Conversational Context Research Articles

Why I Choose This Sticker When Chatting with You: Exploring Design Considerations for Sticker Recommendation Services in Mobile Instant Messengers

Nowadays, many mobile instant messenger (MIM) users employ stickers to express their emotions. Due to the increase in the variety of stickers, many MIM platforms have begun to offer sticker recommendation services, which suggest appropriate stickers based on the context of conversations. Although sticker recommendation services need to consider users' subtle social and psychological needs from the perspective that stickers are an alternative communication language, few attempts have been made to explore this area. Consequently, we investigated sticker usage practices through a qualitative study with 18 KakaoTalk users in South Korea. Based on the results of this study, we revealed user's three common considerations in sticker usage experiences and discussed the design implications for sticker recommendation services in MIM.

Deep emotion recognition in textual conversations: a survey

Emotion Recognition in Conversations (ERC) is a key step towards successful human–machine interaction. While the field has seen tremendous advancement in the last few years, new applications and implementation scenarios present novel challenges and opportunities. These range from leveraging the conversational context, speaker, and emotion dynamics modelling, to interpreting common sense expressions, informal language, and sarcasm, addressing challenges of real-time ERC, recognizing emotion causes, different taxonomies across datasets, multilingual ERC, and interpretability. This survey starts by introducing ERC, elaborating on the challenges and opportunities of this task. It proceeds with a description of the emotion taxonomies and a variety of ERC benchmark datasets employing such taxonomies. This is followed by descriptions comparing the most prominent works in ERC with explanations of the neural architectures employed. Then, it provides advisable ERC practices towards better frameworks, elaborating on methods to deal with subjectivity in annotations and modelling and methods to deal with the typically unbalanced ERC datasets. Finally, it presents systematic review tables comparing several works regarding the methods used and their performance. Benchmarking these works highlights resorting to pre-trained Transformer Language Models to extract utterance representations, using Gated and Graph Neural Networks to model the interactions between these utterances, and leveraging Generative Large Language Models to tackle ERC within a generative framework. This survey emphasizes the advantage of leveraging techniques to address unbalanced data, the exploration of mixed emotions, and the benefits of incorporating annotation subjectivity in the learning phase.

Socio-Pragmatic Insights into Adichie’s Half of a Yellow Sun

The present study focuses on a socio-pragmatic perspective within Discourse Analysis. Grounded in the theoretical frameworks of Grice’s Cooperative Principles (1975), Conversational Implicatures (Levinson, 1983, 2000), and Politeness Principles (Levinson, 1978; Leech, 1983), it examines five selected excerpts from Adichie’s Half of a Yellow Sun to illuminate the varying strategies through which interlocutors navigate language according to shifting conversational contexts. Following a systematic corpus selection process, the quantitative analysis and interpretation of findings reveal how characters work cooperatively to achieve mutual understanding within conversations, often generating different implicatures. The findings suggest that an utterance’s conversational significance is shaped entirely by its specific context of use. Consequently, this study underscores the essential recommendation that, for effective communication, conversational participants must attentively utilize contextual cues to construct and interpret utterances’ ad hoc conversational values, thereby fostering mutual understanding and common ground.

Constructing Talent and Flexibility in Focus Group Discussions on Inequality and Unemployment in Greece

ABSTRACTThis article examines the ways in which participants construct talent and flexibility in accounts of income inequality and (un)employment in Greece. Eight (8) focus groups with forty‐one (41) Greek citizens, aged 18–65 years were conducted. Analysis used the principles of Critical Discursive Social Psychology (CDSP). Participants were asked to evaluate the compensation that should be given to employees from different professional categories and to discuss changes in the Greek (un)employment policy. Findings indicated that in the first conversational context, talent (as one's inherent characteristic or as a byproduct of personal effort) was constructed as important for professional success. Besides, putting the right (talented) person in the right job was constituted as necessary for achieving maximum productivity. On the other hand, in the second conversational context, when participants discussed the new law on unemployment, they underscored the obligation of the unemployed to conform with the labour market's needs, regardless of their qualifications or talents. The ideal unemployed individual was represented as highly active and adaptable, while productivity was linked to flexibility. The discussion relates these potentially contradictory argumentative lines mobilised by participants to dilemmatic aspects of neoliberal ideology and its reproduction.

