通过对话文本分析学习者元认知

No abstract available

No abstract available

Stronger metacognition, or awareness and regulation of thinking, is related to higher academic achievement. Most metacognition research has focused at the level of the individual learner. However, a few studies have shown that students working in small groups can stimulate metacognition in one another, leading to improved learning. Given the increased adoption of interactive group work in life science classrooms, there is a need to study the role of social metacognition, or the awareness and regulation of the thinking of others, in this context. Guided by the frameworks of social metacognition and evidence-based reasoning, we asked: 1) What metacognitive utterances (words, phrases, statements, or questions) do students use during small-group problem solving in an upper-division biology course? 2) Which metacognitive utterances are associated with small groups sharing higher-quality reasoning in an upper-division biology classroom? We used discourse analysis to examine transcripts from two groups of three students during breakout sessions. By coding for metacognition, we identified seven types of metacognitive utterances. By coding for reasoning, we uncovered four categories of metacognitive utterances associated with higher-quality reasoning. We offer suggestions for life science educators interested in promoting social metacognition during small-group problem solving.

No abstract available

This study investigates designs for developing knowledge building (KB) and higher order competencies among academically low-achieving students. Thirty-seven low-achieving students from a ninth-grade visual arts course in Hong Kong participated. The design involved principle-based KB pedagogy, with students writing on Knowledge Forum® (KF), enriched by analytics-supported reflective assessment. Analysis of the discourse on KF showed that the low achievers were able to engage in productive discourse, with evidence of metacognitive, collaborative, and epistemic inquiry. Analysis illustrates how the design supported student engagement, including (1) reflective inquiry and social metacognition; (2) reflective meta- and epistemic talk; (3) evidence-based reflection for collective growth; and (4) reflection embedded in community ethos. Implications of reflective assessment for supporting low achievers for inquiry learning and KB are discussed.

Metacognition refers to a person’s ability to understand and regulate their thinking and learning. For students, metacognitive skills increase awareness of their thought processes. These skills can impact how new and old information is processed and stored in memory and how new and old information is accessed or applied. This research study focuses on identifying metacognitive themes that emerged during debriefing sessions following an interprofessional healthcare simulation which used standardized patients.Teams of health professional students from eight disciplines collaborated to assess and care for a standardized patient in a simulated environment. Following each simulation session, student teams debriefed their experiences and learning outcomes. Debriefing conversations from four independent teams were transcribed and common metacognitive themes were determined by consensus.The themes that emerged were: 1) Collaboration, 2) Peer-to-peer learning, 3) Problem-Solving, and 4) Self-Reflection. Together, these themes suggested that participating students applied metacognitive processes during the team-based simulation session debriefs.Metacognition and metacognitive ability are important learning tools that can be incorporated into interprofessional learning environments through instructional and facilitation strategies. Interprofessional team simulations with standardized patients  provide an optimal activity for encouraging critical discourse and other metacognitive processes.

Contributions: To address the interpretability issues in artificial intelligence (AI)-powered classroom discourse models, we employ explainable AI methods to unpack classroom discourse analysis from deep learning-based models and evaluate the effects of model explanations on STEM teachers. Background: Deep learning techniques have been used to automatically analyze classroom dialogue to provide feedback for teachers. However, these complex models operate as black boxes, lacking clear explanations of the analysis, which may lead teachers, particularly those lacking AI knowledge, to distrust the models and hinder their teaching practice. Therefore, it is crucial to address the interpretability issue in AI-powered classroom discourse models. Research Questions: How to explain deep learning-based classroom discourse models using explainable AI methods? What is the effect of these explanations on teachers’ trust in and technology acceptance of the models? How do teachers perceive the explanations of deep learning-based classroom discourse models? Method: Two explainable AI methods were employed to interpret deep learning-based models that analyzed teacher and student talk moves. A pilot study was conducted, involving seven STEM teachers interested in learning talk moves and receiving classroom discourse analysis. The study assessed changes in teachers’ trust and technology acceptance before and after receiving model explanations. Teachers’ perceptions of the model explanations were investigated. Findings: The AI-powered classroom discourse models were effectively explained using explainable AI methods. The model explanations enhanced teachers’ trust and technology acceptance of the classroom discourse models. The seven STEM teachers expressed satisfaction with the explanations and provided their perception of model explanations.

No abstract available

No abstract available

No abstract available

Abstract This study investigated metacognitive skills manifested during collaborative problem-solving activities in mathematics. It specifically focused on seventh-grade students attending a public secondary school in the Kayonza district in the Eastern Province of Rwanda. Discussion protocols were analyzed using a devised framework, yielding three main findings. First, instances were observed wherein a select few group members led the discussion, triggering metacognition among the remaining members. Second, in situations where individual learning stalled, dialogue proved instrumental in activating interaction among group members by posing a simple question. Third, teachers’ questioning, which hints at how to approach problems, emerged as a crucial factor. Notably, it is suggested that teacher-directed questioning directed at slower learners can promote metacognition across the group. This represents only one empirical study in Rwanda; further case studies are needed. Furthermore, research on the effects of teacher interventions on group learning is necessary.

Intelligence augmentation can offer personalized learning resources and pathways tailored to each student’s unique characteristics and needs. Among these advancements, the large language model (LLM) agent has ushered in a new revolution in education. In this study, we constructed a metacognitive reflective learning scaffold (MRLS) grounded in metacognitive theory and reflective learning principles to provide conceptual support for students during their reflective practices. In addition, we developed a metacognitive reflective learning agent (MRLA) on the Coze platform designed to deliver personalized guidance and assistance throughout the reflective learning process. We conducted a 16-week $2 \times 2$ quasi-experiment study at Z University in China, where participants were randomly assigned to four groups. Throughout the research process, we collected dialogue data from students using the Coze platform, as well as reflection reports submitted via the XueXiTong platform for quantitative analysis. Empirical results demonstrated that both the MRLS and MRLA significantly enhanced students’ metacognition, indicated that the MRLS offers precise guidance for students’ reflective learning processes, enabling them to better comprehend and articulate their reflections. The MRLA equips students with more convenient, efficient, and intelligent resources, significantly augmenting the provision of metacognitive training support that would otherwise be provided by teachers. This study emphasizes the validity and necessity of MRLS and MRLA for the cultivation of students’ metacognitive ability and provides insights for the future application of LLM agent and learning scaffolds for optimizing students’ learning process.

Deep learning is a critical goal of education in the 21st century, and metacognitive scaffolding plays a very key auxiliary role in achieving this goal. To exceed the limitations of traditional teaching scaffolds that lacked customization, they developed a dynamic scaffold framework based on metacognitive theory and generative artificial intelligence technology. Its structure is mainly composed of behavioral examination research, creative dialogue exchanges, and responses to self-awareness. On the Coze platform, users can build an intelligent agent, select a "three-person voting setup" and a "four-person quiz setup" scenario, and then conduct educational exercises. The results showed that this structure enabled learners to perform superior in goal setting, strategy application, and reflective standard, which shows that self-awareness assistance using general artificial intelligence is indeed very promising in helping people to gain deeper understanding.

