The automatic detection of temporal relations among events has been mainly investigated with encoder-only models such as RoBERTa. Large Language Models (LLM) have recently shown promising performance in temporal reasoning tasks such as temporal question answering. Nevertheless, recent studies have tested the LLMs’ performance in detecting temporal relations of closed-source models only, limiting the interpretability of those results. In this work, we investigate LLMs’ performance and decision process in the Temporal Relation Classification task. First, we assess the performance of seven open and closed-sourced LLMs experimenting with in-context learning and lightweight fine-tuning approaches. Results show that LLMs with in-context learning significantly underperform smaller encoder-only models based on RoBERTa. Then, we delve into the possible reasons for this gap by applying explainable methods. The outcome suggests a limitation of LLMs in this task due to their autoregressive nature, which causes them to focus only on the last part of the sequence. Additionally, we evaluate the word embeddings of these two models to better understand their pre-training differences. The code and the fine-tuned models can be found respectively on GitHub.

LLMs acquire knowledge from massive data snapshots collected at different timestamps. Their knowledge is then commonly evaluated using static benchmarks. However, factual knowledge is generally subject to time-sensitive changes, and static benchmarks cannot address those cases. We present an approach to dynamically evaluate the knowledge in LLMs and their time-sensitiveness against Wikidata, a publicly available up-to-date knowledge graph. We evaluate the time-sensitive knowledge in twenty-four private and open-source LLMs, as well as the effectiveness of four editing methods in updating the outdated facts. Our results show that 1) outdatedness is a critical problem across state-of-the-art LLMs; 2) LLMs output inconsistent answers when prompted with slight variations of the question prompt; and 3) the performance of the state-of-the-art knowledge editing algorithms is very limited, as they can not reduce the cases of outdatedness and output inconsistency.

We study the limitations of Large Language Models (LLMs) for the task of response generation in human-machine dialogue. Several techniques have been proposed in the literature for different dialogue types (e.g., Open-Domain). However, the evaluations of these techniques have been limited in terms of base LLMs, dialogue types and evaluation metrics. In this work, we extensively analyze different LLM adaptation techniques when applied to different dialogue types. We have selected two base LLMs, Llama-2 and Mistral, and four dialogue types Open-Domain, Knowledge-Grounded, Task-Oriented, and Question Answering. We evaluate the performance of in-context learning and fine-tuning techniques across datasets selected for each dialogue type. We assess the impact of incorporating external knowledge to ground the generation in both scenarios of Retrieval-Augmented Generation (RAG) and gold knowledge. We adopt consistent evaluation and explainability criteria for automatic metrics and human evaluation protocols. Our analysis shows that there is no universal best-technique for adapting large language models as the efficacy of each technique depends on both the base LLM and the specific type of dialogue. Last but not least, the assessment of the best adaptation technique should include human evaluation to avoid false expectations and outcomes derived from automatic metrics.

Longitudinal Dialogues (LD) are the most challenging type of conversation for human-machine dialogue systems. LDs include the recollections of events, personal thoughts, and emotions specific to each individual in a sparse sequence of dialogue sessions. Dialogue systems designed for LDs should uniquely interact with the users over multiple sessions and long periods of time (e.g. weeks), and engage them in personal dialogues to elaborate on their feelings, thoughts, and real-life events. In this paper, we study the task of response generation in LDs. We evaluate whether general-purpose Pre-trained Language Models (PLM) are appropriate for this purpose. We fine-tune two PLMs, GePpeTto (GPT-2) and iT5, using a dataset of LDs. We experiment with different representations of the personal knowledge extracted from LDs for grounded response generation, including the graph representation of the mentioned events and participants. We evaluate the performance of the models via automatic metrics and the contribution of the knowledge via the Integrated Gradients technique. We categorize the natural language generation errors via human evaluations of contextualization, appropriateness and engagement of the user.