Not in the file: How competency committees work with undocumented contributions.

Competence committees (CCs) centre their work around documentation of trainees' performance; undocumented contributions (i.e. informal, unrecorded material like personal judgements, experiential anecdotes and contextual information) evoke suspicion even though they may play a role in decision making. This qualitative multiple case study incorporates insights from a social practice perspective on writing to examine the use of undocumented contributions by the CCs of two large post-graduate training programmes, one in a more procedural (MP) speciality and the other in a less procedural (LP) one. Data were collected via observations of meetings and semi-structured interviews with CC members. In the analysis, conversations were organised into triptychs of lead-up, undocumented contribution(s), and follow-up. We then created thick descriptions around the undocumented contributions, drawing on conversational context and interview data to assign possible motivations and significance. We found no instances in which undocumented contributions superseded the contents of a trainee's file or stood in for missing documentation. The number of undocumented contributions varied between the MP CC (six instances over two meetings) and the LP CC (22 instances over three meetings). MP CC discussions emphasised Entrustable Professional Activity (EPA) observations, whereas LP CC members paid more attention to narrative data. The divergent orientations of the CCs-adding an 'advis[ing]/guid[ing]' role versus focusing simply on evaluation-offers the most compelling explanation. In lead-ups, undocumented contributions were prompted by missing and flawed documentation, conflicting evidence and documentation at odds with members' perceptions. Recognising other 'red flags' in documentation often required professional experience. In follow-ups, purposes served by undocumented contributions varied with context and were difficult to generalise; we, therefore, provide deeper analysis of two vignettes to illustrate. Our data suggest undocumented contributions often serve best efforts to ground decisions in documentation. We would encourage CC practices and policies be rooted in more nuanced approaches to documentation.

GASCOM: Graph-based Attentive Semantic Context Modeling for Online Conversation Understanding

Online conversation understanding is an important yet challenging NLP problem which has many useful applications (e.g., hate speech detection). However, online conversations typically unfold over a series of posts and replies to those posts, forming a tree structure within which individual posts may refer to semantic context from elsewhere in the tree. Such semantic cross-referencing makes it difficult to understand a single post by itself; yet considering the entire conversation tree is not only difficult to scale but can also be misleading as a single conversation may have several distinct threads or points, not all of which are relevant to the post being considered. In this paper, we propose a Graph-based Attentive Semantic COntext Modeling (GASCOM) framework for online conversation understanding. Specifically, we design two novel algorithms that utilize both the graph structure of the online conversation as well as the semantic information from individual posts for retrieving relevant context nodes from the whole conversation. We further design a token-level multi-head graph attention mechanism to pay different attentions to different tokens from different selected context utterances for fine-grained conversation context modelling. Using this semantic conversational context, we re-examine two well-studied problems: polarity prediction and hate speech detection. Our proposed framework significantly outperforms state-of-the-art methods on both tasks, improving macro-F1 scores by 4.5% for polarity prediction and by 5% for hate speech detection. The GASCOM context weights also enhance interpretability.

IFusionQuad: A novel framework for improved aspect-based sentiment quadruple analysis in dialogue contexts with advanced feature integration and contextual CloBlock

Aspect-based sentiment analysis (ABSA) represents a crucial field of natural language processing (NLP). It focuses on deriving detailed sentiment insights from textual content. Dialogue-level aspect-based sentiment quadruple extraction (DiaASQ) is specifically concerned with pinpointing target-aspect-opinion-emotion quadruples within conversations. DiaASQ is important in industries like e-commerce, social media analytics, and customer feedback. However, Current ABSA approaches predominantly focus on single-text scenarios, often overlooking the complexities involved in sentiment analysis within conversational contexts. To fill this gap, this paper presents the IFusionQuad model, which is specifically designed for the DiaASQ task. Our contributions include the innovative integration of CloBlock in ABSA, enhancing feature representation with context-aware weights. The InteractiveNet Fusion Module further advances dialogue understanding by aggregating dialogue-specific features such as threads, speakers, and replies. Components such as CloBlock, gating mechanism, and Biaffine attention effectively mitigate data noise issues, improving the relevance of feature extraction. Empirical evaluation on standard datasets demonstrates that the IFusionQuad model outperforms baseline methods, achieving substantial improvements in quadruple extraction. Specifically, our model shows a 6.59% increase in micro F1 and a 7.05% increase in identification F1 for Chinese datasets, and a 2.65% and 4.69% increase in micro F1 and identification F1, respectively, for English datasets. The results clearly demonstrate our IFusionQuad model’s efficacy, which consistently outperforms baseline models across all evaluation datasets on the DiaASQ task.