Analysing classroom dialogue is a widely used approach for understanding students' learning, often requiring team‐based collaborative research. This presents a challenge for single researchers due to the labour‐intensive nature of the process. Emerging advancements in large language models (LLMs) such as ChatGPT, enhance qualitative research, particularly in inductive and deductive coding tasks.This study investigates the feasibility of a single researcher, the author of this study, collaborating with ChatGPT‐4o for qualitative coding of classroom dialogue data. The goal is to develop effective human–ChatGPT co‐coding methods and explore how such collaboration can enhance qualitative coding practices and provide insights into students' dialogue patterns.The study analysed 1287 utterances from middle school science classes using a mixed‐method approach. A new codebook was developed through an inductive process using ChatGPT, followed by deductive coding conducted by both the researcher and ChatGPT. Kappa values were compared between human–human and human–ChatGPT coding. Disagreements in code assignments were resolved by the researcher, with reference to ChatGPT's rationale. Coded utterances were analysed using ordered network analysis (ONA) to visualise dialogue patterns in classes.The coding conducted by the researcher and ChatGPT resulted in a Cohen's kappa of 0.56, with the highest level of disagreement observed in the category of Meta‐cognition. The inductively co‐developed codebook helped uncover students dialogue patterns during experimental activities. Although ChatGPT exhibited limitations in interpreting nuanced and context‐dependent utterances, the findings highlight its potential as a valuable collaborator for solo researchers by supporting cognitive processes such as reflective interpretation and the development of new perspectives.

ABSTRACT Recent research has highlighted the critical role in children’s cognitive development of the metacognitive support parents give their children during everyday interactions. Our main goal was to examine whether parents made consistent use of metacognitive talk across different parent – child interaction contexts and to document the effect of this metacognitive talk across contexts on preschoolers’ memory. The relation between children’s and parents’ metacognitive talk was also examined, along with the link between children’s metacognitive talk and their own memory performance. 64 Belgian preschooler – parent dyads (Mage = 40.3 months) were invited to play memory games together and to reminisce about a standardized event. In both contexts, parents’ and children’s utterances were analyzed for their metacognitive content. Children’s episodic memory was assessed through a story-recall task. Results revealed that, in terms of frequency of use, parents exhibited a somewhat consistent metacognitive style across contexts. The nature of the metacognitive comments that were produced, however, varied depending on the context. Moreover, parents’ consistent production of higher rates of metacognitive comments in both contexts was found to be associated with better memory performance in children. Our data also revealed that children’s tendency to make metacognitive comments was related to their parents’ metacognitive production and to their own memory performance. By documenting the relations between parents’ and children’s metacognitive talk and the effect on children’s memory, this study provides promising avenues for a better understanding of how metacognitive processes might be involved in early memory development.

Effective feedback is essential for refining instructional practices in mathematics education, and researchers often turn to advanced natural language processing (NLP) models to analyze classroom dialogues from multiple perspectives. However, utterance-level discourse analysis encounters two primary challenges: (1) multifunctionality, where a single utterance may serve multiple purposes that a single tag cannot capture, and (2) the exclusion of many utterances from domain-specific discourse move classifications, leading to their omission in feedback. To address these challenges, we proposed a multi-perspective discourse analysis that integrates domain-specific talk moves with dialogue act (using the flattened multi-functional SWBD-MASL schema with 43 tags) and discourse relation (applying Segmented Discourse Representation Theory with 16 relations). Our top-down analysis framework enables a comprehensive understanding of utterances that contain talk moves, as well as utterances that do not contain talk moves. This is applied to two mathematics education datasets: TalkMoves (teaching) and SAGA22 (tutoring). Through distributional unigram analysis, sequential talk move analysis, and multi-view deep dive, we discovered meaningful discourse patterns, and revealed the vital role of utterances without talk moves, demonstrating that these utterances, far from being mere fillers, serve crucial functions in guiding, acknowledging, and structuring classroom discourse. These insights underscore the importance of incorporating discourse relations and dialogue acts into AI-assisted education systems to enhance feedback and create more responsive learning environments. Our framework may prove helpful for providing human educator feedback, but also aiding in the development of AI agents that can effectively emulate the roles of both educators and students.

Recently, several multi-turn dialogue benchmarks have been proposed to evaluate the conversational abilities of large language models (LLMs). As LLMs are increasingly recognized as a key technology for advancing intelligent education, owing to their ability to deeply understand instructional contexts and provide personalized guidance, the construction of dedicated teacher-student dialogue benchmarks has become particularly important. To this end, we present EduDial, a comprehensive multi-turn teacher-student dialogue dataset. EduDial covers 345 core knowledge points and consists of 34,250 dialogue sessions generated through interactions between teacher and student agents. Its design is guided by Bloom's taxonomy of educational objectives and incorporates ten questioning strategies, including situational questioning, zone of proximal development (ZPD) questioning, and metacognitive questioning-thus better capturing authentic classroom interactions. Furthermore, we design differentiated teaching strategies for students at different cognitive levels, thereby providing more targeted teaching guidance. Building on EduDial, we further develop EduDial-LLM 32B via training and propose an 11-dimensional evaluation framework that systematically measures the teaching abilities of LLMs, encompassing both overall teaching quality and content quality. Experiments on 17 mainstream LLMs reveal that most models struggle in student-centered teaching scenarios, whereas our EduDial-LLM achieves significant gains, consistently outperforming all baselines across all metrics. The code is available at https://github.com/Mind-Lab-ECNU/EduDial/tree/main.

Abstract This study examined the metacognitive knowledge of Finnish comprehensive school students and explored whether grade-based differences in metacognitive knowledge exist among 6th, 7th and 9th graders. Integrating qualitative (declarative, procedural, conditional) and contextual (person, task, strategy) frameworks, the research aims for a comprehensive understanding. We employed mixed-methods approach; qualitative data from 225 student interviews underwent qualitative theory-driven analysis, followed by quantitative methods to explore variations and relationships. Results showed prevalent procedural and strategic metacognitive knowledge. Utterances on problem-solving often linked to task-specific knowledge, emphasising the need for a varied metacognitive knowledge set. 9th graders excelled in explaining strategy use, while 7th graders demonstrated proficiency in understanding when and why to employ specific strategies. Although a qualitative level approach can aid in understanding the development of metacognitive knowledge, combining qualitative and contextual conceptualisations provides a better overview of metacognitive knowledge. This study suggests that metacognitive knowledge-supporting elements are needed in math learning materials and teachers’ pedagogical actions.

