspired by thesubvertisingpractice of countercultural art.

This paper reports on the development of a hy-brid and simple method based on a machine learning classifier (Naive Bayes), Word Sense Disambiguation and rules, for the automatic assignment of WordNet Domains to nominal entries of a lexicographic dictionary, the Senso Comune De Mauro Lexicon. The system ob-tained an F1 score of 0.58, with a Precision of 0.70. We further used the automatically as-signed domains to filter out word sense align-ments between MultiWordNet and Senso Co-mune. This has led to an improvement in the quality of the sense alignments showing the validity of the approach for domain assign-ment and the importance of domain informa-tion for achieving good sense alignments. 1

Connecting words with senses, namely, sight, hearing, taste, smell and touch, to comprehend the sensorial information in language is a straightforward task for humans by using commonsense knowledge. With this in mind, a lexicon associating words with senses would be crucial for the computational tasks aiming at interpretation of language. However, to the best of our knowledge, there is no systematic attempt in the literature to build such a resource. In this paper, we present a sensorial lexicon that associates English words with senses. To obtain this resource, we apply a computational method based on bootstrapping and corpus statistics. The quality of the resulting lexicon is evaluated with a gold standard created via crowdsourcing. The results show that a simple classifier relying on the lexicon outperforms two baselines on a sensory classification task, both at word and sentence level, and confirm the soundness of the proposed approach for the construction of the lexicon and the usefulness of the resource for computational applications.

This work describes the evaluations of two approaches, Lexical Matching and Sense Similarity, for word sense alignment between MultiWordNet and a lexicographic dictionary, Senso Comune De Mauro, when having few sense descriptions (MultiWordNet) and no structure over senses (Senso Comune De Mauro).The results obtained from the merging of the two approaches are satisfying, with F1 values of 0.47 for verbs and 0.64 for nouns.

We present a qualitative analysis of the lexicon of dialogue acts: we explore the relationship between the communicative goal of an utterance and its affective lexicon as well as the salience of specific word classes for each speech act. Thought not constituting any deep understanding of the dialogue, automatic dialogue act labeling is a task that may be relevant for a wide range of applications in both human-computer and human-human interaction. The experiments described in this paper fit in the scope of a research study whose long-term goal is to build an unsupervised classifier that simply exploits the lexical semantics of utterances to automatically annotate dialogues with the proper speech acts.

In this paper we present a tool for valence shifting of natural language texts, named Valentino (VALENced Text INOculator). Valentino can modify existing textual expressions towards more positively or negatively valenced versions. To this end we built specific resources, gathering valenced terms that are semantically or contextually connected to the original one, and implemented strategies that use these resources in the substitution process. Valentino is meant to be a modular component. It is non-domain specific and it requires as its input a coefficient that represents the desired valence for the final expression.

Vocabulary acquisition constitutes a crucial but difficult and time consuming step of learning a foreign language. There exist several teaching methods which aim to facilitate this step by providing learners with various verbal and visual tips. However, building systems based on these methods is generally very costly since it requires so many resources such as time, money and human labor. In this paper, we introduce a fully automatized vocabulary teaching system which uses state-of-the-art natural language processing (NLP) and information retrieval (IR) techniques. For each foreign word the user is willing to learn, the system is capable of automatically producing memorization tips including key- words, sentences, colors, textual animations and images.

We present a semi-supervised technique for word sense disambiguation that exploits external knowledge acquired in an unsupervised manner. In particular, we use a combination of basic kernel functions to independently estimate syntagmatic and domain similarity, building a set of word-expert classifiers that share a common domain model acquired from a large corpus of unlabeled data. The results show that the proposed approach achieves state-of-the-art performance on a wide range of lexical sample tasks and on the English all-words task of Senseval-3, although it uses a considerably smaller number of training examples than other methods.

Among forms of creative language, verbal humor has received some attention in the computational milieu. Some aspects of irony and wordplay could be experimented in automated systems. For instance, we developed a system that makes fun of existing acronyms, based mainly on lexical reasoning. The dimension of emotion in words is also starting to be understood among computational linguists. The challenge of electronic advertisements offers in particular a great opportunity for getting now deeper into creative language expression and emotion. An advertising message induces in the recipient a positive or negative attitude toward the advertised object. A prototype we have developed for advertising professionals has two steps: (i) the creative variation of familiar expressions, taking into account the effective content of the produced text, (ii) the automatic animation (semantically consistent with the affective text content) of the resulting expression, using kinetic typography techniques. In this way the ironic-affective effect is perceptually emphasized.