Revisiting Conversation Discourse for Dialogue Disentanglement

Dialogue disentanglement aims to detach the chronologically ordered utterances into several independent sessions. Conversation utterances are essentially organized and described by the underlying discourse, and thus dialogue disentanglement requires the full understanding and harnessing of the intrinsic discourse attribute. In this paper, we propose enhancing dialogue disentanglement by taking full advantage of the dialogue discourse characteristics. First of all, in feature encoding stage , we construct the heterogeneous graph representations to model the various dialogue-specific discourse structural features, including the static speaker-role structures (i.e., speaker-utterance and speaker-mentioning structure) and the dynamic contextual structures (i.e., the utterance-distance and partial-replying structure). We then develop a structure-aware framework to integrate the rich structural features for better modeling the conversational semantic context. Second, in model learning stage , we perform optimization with a hierarchical ranking loss mechanism, which groups dialogue utterances into different discourse levels and carries training covering pair-wise and session-wise levels hierarchically. Third, in inference stage , we devise an easy-first decoding algorithm, which performs utterance pairing under the easy-to-hard manner with a global context, breaking the constraint of traditional sequential decoding order. On two benchmark datasets, our overall system achieves new state-of-the-art performances on all evaluations. In-depth analyses further demonstrate the efficacy of each proposed idea and also reveal how our methods help advance the task. Our work has great potential to facilitate broader multi-party multi-thread dialogue applications.

Towards a Conversational Ethics of Large Language Models

Abstract Large Language Models are one of the most prominent examples of current uses of AI, and one of the most urgently pursued normative tasks is to make them and their interactive user-interfaces—open-domain chatbots—safe. However, in this paper, we elaborate first on why such a limited view on the permissibility and desirability of their utterances falls conceptually flat, is philosophically insufficient, and leads to severe technological limits. We then propose a positive normative concept, appropriateness, that can provide the required orientation for determining what chatbots may or may not say, and distinguish between technical-discursive, social, and moral appropriateness. Lastly, we explore the requirements for social and conversational contexts to fulfill appropriateness-requirements, and find that positionality is a key element for their position in discourse.

Adapting the GPT engine for proactive customer insight extraction in product development

Understanding the perceptions of consumers is widely recognized as a critical element in the innovation of new products. Traditional techniques used by companies to collect essential consumer insights have largely remained unchanged. Common practices like interviews and surveys have distinct limitations. Interviews may not always capture the precise needs of consumers due to communication barriers, and surveys tend to encourage incremental changes rather than radical innovations. Service industries face further complexity as they navigate customer feedback on the more intangible aspects of service quality. This study highlights the pioneering use of GPT-3.5 Turbo, a tool known for its exceptional ability to delve into the nuances of conversational context and process data in a chat-centric manner, thereby enhancing the extraction of Voice of the Customer (VoC). Its capability to handle large volumes of data in multiple languages leads to a more thorough and inclusive VoC analysis. The study links these technological advancements with Lean Six Sigma 4.0, suggesting that incorporating GPT-3.5 Turbo could significantly improve the customer-focused strategies in the current industrial landscape. This breakthrough in VoC analysis suggests possibilities for more perceptive, immediate data-driven approaches in customer service, and lays a stronger groundwork for decisions in product evolution and process enhancement. The paper concludes by urging further investigation to confirm these initial results and to explore the ethical aspects of employing such advanced natural language processing technologies.