Collaborative problem solving (CPS) is a critical competency in the Artificial intelligence (AI) era, requiring the integration of cognitive and social skills through real-time dialogue and coordination. While prior studies have explored CPS behaviours using human-coded text from online platforms, limited research has examined how machine learning (ML) and deep learning (DL) models perform on spoken peer dialogue in face-to-face (F2F) classroom settings. This study investigates the automatic classification of CPS phases using a validated coding framework applied to two classroom tasks—one supported by a GenAI assistant and one not. A total of 7,744 utterances were manually labelled across nine CPS subskills and three broader facets. Six ML and five DL models were evaluated, including lightweight BERT variants combined with various classifiers. Results show that BERT-based models significantly outperform traditional ML approaches. Specifically, BERT+ANN achieved better overall performance in smaller, imbalanced datasets, while BERT+CNN performed better in larger datasets. Reducing label granularity from nine subskills to three facets consistently improved classification accuracy and F1 scores. Both models achieved AUROC scores around 0.90, indicating strong discriminative capability. Several key insights emerged from the findings: Model architecture matters: Simpler classifiers like ANN preserve BERT’s semantic representations and offer stable performance, especially in smaller or imbalanced datasets. Task context influences CPS behaviour: Different tasks elicit distinct CPS skill distributions, with task regulation dominating in technical tasks and communicative participation more prevalent in reflective tasks. Label granularity affects performance: Reducing the number of classification labels (e.g., from 9 subskills to 3 facets) significantly improves model accuracy and generalizability. Lightweight models are viable: Even with a reduced-capacity BERT model, competitive performance was achieved, suggesting potential for real-time, resource-efficient deployment in educational settings. This study contributes to educational AI by introducing a novel oral CPS dataset, benchmarking multiple models, and demonstrating the feasibility of lightweight architectures for real-time deployment. Limitations include the small sample size and single-modality input. Future work should explore multimodal features, larger and more diverse classrooms, and teacher-facing dashboards for actionable feedback. The findings support the development of scalable, ethical, and human-centered learning analytics tools that enhance collaborative learning in AI-enhanced education.

The growing use of virtual learning environments (VLEs) in primary education offers new opportunities for enhancing student learning. However, understanding and analysing student behaviour in these environments is still challenging, especially in collaborative science education. This research aims to develop and evaluate a multimodal learning analytics (MMLA) approach tailored for primary science education. The study will focus on four key questions: understanding the current state of learning analytics (LA) in VLEs, identifying which data types in MMLA most effectively contribute to insights into student learning behaviours, creating an MMLA approach to support collaborative problem solving (CPS) and metacognitive strategies in VLEs, and improving the visualization of MMLA results for educators and students. To achieve these goals, the research will use a series of studies that collect and analyse multimodal data, including eye-tracking, behaviour logs, and dialogue text. Machine learning and deep learning techniques will be applied to identify critical data types, and these insights will inform the creation of an MMLA approach specifically designed to support CPS and metacognitive strategies. A user-centred design will guide the creation of a visualisation dashboard. This research is expected to contribute to theory by expanding the application of the MMLA approach, critically reflecting on the potential to innovate CPS and metacognitive strategies within VLEs and improving data analysis and dashboard design, and to practice by enhancing tools for primary science education.

In the learning of concepts related to the processes of separating mixtures in elementary school (final years) some difficulties may arise related to the lack of use of some metacognitive strategies that help in the approximation of a meaningful learning. In this context, it becomes relevant to elaborate didactic intervention proposals that show students metacognitive strategies that facilitate learning and provide autonomy. This article aims to raise an environment for reflection and dialogue in the face of an intervention proposal that establishes, in a structured way, a sequence of metacognitive questions linked to the learning of the aforementioned concepts. The results show some potential that can be generated with the implementation of the suggested didactic intervention, except for the necessary adaptations to the school context in which the proposal is intended to be applied.

Automating the classification of instructional strategies from a large-scale online tutorial dialogue corpus is indispensable to the design of dialogue-based intelligent tutoring systems. Despite many existing studies employing supervised machine learning (ML) models to automate the classification process, they concluded that building a well-performed ML model is nontrivial since the sample size is commonly limited for model training. Based on this fact, we posited that the model performance can be further optimized by the design of input features and the selection of informative instances for model training. By reviewing the existing dialogue-related research, we found that contextual information (i.e., the content of preceding utterances) was a potential feature in many text classification tasks but underexplored in the classification of an instructional strategy. In addition, training a well-performed ML model (e.g., deep neural network) to recognize instructional strategies usually requires a large amount of manually annotated data, which are labor intensive. To alleviate the demand of manual annotation, researchers proposed the use of statistical active learning methods to select informative instances, but this method was rarely used for recognizing the instructional strategy in online tutoring dialogue. Therefore, our study aimed to investigate the improvement of automating the classification of dialogue acts—a popular approach to the detection of instructional strategies—from two perspectives. First, we explored whether and to what extent the incorporation of contextual information can boost a model's prediction performance. Then, we investigated the extent to which the recent active learning methods can alleviate the labor-intensive issues in training the ML model for recognizing the instructional strategies. Our study showed that: 1) the ML models trained on the features that included the contextual information achieved better performance than that of the models excluding it; 2) the effectiveness of the contextual information decayed after the ML model achieved an optimal performance; and 3) compared with the random baseline, active learning methods can select informative samples from the training dataset to train ML models, which can alleviate the labor-intensive issues.

No abstract available

Abstract The first purpose of the present study was to characterize progressive and metacognitive dialogue in knowledge-building classrooms. The second purpose was to examine the association between the quality of classroom dialogue and students’ learning outcomes, including domain knowledge and conception of collaboration. One hundred thirty college students and their teacher participated in this study. A total of 73 lessons were video-recorded. Qualitative analysis of classroom dialogue identified dialogic moves into three categories—articulation and elaboration, building-on and connection, and reflection and coordination. The findings provide a rich illustration of how students engage in metacognitive and progressive dialogue. Quantitative analysis indicated that student engagement in classroom dialogue was positively associated with their domain knowledge and conception of collaboration. The results revealed that considering student involvement in classroom dialogue is essential for understanding how knowledge building occurs in classroom contexts and how collective knowledge and conception of collaboration are developed through classroom dialogue.

Abstract Reciprocal teaching (RT) involves collaborative dialogue between the teacher and students, with the aim of jointly constructing the meaning from a shared text. This study employs RT as a cognitive and metacognitive set of strategies for enhancing students’ reading comprehension and monitoring skills. Thirty-four students were assigned to the RT group for reading instruction, whereas 30 students were assigned to the control group for regular reading instruction. The four RT strategies (predicting, questioning, clarifying, and summarizing) empowered the students to engage in a non-linear approach to comprehend texts and monitor their reading comprehension. They afforded the students the chance to engage in pre-, during-, and post-reading cognitive and metacognitive processes, collaborate with their peers, read with intentionality and focus, and assume shared responsibility for monitoring their advancement in reading comprehension. The students engaged in the scaffolded activities of read-aloud, dialogue, structured comprehension, and peer teaching to develop their reading comprehension and monitoring skills. The results showed that the RT group outperformed the control group on the post-test in both total score of reading comprehension and individual reading skills. Therefore, language instructors are recommended to use an array of effective RT strategies to enhance their students’ reading comprehension and monitoring skills.