The valence analysis of speakers’ utterances or written posts helps to understand the activation and variations of the emotional state throughout the conversation. More recently, the concept of Emotion Carriers (EC) has been introduced to explain the emotion felt by the speaker and its manifestations. In this work, we investigate the natural inter-dependency of valence and ECs via a multi-task learning approach. We experiment with Pre-trained Language Models (PLM) for single-task, two-step, and joint settings for the valence and EC prediction tasks. We compare and evaluate the performance of generative (GPT-2) and discriminative (BERT) architectures in each setting. We observed that providing the ground truth label of one task improves the prediction performance of the models in the other task. We further observed that the discriminative model achieves the best trade-off of valence and EC prediction tasks in the joint prediction setting. As a result, we attain a single model that performs both tasks, thus, saving computation resources at training and inference times.

Narratives include a rich source of events unfolding over time and context. Automatic understanding of these events provides a summarised comprehension of the narrative for further computation (such as reasoning). In this paper, we study the Information Status (IS) of the events and propose a novel challenging task: the automatic identification of new events in a narrative. We define an event as a triplet of subject, predicate, and object. The event is categorized as new with respect to the discourse context and whether it can be inferred through commonsense reasoning. We annotated a publicly available corpus of narratives with the new events at sentence level using human annotators. We present the annotation protocol and study the quality of the annotation and the difficulty of the task. We publish the annotated dataset, annotation materials, and machine learning baseline models for the task of new event extraction for narrative understanding.

Human Evaluation (HE) of automatically generated responses is necessary for the advancement of human-machine dialogue research. Current automatic evaluation measures are poor surrogates, at best. There are no agreed-upon HE protocols and it is difficult to develop them. As a result, researchers either perform non-replicable, non-transparent and inconsistent procedures or, worse, limit themselves to automated metrics. We propose to standardize the human evaluation of response generation models by publicly sharing a detailed protocol. The proposal includes the task design, annotators recruitment, task execution, and annotation reporting. Such protocol and process can be used as-is, as-a-whole, in-part, or modified and extended by the research community. We validate the protocol by evaluating two conversationally fine-tuned state-of-the-art models (GPT-2 and T5) for the complex task of personalized response generation. We invite the community to use this protocol - or its future community amended versions - as a transparent, replicable, and comparable approach to HE of generated responses.

Sentiment analysis is one of the most widely studied tasks in natural language processing. While BERT-based models have achieved state-of-the-art results in this task, little attention has been given to its performance variability across class labels, multi-source and multi-domain corpora. In this paper, we present an improved state-of-the-art and comparatively evaluate BERT-based models for sentiment analysis on Italian corpora. The proposed model is evaluated over eight sentiment analysis corpora from different domains (social media, finance, e-commerce, health, travel) and sources (Twitter, YouTube, Facebook, Amazon, Tripadvisor, Opera and Personal Healthcare Agent) on the prediction of positive, negative and neutral classes. Our findings suggest that BERT-based models are confident in predicting positive and negative examples but not as much with neutral examples. We release the sentiment analysis model as well as a newly financial domain sentiment corpus.

Personal Narrative (PN) is the recollection of individuals’ life experiences, events, and thoughts along with the associated emotions in the form of a story. Compared to other genres such as social media texts or microblogs, where people write about experienced events or products, the spoken PNs are complex to analyze and understand. They are usually long and unstructured, involving multiple and related events, characters as well as thoughts and emotions associated with events, objects, and persons. In spoken PNs, emotions are conveyed by changing the speech signal characteristics as well as the lexical content of the narrative. In this work, we annotate a corpus of spoken personal narratives, with the emotion valence using discrete values. The PNs are segmented into speech segments, and the annotators annotate them in the discourse context, with values on a 5-point bipolar scale ranging from -2 to +2 (0 for neutral). In this way, we capture the unfolding of the PNs events and changes in the emotional state of the narrator. We perform an in-depth analysis of the inter-annotator agreement, the relation between the label distribution w.r.t. the stimulus (positive/negative) used for the elicitation of the narrative, and compare the segment-level annotations to a baseline continuous annotation. We find that the neutral score plays an important role in the agreement. We observe that it is easy to differentiate the positive from the negative valence while the confusion with the neutral label is high. Keywords: Personal Narratives, Emotion Annotation, Segment Level Annotation

Deep Neural Networks (DNN) models have achieved acceptable performance in sentiment prediction of written text. However, the output of these machine learning (ML) models cannot be natively interpreted. In this paper, we study how the sentiment polarity predictions by DNNs can be explained and compare them to humans’ explanations. We crowdsource a corpus of Personal Narratives and ask human judges to annotate them with polarity and select the corresponding token chunks - the Emotion Carriers (EC) - that convey narrators’ emotions in the text. The interpretations of ML neural models are carried out through Integrated Gradients method and we compare them with human annotators’ interpretations. The results of our comparative analysis indicate that while the ML model mostly focuses on the explicit appearance of emotions-laden words (e.g. happy, frustrated), the human annotator predominantly focuses the attention on the manifestation of emotions through ECs that denote events, persons, and objects which activate narrator’s emotional state.