Meaning Shifts of Culture-Specific Expressions: A Case Study of Facebook

This research aimed to examine the effects of meaning shifts in culture-specific content on Facebook. Specifically, it sought to explore the linguistic, cultural, and grammatical similarities between English and Kiswahili, while highlighting the significant mismatches between these two languages. The study used the theory of Principle of Cultural Translation to provide the theoretical relevance. The target population comprised Facebook pages with a significant following and a focus on CSIs posted within the context of conversations on those pages. The study aimed to examine CSIs posted on Facebook pages based in Kenya on foods, fashion and cultural architecture where English-Kiswahili translations were assessed for errors. The study employed purposive sampling to select a sample of 11 Facebook pages for content analysis. Primary data was gathered through the content analysis method, which involved the systematic preparation, organization, and reporting of the data collected. The study found that lexical, rhetorical, and grammatical mis-matches contributed to misunderstandings and misrepresentations in a variety of domains, including food, fashion, and cultural architecture. Based on the analysis, 80% of the Facebook sources revealed food items like "ugali" and "Nyama Choma" were often reduced to overly simplified translations like "cornmeal" and "grilled meat," failing to capture their rich cultural significance. With regards of fashion, the study found that 70% of the Facebook pages mistranslated the garments such as "kanga" led to a loss of their deeper meanings, contributing to cultural appropriation. Similarly, architectural terms suffered from meaning shifts; 55% of the sources depicted that terms describing unique Kenyan architectural styles were reduced to words like "huts," ignoring the cultural and historical narratives they encapsulate. The study concludes that the meaning shifts of culture-specific expressions related to food, fashion, and architecture in Kenya on Facebook have various negative implications as it leads to misunderstandings that go beyond mere words. The study thus recommends that stickers that contain only the most basic information should be made more useful by scaling them down or making them smaller while figurative language, such as idioms, should be used to describe complex ideas. Keywords: Culture-specific expressions, Meaning shifts, Mistranslation, Lexical mismatches, Facebook content

EAV: EEG-Audio-Video Dataset for Emotion Recognition in Conversational Contexts

Understanding emotional states is pivotal for the development of next-generation human-machine interfaces. Human behaviors in social interactions have resulted in psycho-physiological processes influenced by perceptual inputs. Therefore, efforts to comprehend brain functions and human behavior could potentially catalyze the development of AI models with human-like attributes. In this study, we introduce a multimodal emotion dataset comprising data from 30-channel electroencephalography (EEG), audio, and video recordings from 42 participants. Each participant engaged in a cue-based conversation scenario, eliciting five distinct emotions: neutral, anger, happiness, sadness, and calmness. Throughout the experiment, each participant contributed 200 interactions, which encompassed both listening and speaking. This resulted in a cumulative total of 8,400 interactions across all participants. We evaluated the baseline performance of emotion recognition for each modality using established deep neural network (DNN) methods. The Emotion in EEG-Audio-Visual (EAV) dataset represents the first public dataset to incorporate three primary modalities for emotion recognition within a conversational context. We anticipate that this dataset will make significant contributions to the modeling of the human emotional process, encompassing both fundamental neuroscience and machine learning viewpoints.

An Accurate and Rapidly Calibrating Speech Neuroprosthesis

BackgroundBrain–computer interfaces can enable communication for people with paralysis by transforming cortical activity associated with attempted speech into text on a computer screen. Communication with brain–computer interfaces has been restricted by extensive training requirements and limited accuracy.MethodsA 45-year-old man with amyotrophic lateral sclerosis (ALS) with tetraparesis and severe dysarthria underwent surgical implantation of four microelectrode arrays into his left ventral precentral gyrus 5 years after the onset of the illness; these arrays recorded neural activity from 256 intracortical electrodes. We report the results of decoding his cortical neural activity as he attempted to speak in both prompted and unstructured conversational contexts. Decoded words were displayed on a screen and then vocalized with the use of text-to-speech software designed to sound like his pre-ALS voice.ResultsOn the first day of use (25 days after surgery), the neuroprosthesis achieved 99.6% accuracy with a 50-word vocabulary. Calibration of the neuroprosthesis required 30 minutes of cortical recordings while the participant attempted to speak, followed by subsequent processing. On the second day, after 1.4 additional hours of system training, the neuroprosthesis achieved 90.2% accuracy using a 125,000-word vocabulary. With further training data, the neuroprosthesis sustained 97.5% accuracy over a period of 8.4 months after surgical implantation, and the participant used it to communicate in self-paced conversations at a rate of approximately 32 words per minute for more than 248 cumulative hours.ConclusionsIn a person with ALS and severe dysarthria, an intracortical speech neuroprosthesis reached a level of performance suitable to restore conversational communication after brief training. (Funded by the Office of the Assistant Secretary of Defense for Health Affairs and others; BrainGate2 ClinicalTrials.gov number, NCT00912041.)