The core challenge in learning has shifted from knowledge acquisition to effective Self-Regulated Learning (SRL): planning, monitoring, and reflecting on one's learning. Existing digital tools, however, inadequately support metacognitive reflection. Spaced Repetition Systems (SRS) use de-contextualized review, overlooking the role of context, while Personal Knowledge Management (PKM) tools require high manual maintenance. To address these challenges, this paper introduces"Insight Recall,"a novel paradigm that conceptualizes the context-triggered retrieval of personal past insights as a metacognitive scaffold to promote SRL. We formalize this paradigm using the Just-in-Time Adaptive Intervention (JITAI) framework and implement a prototype system, Irec, to demonstrate its feasibility. At its core, Irec uses a dynamic knowledge graph of the user's learning history. When a user faces a new problem, a hybrid retrieval engine recalls relevant personal"insights."Subsequently, a large language model (LLM) performs a deep similarity assessment to filter and present the most relevant scaffold in a just-in-time manner. To reduce cognitive load, Irec features a human-in-the-loop pipeline for LLM-based knowledge graph construction. We also propose an optional"Guided Inquiry"module, where users can engage in a Socratic dialogue with an expert LLM, using the current problem and recalled insights as context. The contribution of this paper is a solid theoretical framework and a usable system platform for designing next-generation intelligent learning systems that enhance metacognition and self-regulation.

In an English as a foreign language (EFL) environment, many students only get to learn or use English in the classroom. To give those students more authentic experiences in language use, some university courses are adopting hands-on learning approaches such as project-based language learning (PBLL). One of the biggest benefits of PBLL is the role of social interaction as students work together in a situated activity to construct shared understanding through sharing, using, and debating ideas with peers. While this learning approach can enhance students’ language learning through joint collaborative efforts, it can cause a variety of challenges in the process as they try to engage with their peers through their L2. This paper reports on the metacognitive process of Japanese university students of intermediate to advanced proficiency in English in their attempts to engage in collaborative dialogue with their peers during a group project about current international affairs. The interview data collected from the students after the project were analyzed using the Modified Grounded Theory Approach (M-GTA). The results suggest that students constantly struggled to say something convincing and pleasing to peers so as not to disturb the peaceful atmosphere within the group. It was also found that while students were faced with various internal obstacles arising from multiple emotions, they tried hard to maintain good relationships with their peers. They sometimes spoke in Japanese when doing so would facilitate their collaborative dialogue.

No abstract available

No abstract available

Dialogue data has been a key source for understanding learning processes, offering critical insights into how students engage in collaborative discussions and how these interactions shape their knowledge construction. The advent of Large Language Models (LLMs) has introduced promising opportunities for advancing qualitative research, particularly in the automated coding of dialogue data. However, the inherent contextual complexity of dialogue presents unique challenges for these models, especially in understanding and interpreting complex contextual information. This study addresses these challenges by developing a novel LLM-assisted automated coding approach for dialogue data. The novelty of our proposed framework is threefold: 1) We predict the code for an utterance based on dialogue-specific characteristics -- communicative acts and communicative events -- using separate prompts following the role prompts and chain-of-thoughts methods; 2) We engaged multiple LLMs including GPT-4-turbo, GPT-4o, DeepSeek in collaborative code prediction; 3) We leveraged the interrelation between events and acts to implement consistency checking using GPT-4o. In particular, our contextual consistency checking provided a substantial accuracy improvement. We also found the accuracy of act predictions was consistently higher than that of event predictions. This study contributes a new methodological framework for enhancing the precision of automated coding of dialogue data as well as offers a scalable solution for addressing the contextual challenges inherent in dialogue analysis.

As teachers in elementary school classes have limited time for reflection, it becomes crucial to make this process of reflection systematized. Therefore, in this study, we transcribe the utterances into text form during an elementary school mathematics class and analyze the utterances of teachers and children using a dialogue model with a neural network. We here analyze utterance categories and scenes.

Two decades of L2 self-regulated learning (SRL) research has focused heavily on reading and writing, leaving speaking an under-explored area. While speaking is an important yet difficult communicative skill to master that often generates high anxiety for English learners, EFL contexts seldom provide enough practice opportunities beyond the classroom. The study introduces an innovative pedagogy – digital oral dialogue journaling (DODJ) – and investigates its effectiveness in cultivating SRL speaking strategy use in a Hong Kong secondary school. A total of 95 recorded journal videos were created by a class of 19 students and uploaded to the digital platform Flipgrid over five weeks to receive dialogic feedback from the teacher on a weekly basis. Students’ speaking strategy use was conceptualized from a social cognitive SRL framework, and changes were measured using a pre- and post-intervention questionnaire as well as through two semi-structured interviews. Results of repeated measures MANOVA and ANOVA tests reveal significant increases in multiple cognitive and metacognitive strategies such as rehearsing, memorizing, and evaluating. Interview findings further show that certain DODJ design features facilitated the changes, including flexible topic selection and scheduling, unlimited video submission and review, and emoticon masking and dialogic teacher feedback. Pedagogical suggestions are offered on how to design effective DODJ practice to cultivate L2 learners to become strategic, self-regulated speakers.