The acquisition of a dialogue corpus is a key step in the process of training a dialogue model. In this context, corpora acquisitions have been designed either for open-domain information retrieval or slot-filling (e.g. restaurant booking) tasks. However, there has been scarce research in the problem of collecting personal conversations with users over a long period of time. In this paper we focus on the types of dialogues that are required for mental health applications. One of these types is the follow-up dialogue that a psychotherapist would initiate in reviewing the progress of a Cognitive Behavioral Therapy (CBT) intervention. The elicitation of the dialogues is achieved through textual stimuli presented to dialogue writers. We propose an automatic algorithm that generates textual stimuli from personal narratives collected during psychotherapy interventions. The automatically generated stimuli are presented as a seed to dialogue writers following principled guidelines. We analyze the linguistic quality of the collected corpus and compare the performances of psychotherapists and non-expert dialogue writers. Moreover, we report the human evaluation of a corpus-based response-selection model.

We are interested in the problem of understanding personal narratives (PN) - spoken or written - recollections of facts, events, and thoughts. For PNs, we define emotion carriers as the speech or text segments that best explain the emotional state of the narrator. Such segments may span from single to multiple words, containing for example verb or noun phrases. Advanced automatic understanding of PNs requires not only the prediction of the narrator’s emotional state but also to identify which events (e.g. the loss of a relative or the visit of grandpa) or people (e.g. the old group of high school mates) carry the emotion manifested during the personal recollection. This work proposes and evaluates an annotation model for identifying emotion carriers in spoken personal narratives. Compared to other text genres such as news and microblogs, spoken PNs are particularly challenging because a narrative is usually unstructured, involving multiple sub-events and characters as well as thoughts and associated emotions perceived by the narrator. In this work, we experiment with annotating emotion carriers in speech transcriptions from the Ulm State-of-Mind in Speech (USoMS) corpus, a dataset of PNs in German. We believe this resource could be used for experiments in the automatic extraction of emotion carriers from PN, a task that could provide further advancements in narrative understanding.

In this work, we investigate the human perception of coherence in open-domain dialogues. In particular, we address the problem of annotating and modeling the coherence of next-turn candidates while considering the entire history of the dialogue. First, we create the Switchboard Coherence (SWBD-Coh) corpus, a dataset of human-human spoken dialogues annotated with turn coherence ratings, where next-turn candidate utterances ratings are provided considering the full dialogue context. Our statistical analysis of the corpus indicates how turn coherence perception is affected by patterns of distribution of entities previously introduced and the Dialogue Acts used. Second, we experiment with different architectures to model entities, Dialogue Acts and their combination and evaluate their performance in predicting human coherence ratings on SWBD-Coh. We find that models combining both DA and entity information yield the best performances both for response selection and turn coherence rating.

The increase in the prevalence of mental health problems has coincided with a growing popularity of health related social networking sites. Regardless of their therapeutic potential, on-line support groups (OSGs) can also have negative effects on patients. In this work we propose a novel methodology to automatically verify the presence of therapeutic factors in social networking websites by using Natural Language Processing (NLP) techniques. The methodology is evaluated on on-line asynchronous multi-party conversations collected from an OSG and Twitter. The results of the analysis indicate that therapeutic factors occur more frequently in OSG conversations than in Twitter conversations. Moreover, the analysis of OSG conversations reveals that the users of that platform are supportive, and interactions are likely to lead to the improvement of their emotional state. We believe that our method provides a stepping stone towards automatic analysis of emotional states of users of online platforms. Possible applications of the method include provision of guidelines that highlight potential implications of using such platforms on users’ mental health, and/or support in the analysis of their impact on specific individuals.