This paper investigates conceptually and empirically the novel sense matching task, which requires to recognize whether the senses of two synonymous words match in context. We suggest direct approaches to the problem, which avoid the intermediate step of explicit word sense disambiguation, and demonstrate their appealing advantages and stimulating potential for future research.

Cross-language Text Categorization is the task of assigning semantic classes to documents written in a target language (e.g. English) while the system is trained using labeled documents in a source language (e.g. Italian).In this work we present many solutions according to the availability of bilingual resources, and we show that it is possible to deal with the problem even when no such resources are accessible. The core technique relies on the automatic acquisition of Multilingual Domain Models from comparable corpora.Experiments show the effectiveness of our approach, providing a low cost solution for the Cross Language Text Categorization task. In particular, when bilingual dictionaries are available the performance of the categorization gets close to that of monolingual text categorization.

Humor is one of the most interesting and puzzling aspects of human behavior. Despite the attention it has received in fields such as philosophy, linguistics, and psychology, there have been only few attempts to create computational models for humor recognition or generation. In this article, we bring empirical evidence that computational approaches can be successfully applied to the task of humor recognition. Through experiments performed on very large data sets, we show that automatic classification techniques can be effectively used to distinguish between humorous and non‐humorous texts, with significant improvements observed over a priori known baselines.

In this paper we present a supervised Word Sense Disambiguation methodology, that exploits kernel methods to model sense distinctions. In particular a combination of kernel functions is adopted to estimate independently both syntagmatic and domain similarity. We defined a kernel function, namely the Domain Kernel, that allowed us to plug "external knowledge" into the supervised learning process. External knowledge is acquired from unlabeled data in a totally unsupervised way, and it is represented by means of Domain Models. We evaluated our methodology on several lexical sample tasks in different languages, outperforming significantly the state-of-the-art for each of them, while reducing the amount of labeled training data required for learning.

Computational humor will be needed in interfaces, no less than other cognitive capabilities. There are many practical settings where computational humor will add value. Among them there are: business world applications (such as advertisement, e-commerce, etc...), general computer-mediated communication and human-computer interaction, increase in the friendliness of natural language interfaces, educational and edutainment systems. In particular in the educational field it is an important resource for getting selective attention, help in memorizing names and situations etc. And we all know how well it works with children.Automated humor production in general is a very difficult task but we wanted to prove that some results can be achieved even in short time. We have worked at a concrete limited problem, as the core of the European Project HAHAcronym. The main goal of HAHAcronym has been the realization of an acronym ironic re-analyzer and generator as a proof of concept in a focalized but non restricted context. In order to implement this system some general tools have been adapted, or developed for the humorous context. Systems output has been submitted to evaluation by human subjects, with a very positive result

This paper explores the role of domain information in word sense disambiguation. The underlying hypothesis is that domain labels, such as MEDICINE, ARCHITECTURE and SPORT, provide a useful way to establish semantic relations among word senses, which can be profitably used during the disambiguation process. Results obtained at the SENSEVAL-2 initiative confirm that for a significant subset of words domain information can be used to disambiguate with a very high level of precision.

Multimodality is a powerful concept for dealing with dialogue cohesion in a human–computer natural language (NL)‐centered system. This work is a modest step toward more effective exploitation of the potentially large bandwidth of communication provided by this situation. The relations between exploration, navigation, and NL‐based communication are discussed in general and with reference to two prototypes. Light cognitive load feedback and direct manipulation are proposed so that user and system can cooperate in mutually establishing the structure of the ongoing dialogue. The main points are: (i) use of an appropriate dialogue structure to constrain inference in the anaphora resolution process; (ii) use of a graphical representation of the structure, to limit the problem of opacity; (iii) allowance for the possibility of direct manipulation on this representation, to avoid the necessity of operating linguistically at the metalevel. The context of the work is within NL‐centered multimodal information access systems, in which basic entities are pairs (most commonly question and answer). A dialogue model is provided by a modified version of the centering model; it is both sufficiently simple to be displayed in an intuitive fashion on the screen, and sufficiently powerful to give accurate results. An extension of the discourse model, oriented to the treatment of deixis, is also proposed. Finally, steps toward an overall approach to the integration of navigational and mediated aspects of interaction are discussed.