No abstract available

The relevance of this study is due to the need to improve the methodology of teaching high school students writing in a foreign language. Modern domestic and foreign methods consider the teaching of writing as one of the essential aspects of the practical purpose of teaching a foreign language. Written speech, differing from oral speech in its structure, logic, accuracy and concreteness, places high demands on the writer. The communicative and cognitive need underlying written speech reflects the desire of the writer to convey information, express his opinion, views, and form a certain impression of himself. This need can be effectively realized if the student has developed the skills of independent management of the learning process: planning, selfcontrol and self-assessment. In Russian educational practice, when teaching writing, it is traditionally customary to focus more on the result than on the process. This fact, as well as the complexity and diversity of this type of speech activity, often lead to low efficiency in teaching students writing. An effective way to solve this problem may be to purposefully teach students metacognitive strategies. The study of Russian methodological literature has shown that the problem of formation of foreign language writing skills using metacognitive strategies has not been formulated and sufficiently studied. The aim of the study was to develop a model for teaching writing within the framework of a procedural approach, which does not exclude, but complements the modern rhetorical approach as a step on the way from completely teacher-controlled written assignments to uncontrolled writing. The latter implies the formation of metacognitive learning strategies among schoolchildren, most of which are universal for a wide range of tasks. Familiarization with foreign scientific sources allowed us to identify two main models developed specifically for teaching writing using metacognitive strategies in which the factor of a non-native language was not taken into account – the K. Englert model, the K. Harris and S. Graham model. K. Englert's model emphasizes that internal dialogue, speaking thoughts aloud in the process of writing a text and subsequent discussion allow the student to become aware of previously unconscious cognitive processes and develop metacognitive skills. The model also provides metered adult care. Various supports are offered to help the student use a particular metacognitive strategy (memos, flashcards, computer programs, etc.). Various supports are offered to help the student use a particular metacognitive strategy (memos, flashcards, computer programs, etc.). Teaching writing using metacognitive strategies within the framework of K. Englert's model involves active cooperation of students. A distinctive feature of the K. Harris and S. Graham model is explicit learning strategies in the process of forming writing skills, which scientists divide into two types: writing strategies (including strategies specific to a particular genre, as well as strategies necessary to create any written statement) and self-regulation strategies (goal setting, self-control, self-assessment, management internal dialogue, etc.). The authors of the model emphasize that almost any strategy can be considered metacognitive if it is understood by students. The analysis showed that in the process of teaching foreign language writing, it is important to take into account the following aspects: explicit teaching of metacognitive strategies, scaffolding using supports, group interaction and modeling of internal dialogue in the process of creating a written statement. This allowed the authors to propose their own methodology for teaching metacognitive strategies, which includes six key stages and takes into account all aspects of writing from the point of view of a procedural approach: analysis of the sample proposed by the teacher; group discussion and choice of strategies; modeling the use of strategies; guided practice (feedback from the teacher); independent practice; reflection. The developed methodology takes into account cognitive and motivational-affective processes, as well as the improvement of short-term and long-term memory and the creation of an environment that brings academic writing as close as possible to real written communication. The conducted research allowed us to conclude about the advantages associated with the use of metacognitive strategies in teaching writing at school: the development of self-regulation and self-esteem, which is expressed in the ability to set goals and monitor one's own progress in mastering foreign language writing; improving the ability to interact with educational material, more effective assimilation of material; increasing self-confidence and motivation to further work; the formation of critical thinking skills based on the analysis and evaluation of one's thinking and actions; increasing the accessibility of education in general, since the formed metacognitive skills are successfully applied in various educational contexts and formats (inclusion, distance learning). The skills of independent and responsible attitude to learning acquired by schoolchildren bring them closer to self-regulation of the educational process, to autonomy in mastering academic disciplines.

No abstract available

No abstract available

This study examined the factors of learning behaviors in digital-learning-material reader that predict university students’ productive participation identified by social presence indicators in online collaborative learning. Data were collected from 76 Frist-year university students, including learning behaviors in digital learning-material readers, and frequency of each social presence indicator classified by coding discussion scripts with automated text classification. The results of the stepwise multiple regression analysis revealed that the learning behavior of drawing red markers, highlighting important content predicts social presence. Furthermore, the study’s findings showed that different learning behaviors predict different social presence indicators, which are related to metacognition and critical thinking.

This study explored the application of the Community of Inquiry (CoI) framework in a large enrollment online course, which focuses specifically on collaborative learning and student engagement. The CoI framework, which comprises social, cognitive, and teaching presence, provides a theoretical foundation for understanding how students construct personal meaning and confirm mutual understanding. While metacognition has been an important factor in learning, its role in collaborative online learning environments is poorly understood. This mixed-methods study investigates the impact of shared metacognition on collaboration in a large-enrollment online Academic Writing course at an open distance learning university in South Africa. The study employed the Shared Metacognition survey, which was developed from the CoI framework, to collect data from 1200 students at three stages: pre-module, midpoint, and post-module. Statistical analysis and qualitative content analysis were used to examine self-regulation and co-regulation dimensions of metacognition. The findings highlight the significance of teaching presence in predicting student success and satisfaction. Shared metacognition emerged as a significant factor in developing collaborative learning environments. Students’ awareness of their thinking and learning processes improved through critical discourse and peer interaction. This research contributes to the understanding of collaborative learning in large courses and emphasises the importance of metacognitive awareness and shared regulatory functions. The study has significant implications for lecturers, practitioners, and instructional designers in large modules who seek to enhance collaborative learning experiences in online education.

No abstract available

Metacognitive awareness is knowing about learners’ own thinking and learning, facilitated by introspection and self-evaluation. Although metacognitive functions are personal, they cannot be explained simply by individual conceptions, especially in a collaborative group learning context. This study considers metacognitive awareness on multiple levels. It investigates how metacognitive awareness at the individual, social, and environmental levels are associated with collaborative problem solving (CPS). Seventy-seven higher education students collaborated in triads on a computer-based simulation about running a fictional company for 12 simulated months. The individual level of metacognitive awareness was measured using the Metacognitive Awareness Inventory. The social level of metacognitive awareness was measured multiple times during CPS through situated self-reports, that is, metacognitive judgements and task difficulty. The environmental level of metacognitive awareness was measured via a complex CPS process so that group members’ interactions were video recorded and facial expression data were created by post-processing video-recorded data. Perceived individual and group performance were measured with self-reports at the end of the CPS task. In the analysis, structural equation modelling was conducted to observe the relationships between multiple levels of metacognitive awareness and CPS task performance. Three-level multilevel modelling was also used to understand the effect of environmental-level metacognitive awareness. The results reveal that facial expression recognition makes metacognitive awareness visible in a collaborative context. This study contributes to research on metacognition by displaying both the relatively static and dynamic aspects of metacognitive awareness in CPS.

No abstract available

No abstract available

No abstract available

No abstract available

The development of reflexive skills among students is a central concern in modern educational theory and practice, particularly in the domain of literary education. Reflexivity, often associated with metacognition and critical reflection, empowers learners to examine their thought processes, challenge assumptions, and deepen engagement with texts. In literary studies, reflexivity enables students to situate themselves within interpretive practices, link literature to personal experiences, and recognize broader social and cultural dimensions. This article examines theoretical foundations, pedagogical strategies, and practical methods for developing reflexive skills in literary education. Drawing on insights from philosophy, psychology, and pedagogy, the study highlights the transformative role of reflexivity in fostering interpretive depth, self-awareness, and critical literacy. By analyzing classroom practices such as reflective writing, dialogic teaching, and collaborative learning, the article proposes effective models for integrating reflexivity into literary curricula. The findings suggest that reflexive pedagogy contributes not only to academic achievement but also to the holistic development of students as independent, critical, and empathetic thinkers.

The transition from school to university mathematics presents academic, emotional and social challenges for many students. In response, Mathematical Thinking Workshops were introduced at the University of Cape Town to support first-year science students in developing reflective, strategic approaches to mathematical learning. The workshops aimed to foster conceptual understanding, metacognitive regulation and a stronger sense of mathematical identity through collaborative problem-solving and structured reflection. This study qualitatively evaluates how students perceive the workshop's impact, which aspects are most beneficial, and what challenges arise. Data were collected through six focus group interviews with 17 workshop participants. Reflexive thematic analysis was used, informed by theories of identity, metacognition, adaptive expertise and sociocultural learning. Five themes emerged: (1) Reflective and Strategic Mathematical Thinking, (2) Communication and Problem-Solving Confidence, (3) Reconnecting with Mathematics Through Alignment, (4) Facilitator Influence and Supportive Learning Environment and (5) Student-identified Hurdles. Students reported increased confidence, deeper conceptual engagement and stronger peer belonging. Benefits included metacognitive growth, collaborative learning and emotional safety, though barriers, such as venue discomfort and time pressures, were noted. The findings suggest that theoretically grounded interventions like these workshops can transform students’ mathematical learning and identity, and should be scaled and embedded into mainstream curricula.