Dialogue Act (DA) tagging is crucial for spoken language understanding systems, as it provides a general representation of speakers’ intents, not bound to a particular dialogue system. Unfortunately, publicly available data sets with DA annotation are all based on different annotation schemes and thus incompatible with each other. Moreover, their schemes often do not cover all aspects necessary for open-domain human-machine interaction. In this paper, we propose a methodology to map several publicly available corpora to a subset of the ISO standard, in order to create a large task-independent training corpus for DA classification. We show the feasibility of using this corpus to train a domain-independent DA tagger testing it on out-of-domain conversational data, and argue the importance of training on multiple corpora to achieve robustness across different DA categories.

Summarization of spoken conversations is a challenging task, since it requires deep understanding of dialogs. Abstractive summarization techniques rely on linking the summary sentences to sets of original conversation sentences, i.e. communities. Unfortunately, such linking information is rarely available or requires trained annotators. We propose and experiment automatic community creation using cosine similarity on different levels of representation: raw text, WordNet SynSet IDs, and word embeddings. We show that the abstractive summarization systems with automatic communities significantly outperform previously published results on both English and Italian corpora.

Silence is an integral part of the most frequent turn-taking phenomena in spoken conversations. Silence is sized and placed within the conversation flow and it is coordinated by the speakers along with the other speech acts. The objective of this analytical study is twofold: to explore the functions of silence with duration of one second and above, towards information flow in a dyadic conversation utilizing the sequences of dialog acts present in the turns surrounding the silence itself; and to design a feature space useful for clustering the silences using a hierarchical concept formation algorithm. The resulting clusters are manually grouped into functional categories based on their similarities. It is observed that the silence plays an important role in response preparation, also can indicate speakers’ hesitation or indecisiveness. It is also observed that sometimes long silences can be used deliberately to get a forced response from another speaker thus making silence a multi-functional and an important catalyst towards information flow.

On June 23rd 2016, UK held the referendum which ratified the exit from the EU. While most of the traditional pollsters failed to forecast the final vote, there were online systems that hit the result with high accuracy using opinion mining techniques and big data. Starting one month before, we collected and monitored millions of posts about the referendum from social media conversations, and exploited Natural Language Processing techniques to predict the referendum outcome. In this paper we discuss the methods used by traditional pollsters and compare it to the predictions based on different opinion mining techniques. We find that opinion mining based on agreement/disagreement classification works better than opinion mining based on polarity classification in the forecast of the referendum outcome.

In this paper, we address the issue of automatic prediction of readers’ mood from newspaper articles and comments. As online newspapers are becoming more and more similar to social media platforms, users can provide affective feedback, such as mood and emotion. We have exploited the self-reported annotation of mood categories obtained from the metadata of the Italian online newspaper corriere.it to design and evaluate a system for predicting five different mood categories from news articles and comments: indignation, disappointment, worry, satisfaction, and amusement. The outcome of our experiments shows that overall, bag-of-word-ngrams perform better compared to all other feature sets; however, stylometric features perform better for the mood score prediction of articles. Our study shows that self-reported annotations can be used to design automatic mood prediction systems.

Spoken conversation corpora often adapt existing Dialogue Act (DA) annotation specifications, such as DAMSL, DIT++, etc., to task specific needs, yielding incompatible annotations; thus, limiting corpora re-usability. Recently accepted ISO standard for DA annotation – Dialogue Act Markup Language (DiAML) – is designed as domain and application independent. Moreover, the clear separation of dialogue dimensions and communicative functions, coupled with the hierarchical organization of the latter, allows for classification at different levels of granularity. However, re-annotating existing corpora with the new scheme might require significant effort. In this paper we test the utility of the ISO standard through comparative evaluation of the corpus-specific legacy and the semi-automatically transferred DiAML DA annotations on supervised dialogue act classification task. To test the domain independence of the resulting annotations, we perform cross-domain and data aggregation evaluation. Compared to the legacy annotation scheme, on the Italian LUNA Human-Human corpus, the DiAML annotation scheme exhibits better cross-domain and data aggregation classification performance, while maintaining comparable in-domain performance.