Assessing language models’ task and language transfer capabilities for sentiment analysis in dialog data

Our work explores the differences between GRU-based and transformer-based approaches in the context of sentiment analysis on text dialog. In addition to the overall performance on the downstream task, we assess the knowledge transfer capabilities of the models by applying a thorough zero-shot analysis at task level, and on the cross-lingual performance between five European languages. The ability to generalize over different tasks and languages is of high importance, as the data needed for a particular application may be scarce or non existent. We perform evaluations on both known benchmark datasets and a novel synthetic dataset for dialog data, containing Romanian call-center conversations. We study the most appropriate combination of synthetic and real data for fine-tuning on the downstream task, enabling our models to perform in low-resource environments. We leverage the informative power of the conversational context, showing that appending the previous four utterances of the same speaker to the input sequence has the greatest benefit on the inference performance. The cross-lingual and cross-task evaluations have shown that the transformer-based models possess superior transfer abilities to the GRU model, especially in the zero-shot setting. Considering its prior intensive fine-tuning on multiple labeled datasets for various tasks, FLAN-T5 excels in the zero-shot task experiments, obtaining a zero-shot accuracy of 51.27% on the IEMOCAP dataset, alongside the classical BERT that obtained the highest zero-shot accuracy on the MELD dataset with 55.08%.

Intent mining framework for understanding online conversations on vaping to inform social media-based intervention design

ABSTRACT The recent surge in the usage of e-cigarettes amongst youth has highlighted a long-standing societal crisis. To assist public health agencies in policymaking, past research often employed traditional survey-based methods to understand youth behavior, which suffer from response biases and scalability, are time-consuming, and their findings often lag the fast-changing public behavior. Our study fills this gap by using social media as a complementary data source to understand user intentions for vape usage at a large scale, thus, providing an alternative to traditional survey-based methods. In this paper, we propose a novel user intent mining framework under the guidance of social cognitive theory for health behavioral interventions that helps study user intentions across different social media platforms. We then employ this framework to investigate the feasibility of automated intent mining on social media by formulating a multi-class classification task, employing machine learning algorithms to classify a social media message across relevant intent classes: Accusational, Anecdotal, Informational, Justificational and Promotional. The analyses indicate that Accusational tweets and Anecdotal messages were most prevalent on X/Twitter and Reddit respectively. We further provide novel insights on the conversational context using topic modeling analysis and psychometric analysis consequently, informing intervention designs and assisting health analysts.

Using ChatGPT as a Tutor to learn Mathematics Application Program

The ability of ChatGPT to understand conversational context and provide relevant responses makes it very useful in human-computer interactions. ChatGPT can be a very useful tool for mathematics teachers in their role as tutors to learn and teach various applications related to mathematics. In this digital era, mathematical applications and software like GeoGebra, Matlab, Augmented Reality, and SPSS are becoming increasingly important in teaching and learning mathematics. ChatGPT can help teachers understand how to use these tools by providing detailed explanations, tutorials, and examples of their application in teaching. This research aims to determine the effectiveness and efficiency of ChatGPT in helping teachers understand and apply technologies such as GeoGebra, Matlab, Augmented Reality, and SPSS. Participants in this case study research consisted of 5 teachers. The categories of participant questions asked to ChatGPT included (1) General explanation of the application, (2) How to download the application, (3) Equipment specifications required, (4) General use of the application, and (5) Requesting sample commands in the application. The results of this study indicate that participants will successfully use ChatGPT as a tutor for learning Mathematics Application Programs if they have the necessary equipment specifications required by the applications and are able to download the applications.