No abstract available

No abstract available

No abstract available

No abstract available

No abstract available

Abstract This classroom-based observational study explored the relationship among young adolescent L2 learners’ self-regulated learning (SRL), SRL behaviours, and L2 use during group work. Although recent research has examined SRL concerning specific L2 skills (e.g., listening) via questionnaires, learners’ actual behaviours during interaction have rarely been investigated. Participants were Grades 7 and 8 (12–14 years old) EFL learners in Chile. First, participants answered an SRL questionnaire. Then, they engaged in a group project (20 groups in total) whose objective was to create a newsletter for the school. Their interactions during the group project were video-recorded (over 300 min in total). A coding scheme for SRL behaviours was developed by focusing on both verbal and nonverbal cues related to SRL. Their L2 use was analysed by focusing on the number of words, turns, and language-related episodes (LREs). The correlational analyses showed that the more SRL behaviours, the more L2 use, implying that when learners engage in SRL behaviours, it is more likely that they would benefit from group work for their L2 learning. Implications for pedagogical interventions for SRL are discussed.

Advancements in conversational artificial intelligence (AI) enable scalable, web-based training environments beyond traditional classroom settings to support learning. In emergency communication centers such as 9-1-1, trainees have limited opportunities to practice diverse call-handling scenarios, as it relies on role-play instruction and constrained instructor availability. We introduce SHIELD (Strengthening Human Intervention in Emergencies through Learning with Data and AI), a framework and online conversational AI system designed to support scenario-based training through interactive simulations, adaptive feedback, and learning analytics. SHIELD integrates AI-generated call simulations with data science techniques to capture fine-grained trainee interaction data, including decision-making behavior, response timing, and corrective actions. These logged interactions are analyzed to provide real-time metacognitive nudges during simulated calls and post-performance feedback through AI-assisted analytics. The framework is informed by principles of self-regulated learning to structure training across planning, execution, and reflection phases. This demo showcases SHIELD as a web-based prototype deployed for emergency call-taker training and discusses design insights from initial testing sessions conducted in collaboration with a public safety communications agency. The system also highlights potential applicability to other high-stakes operational training settings.

Language learning strategies have been of interest to researchers and educators alike for half a century. Recent research seeks to reinterpret and further theorise strategies, questioning whether strategies have to be self-regulated, or can instead be other-regulated. This study draws on data from a naturalistic English Medium Instruction classroom to investigate how learners use listening strategies when comprehending teacher input. Lesson observations and stimulated recall interviews with eight students provided in-depth data on how students’ strategies were prompted by the teacher (e.g. teacher asking students to recall their prior knowledge) during classroom interaction. The findings empirically support state-of-the-art thinking in the field that listening strategies do not have to be self-regulated and can be prompted by the other. Pedagogical implications are discussed in terms of the interactive nature of learners’ strategy use in relation to teacher input, and how teachers can prompt learners’ use of strategies to facilitate uptake.

No abstract available

Abstract Interaction is crucial for online self-regulated learning (OSRL) ability and learning outcomes. The absence of social interaction might lead to high dropout rates in online learning environments. Danmaku holds great potential to enhance online learning by fostering interaction. This study explores the motivations behind university students’ use of danmaku and its influence on their OSRL. Using a mixed methods approach through surveys and interviews with 100 university students from two universities, we found that danmaku promotes social interaction by fulfilling students’ information and entertainment needs. Additionally, engagement with danmaku supports self-regulated learning through reflection and responding strategies and enhances enjoyment by increasing self-efficacy in contributing to the online learning community. This study expands understanding of the role interactive tools like danmaku can play in enhancing social interaction and OSRL, and highlights the potential of danmaku to improve student engagement and reduce dropout rates for quality education.

Recent research underscores the importance of inquiry learning for effective science education. Inquiry learning involves self-regulated learning (SRL), for example when students conduct investigations. Teachers face challenges in orchestrating and tracking student learning in such instruction; making it hard to adequately support students. Using AI methods such as machine learning (ML), the data that is generated when students interact in technology-enhanced classrooms can be used to track their learning and subsequently to inform teachers so that they can better support student learning. This study implemented digital workbooks in an inquiry-based physics unit, collecting cognitive, metacognitive, and affective data from 214 students. Using ML methods, an early warning system was developed to predict students’ learning outcomes. Explainable ML methods were used to unpack these predictions and analyses were conducted for potential biases. Results indicate that an integration of cognitive, metacognitive, and affective data can predict students’ productivity with an accuracy ranging from 60 to 100% as the unit progresses. Initially, affective and metacognitive variables dominate predictions, with cognitive variables becoming more significant later. Using only affective and metacognitive data, predictive accuracies ranged from 60 to 80% throughout. Bias was found to be highly dependent on the ML methods being used. The study highlights the potential of digital student workbooks to support SRL in inquiry-based science education, guiding future research and development to enhance instructional feedback and teacher insights into student engagement. Further, the study sheds new light on the data needed and the methodological challenges when using ML methods to investigate SRL processes in classrooms.

The paper describes a theoretical framework for the study of teachers’ promotion of self-regulated learning in the classroom. The Self-Regulated Learning Teacher Promotion Framework (SRL-TPF) utilizes the ICAP theory to assess the affordances of the learning environment for the indirect promotion of SRL, proposes new variables in the investigation of the direct promotion of SRL, and examines how these two ways to promote SRL are related. The SRL-TPF was used to examine the direct and indirect promotion of SRL in filmed observations of 23 Australian classrooms. The results revealed a paucity in the design of Constructive and Interactive lesson tasks that support the indirect promotion of SRL and a preference for the direct support of SRL through implicit strategy instruction and the provision of metacognitive reflection and support. There were important teacher differences in both the direct and indirect promotion of SRL, but the teachers who were more likely to design Constructive and Interactive lesson tasks did not necessarily promote SRL directly and vice versa. The research contributes to a better understanding of the relationship between teaching what to learn (subject content) and how to learn (SRL knowledge and strategies).

This research investigated the details and effects of a short online Professional Learning Program designed to develop teacher education students’ knowledge about how to promote self-regulated learning (SRL) in the classroom. The Program was based on a new framework for how teachers can promote SRL, the SRL Teacher Promotion Framework (SRL-TPF), which focused on the promotion of SRL strategies, students’ knowledge about learning, and students’ metacognition. It consisted of seven modules describing the different SRL promotion types and SRL capabilities and ways to promote them through teacher talk and action. Modules included written information and video examples taken from observations of real classrooms, which were used to illustrate the transfer of SRL theory to instructional practice. Each module concluded with several assessment items. During the Program the participants, 91 teacher education students, were asked to use a simplified scoring system based on the SRL-TPF to code lesson transcripts taken from classroom observations. The results showed that by the end of the program over 85% of the participants were able to provide teacher instructions that included explicit SRL promotion and/or promoted students’ SRL knowledge. Our study contributes to research findings on teacher education students’ knowledge of SRL, their promotion of SRL to students, and the contribution of short duration SRL professional development.