In this paper, we present a corpus of news blog conversations in Italian annotated with gold standard agreement/disagreement relations at message and sentence levels. This is the first resource of this kind in Italian. From the analysis of ADRs at the two levels emerged that agreement annotated at message level is consistent and generally reflected at sentence level, moreover, the argumentation structure of disagreement is more complex than agreement. The manual error analysis revealed that this resource is useful not only for the analysis of argumentation, but also for the detection of irony/sarcasm in online debates. The corpus and annotation tool are available for research purposes on request.

Annotating and predicting behavioural aspects in conversations is becoming critical in the conversational analytics industry. In this paper we look into inter-annotator agreement of agent behaviour dimensions on two call center corpora. We find that the task can be annotated consistently over time, but that subjectivity issues impacts the quality of the annotation. The reformulation of some of the annotated dimensions is suggested in order to improve agreement.

In this paper, we aim to investigate the coordination of interlocutors behavior in different emotional segments. Conversational coordination between the interlocutors is the tendency of speakers to predict and adjust each other accordingly on an ongoing conversation. In order to find such a coordination, we investigated 1) lexical similarities between the speakers in each emotional segments, 2) correlation between the interlocutors using psycholinguistic features, such as linguistic styles, psychological process, personal concerns among others, and 3) relation of interlocutors turn-taking behaviors such as competitiveness. To study the degree of coordination in different emotional segments, we conducted our experiments using real dyadic conversations collected from call centers in which agent’s emotional state include empathy and customer’s emotional states include anger and frustration. Our findings suggest that the most coordination occurs between the interlocutors inside anger segments, where as, a little coordination was observed when the agent was empathic, even though an increase in the amount of non-competitive overlaps was observed. We found no significant difference between anger and frustration segment in terms of turn-taking behaviors. However, the length of pause significantly decreases in the preceding segment of anger where as it increases in the preceding segment of frustration.

The development of annotated corpora is a critical process in the development of speech applications for multiple target languages. While the technology to develop a monolingual speech application has reached satisfactory results (in terms of performance and effort), porting an existing application from a source language to a target language is still a very expensive task. In this paper we address the problem of creating multilingual aligned corpora and its evaluation in the context of a spoken language understanding (SLU) porting task. We discuss the challenges of the manual creation of multilingual corpora, as well as present the algorithms for the creation of multilingual SLU via Statistical Machine Translation (SMT).

We describe two constraint-based methods that can be used to improve the recall of a shallow discourse parser based on conditional random field chunking. These method uses a set of natural structural constraints as well as others that follow from the annotation guidelines of the Penn Discourse Treebank. We evaluated the resulting systems on the standard test set of the PDTB and achieved a rebalancing of precision and recall with improved F-measures across the board. This was especially notable when we used evaluation metrics taking partial matches into account; for these measures, we achieved F-measure improvements of several points.

In this paper, we make a qualitative and quantitative analysis of discourse relations within the LUNA conversational spoken dialog corpus. In particular, we first describe the Penn Discourse Treebank (PDTB) and then we detail the adaptation of its annotation scheme to the LUNA corpus of Italian task-oriented dialogs in the domain of software/hardware assistance. We discuss similarities and differences between our approach and the PDTB paradigm and point out the peculiarities of spontaneous dialogs w.r.t. written text, which motivated some changes in the annotation strategy. In particular, we introduced the annotation of relations between non-contiguous arguments and we modified the sense hierarchy in order to take into account the important role of pragmatics in dialogs. In the final part of the paper, we present a comparison between the sense and connective frequency in a representative subset of the LUNA corpus and in the PDTB. Such analysis confirmed the differences between the two corpora and corroborates our choice to introduce dialog-specific adaptations.

In this paper we present an active approach to annotate with lexical and semantic labels an Italian corpus of conversational human-human and Wizard-of-Oz dialogues. This procedure consists in the use of a machine learner to assist human annotators in the labeling task. The computer assisted process engages human annotators to check and correct the automatic annotation rather than starting the annotation from un-annotated data. The active learning procedure is combined with an annotation error detection to control the reliablity of the annotation. With the goal of converging as fast as possible to reliable automatic annotations minimizing the human effort, we follow the active learning paradigm, which selects for annotation the most informative training examples required to achieve a better level of performance. We show that this procedure allows to quickly converge on correct annotations and thus minimize the cost of human supervision.