Speech Emotion Recognition Incorporating Relative Difficulty and Labeling Reliability.

Emotions in speech are expressed in various ways, and the speech emotion recognition (SER) model may perform poorly on unseen corpora that contain different emotional factors from those expressed in training databases. To construct an SER model robust to unseen corpora, regularization approaches or metric losses have been studied. In this paper, we propose an SER method that incorporates relative difficulty and labeling reliability of each training sample. Inspired by the Proxy-Anchor loss, we propose a novel loss function which gives higher gradients to the samples for which the emotion labels are more difficult to estimate among those in the given minibatch. Since the annotators may label the emotion based on the emotional expression which resides in the conversational context or other modality but is not apparent in the given speech utterance, some of the emotional labels may not be reliable and these unreliable labels may affect the proposed loss function more severely. In this regard, we propose to apply label smoothing for the samples misclassified by a pre-trained SER model. Experimental results showed that the performance of the SER on unseen corpora was improved by adopting the proposed loss function with label smoothing on the misclassified data.

A multi-level multi-task progressive framework based on relational graph convolutional networks for causal emotion entailment

Causal Emotion Entailment (CEE) aims to identify the corresponding causal utterances given a non-neutral target utterance and its conversational history. Previous studies have been devoted to modeling conversational context by infusing emotion and speaker information separately. However, for target utterances with different emotions, the joint distributions of speaker identities and emotion types for corresponding causal utterances are significantly different. In addition, a single non-neutral target utterance may correspond to multiple causal utterances and vice versa (i.e., one-to-many and many-to-one relations). These factors make the corresponding relation between the non-neutral target utterances and causal utterances more diverse and difficult to model, hindering the performance of existing methods. To this end, we design an Utterance-level Interaction Graph (UIG) and a Pair-level Interaction Graph (PIG). The former explicitly captures differences in the joint distribution of speaker identities and emotion types for target utterances with different emotions by utterance-level interactions. The latter models the correlations between utterances resulting from one-to-many and many-to-one relations by pair-level interactions. Furthermore, we propose a Multi-level Multi-task Progressive Framework (MMPF) for the CEE task, which achieves UIG and PIG using Relational Graph Convolutional Networks. Extensive experiments on RECCON-DD datasets demonstrate that MMPF obtains state-of-the-art performance.

Culture and class in contemporary Chinese psychotherapy

Working with US‐based psychoanalysts who instruct Chinese psychoanalytic trainees, some reject any discussion of ‘culture’ as a means by which patients avoid difficult material, whilst others draw a sharp contrast between ‘the Chinese mind’ and ‘the Western mind’. My discussion with Chinese psychoanalytic trainees suggests that although they often use a ‘China/West’ dichotomy to understand clinical work, they are more likely than their predominantly Euro‐American analytic trainers to look at this distinction in fluid terms affected by other differences. They often draw attention to the ‘type’ of person who is the typical client of psychoanalytic therapy in contemporary China: young, urban, cosmopolitan and financially solvent. This article investigates how Chinese psychoanalytic trainees’ perceptions of relevant differences in their practice vary depending on the conversational context, challenging both the universalist claims of traditional psychoanalytic theory and the recent emphasis on cultural differences in the field.

CitizenConnectBot(CCB): The interactive guide to government schemes

In the ever-evolving landscape of digital governance, the Generative CitizenConnectBot (CCB) emerges as a transformative tool, seamlessly bridging the gap between citizens and government schemes.The absence of real-time updates results in outdated information, undermining the effectiveness of the service. Moreover, the current translation feature limits the accessibility of scheme details for a diverse audience, creating barriers to understanding and participation. Additionally, the lack of a chatbot on schemes dissemination in the available websites complicates the process for citizens to obtain necessary information and determine eligibility, impeding their ability to benefit from available support.The CitizenConnectBot(CCB), integrated into a user-friendly website, acts as a comprehensive information hub, offering citizens tailored insights into various government initiatives.The bot's adaptability ensures that information is generated according to the unique requirements of each user.This paper employs natural language processing (NLP) techniques, including text classification with SVM, TF-IDF vectorization, regular expressions for pattern matching, and session management for maintaining conversation context.