Metacognition is an important aspect in creative problem solving (CPS) and through this chapter we analyse the meta-reasoning aspects applied in the different processes of monitoring the progress of learners' reasoning and CPS activities. Meta-reasoning monitors the way that problem-solving processes advance and regulate time and efforts towards a solution. In the context of an ill-defined problem, exploration is required to develop a better-defined problem space and advance towards the solution space. The way learners engage in exploration and exploitations is regulated by the meta-reasoning within the CPS activity. The objective of this chapter is to examine and identify the CPS process with educational robots through a metacognitive and interactionist approach. This chapter presents a case study, where, to solve a problem, a participant had to explore a set of robot cubes to develop the technological knowledge associated with each single component of the system, but also conceptualize a system-level behaviour of the cubes when they are assembled. The chapter presents the emergence of knowledge through the metacognitive regulation of the process of exploration and exploitation of prior knowledge and emergent knowledge until finding a solution

Previous data-driven work investigating the types and distributions of discourse relation signals, including discourse markers such as 'however' or phrases such as 'as a result' has focused on the relative frequencies of signal words within and outside text from each discourse relation. Such approaches do not allow us to quantify the signaling strength of individual instances of a signal on a scale (e.g. more or less discourse-relevant instances of 'and'), to assess the distribution of ambiguity for signals, or to identify words that hinder discourse relation identification in context ('anti-signals' or 'distractors'). In this paper we present a data-driven approach to signal detection using a distantly supervised neural network and develop a metric, Delta s (or 'delta-softmax'), to quantify signaling strength. Ranging between -1 and 1 and relying on recent advances in contextualized words embeddings, the metric represents each word's positive or negative contribution to the identifiability of a relation in specific instances in context. Based on an English corpus annotated for discourse relations using Rhetorical Structure Theory and signal type annotations anchored to specific tokens, our analysis examines the reliability of the metric, the places where it overlaps with and differs from human judgments, and the implications for identifying features that neural models may need in order to perform better on automatic discourse relation classification.

With the increasing use of generative Artificial Intelligence (AI) methods to support science workflows, we are interested in the use of discourse-level information to find supporting evidence for AI generated scientific claims. A first step towards this objective is to examine the task of inferring discourse structure in scientific writing. In this work, we present a preliminary investigation of pretrained language model (PLM) and Large Language Model (LLM) approaches for Discourse Relation Classification (DRC), focusing on scientific publications, an under-studied genre for this task. We examine how context can help with the DRC task, with our experiments showing that context, as defined by discourse structure, is generally helpful. We also present an analysis of which scientific discourse relation types might benefit most from context.

We introduce an attention-based Bi-LSTM for Chinese implicit discourse relations and demonstrate that modeling argument pairs as a joint sequence can outperform word order-agnostic approaches. Our model benefits from a partial sampling scheme and is conceptually simple, yet achieves state-of-the-art performance on the Chinese Discourse Treebank. We also visualize its attention activity to illustrate the model's ability to selectively focus on the relevant parts of an input sequence.

This paper aims to quantitatively evaluate the performance of ChatGPT, an interactive large language model, on inter-sentential relations such as temporal relations, causal relations, and discourse relations. Given ChatGPT's promising performance across various tasks, we proceed to carry out thorough evaluations on the whole test sets of 11 datasets, including temporal and causal relations, PDTB2.0-based, and dialogue-based discourse relations. To ensure the reliability of our findings, we employ three tailored prompt templates for each task, including the zero-shot prompt template, zero-shot prompt engineering (PE) template, and in-context learning (ICL) prompt template, to establish the initial baseline scores for all popular sentence-pair relation classification tasks for the first time. Through our study, we discover that ChatGPT exhibits exceptional proficiency in detecting and reasoning about causal relations, albeit it may not possess the same level of expertise in identifying the temporal order between two events. While it is capable of identifying the majority of discourse relations with existing explicit discourse connectives, the implicit discourse relation remains a formidable challenge. Concurrently, ChatGPT demonstrates subpar performance in the dialogue discourse parsing task that requires structural understanding in a dialogue before being aware of the discourse relation.

Recent research on discourse relations has found that they are cued not only by discourse markers (DMs) but also by other textual signals and that signaling information is indicative of genres. While several corpora exist with discourse relation signaling information such as the Penn Discourse Treebank (PDTB, Prasad et al. 2008) and the Rhetorical Structure Theory Signalling Corpus (RST-SC, Das and Taboada 2018), they both annotate the Wall Street Journal (WSJ) section of the Penn Treebank (PTB, Marcus et al. 1993), which is limited to the news domain. Thus, this paper adapts the signal identification and anchoring scheme (Liu and Zeldes, 2019) to three more genres, examines the distribution of signaling devices across relations and genres, and provides a taxonomy of indicative signals found in this dataset.

In this paper, we present a method for identifying discourse marker usage in spontaneous speech based on machine learning. Discourse markers are denoted by special POS tags, and thus the process of POS tagging can be used to identify discourse markers. By incorporating POS tagging into language modeling, discourse markers can be identified during speech recognition, in which the timeliness of the information can be used to help predict the following words. We contrast this approach with an alternative machine learning approach proposed by Litman (1996). This paper also argues that discourse markers can be used to help the hearer predict the role that the upcoming utterance plays in the dialog. Thus discourse markers should provide valuable evidence for automatic dialog act prediction.

This paper investigates the influence of discourse features on text complexity assessment. To do so, we created two data sets based on the Penn Discourse Treebank and the Simple English Wikipedia corpora and compared the influence of coherence, cohesion, surface, lexical and syntactic features to assess text complexity. Results show that with both data sets coherence features are more correlated to text complexity than the other types of features. In addition, feature selection revealed that with both data sets the top most discriminating feature is a coherence feature.

Centering was formulated as a model of the relationship between attentional state, the form of referring expressions, and the coherence of an utterance within a discourse segment (Grosz, Joshi and Weinstein, 1986; Grosz, Joshi and Weinstein, 1995). In this chapter, I argue that the restriction of centering to operating within a discourse segment should be abandoned in order to integrate centering with a model of global discourse structure. The within-segment restriction causes three problems. The first problem is that centers are often continued over discourse segment boundaries with pronominal referring expressions whose form is identical to those that occur within a discourse segment. The second problem is that recent work has shown that listeners perceive segment boundaries at various levels of granularity. If centering models a universal processing phenomenon, it is implausible that each listener is using a different centering algorithm.The third issue is that even for utterances within a discourse segment, there are strong contrasts between utterances whose adjacent utterance within a segment is hierarchically recent and those whose adjacent utterance within a segment is linearly recent. This chapter argues that these problems can be eliminated by replacing Grosz and Sidner's stack model of attentional state with an alternate model, the cache model. I show how the cache model is easily integrated with the centering algorithm, and provide several types of data from naturally occurring discourses that support the proposed integrated model. Future work should provide additional support for these claims with an examination of a larger corpus of naturally occurring discourses.

The goal of argumentation mining, an evolving research field in computational linguistics, is to design methods capable of analyzing people's argumentation. In this article, we go beyond the state of the art in several ways. (i) We deal with actual Web data and take up the challenges given by the variety of registers, multiple domains, and unrestricted noisy user-generated Web discourse. (ii) We bridge the gap between normative argumentation theories and argumentation phenomena encountered in actual data by adapting an argumentation model tested in an extensive annotation study. (iii) We create a new gold standard corpus (90k tokens in 340 documents) and experiment with several machine learning methods to identify argument components. We offer the data, source codes, and annotation guidelines to the community under free licenses. Our findings show that argumentation mining in user-generated Web discourse is a feasible but challenging task.

This paper describes our submission to the DISRPT2021 Shared Task on Discourse Unit Segmentation, Connective Detection, and Relation Classification. Our system, called DisCoDisCo, is a Transformer-based neural classifier which enhances contextualized word embeddings (CWEs) with hand-crafted features, relying on tokenwise sequence tagging for discourse segmentation and connective detection, and a feature-rich, encoder-less sentence pair classifier for relation classification. Our results for the first two tasks outperform SOTA scores from the previous 2019 shared task, and results on relation classification suggest strong performance on the new 2021 benchmark. Ablation tests show that including features beyond CWEs are helpful for both tasks, and a partial evaluation of multiple pre-trained Transformer-based language models indicates that models pre-trained on the Next Sentence Prediction (NSP) task are optimal for relation classification.

In languages such as Japanese, the use of {\it zeros}, unexpressed arguments of the verb, in utterances that shift the topic involves a risk that the meaning intended by the speaker may not be transparent to the hearer. However, this potentially undesirable conversational strategy often occurs in the course of naturally-occurring discourse. In this chapter, I report on an empirical study of 250 utterances with {\it zeros} in 20 Japanese newspaper articles. Each utterance is analyzed in terms of centering transitions and the form in which centers are realized by referring expressions. I also examine lexical subcategorization information, and tense and aspect in order to test the hypothesis that the speaker expects the hearer to use this information in determining global discourse structure. I explain the occurrence of {\it zeros} in {\sc retain} and {\sc rough-shift} centering transitions, by claiming that a {\it zero} can only be used in these cases when the shift of centers is supported by contextual information such as lexical semantics, tense and aspect, and agreement features. I then propose an algorithm by which centering can incorporate these observations to integrate centering with global discourse structure, and thus enhance its ability for non-local pronoun resolution.

This paper presents DeDisCo, Georgetown University's entry in the DISRPT 2025 shared task on discourse relation classification. We test two approaches, using an mt5-based encoder and a decoder based approach using the openly available Qwen model. We also experiment on training with augmented dataset for low-resource languages using matched data translated automatically from English, as well as using some additional linguistic features inspired by entries in previous editions of the Shared Task. Our system achieves a macro-accuracy score of 71.28, and we provide some interpretation and error analysis for our results.

In this paper, we introduce Dependency Dialogue Acts (DDA), a novel framework for capturing the structure of speaker-intentions in multi-party dialogues. DDA combines and adapts features from existing dialogue annotation frameworks, and emphasizes the multi-relational response structure of dialogues in addition to the dialogue acts and rhetorical relations. It represents the functional, discourse, and response structure in multi-party multi-threaded conversations. A few key features distinguish DDA from existing dialogue annotation frameworks such as SWBD-DAMSL and the ISO 24617-2 standard. First, DDA prioritizes the relational structure of the dialogue units and the dialog context, annotating both dialog acts and rhetorical relations as response relations to particular utterances. Second, DDA embraces overloading in dialogues, encouraging annotators to specify multiple response relations and dialog acts for each dialog unit. Lastly, DDA places an emphasis on adequately capturing how a speaker is using the full dialog context to plan and organize their speech. With these features, DDA is highly expressive and recall-oriented with regard to conversation dynamics between multiple speakers. In what follows, we present the DDA annotation framework and case studies annotating DDA structures in multi-party, multi-threaded conversations.

In deductive domains, three metacognitive knowledge types in ascending order are declarative, procedural, and conditional learning. This work leverages Deep Reinforcement Learning (DRL) in providing adaptive metacognitive interventions to bridge the gap between the three knowledge types and prepare students for future learning across Intelligent Tutoring Systems (ITSs). Students received these interventions that taught how and when to use a backward-chaining (BC) strategy on a logic tutor that supports a default forward-chaining strategy. Six weeks later, we trained students on a probability tutor that only supports BC without interventions. Our results show that on both ITSs, DRL bridged the metacognitive knowledge gap between students and significantly improved their learning performance over their control peers. Furthermore, the DRL policy adapted to the metacognitive development on the logic tutor across declarative, procedural, and conditional students, causing their strategic decisions to be more autonomous.

Developing students' ability to troubleshoot is an important learning outcome for many undergraduate physics lab courses, especially electronics courses. In other work, metacognition has been identified as an important feature of troubleshooting. However, that work has focused primarily on individual students' metacognitive processes or troubleshooting abilities. In contrast, electronics courses often require students to work in pairs, and hence students' in-class experiences likely have significant social dimensions that are not well understood. In this work, we use an existing framework for socially mediated metacognition to analyze audiovisual data from think-aloud activities in which eight pairs of students from two institutions attempted to diagnose and repair a malfunctioning electric circuit. In doing so, we provide insight into some of the social metacognitive dynamics that arise during collaborative troubleshooting. We find that students engaged in socially mediated metacognition at multiple key transitions during the troubleshooting process. Reciprocated metacognitive dialogue arose when students were collectively strategizing about which measurements to perform, or reaching a shared understanding of the circuit's behavior. In addition to elaborating upon these findings, we discuss implications for instruction, and we identify areas for potential future investigation.

Contributing to the literature on aptitude-treatment interactions between worked examples and problem-solving, this paper addresses differential learning from the two approaches when students are positioned as domain experts learning new concepts. Our evaluation is situated in a team project that is part of an advanced software engineering course. In this course, students who possess foundational domain knowledge but are learning new concepts engage alternatively in programming followed by worked example-based reflection. They are either allowed to finish programming or are curtailed after a pre-specified time to participate in a longer worked example-based reflection. We find significant pre- to post-test learning gains in both conditions. Then, we not only find significantly more learning when students participated in longer worked example-based reflections but also a significant performance improvement on a problem-solving transfer task. These findings suggest that domain experts learning new concepts benefit more from worked example-based reflections than from problem-solving.

Many language learners need to be supported in acquiring a second or foreign language quickly and effectively across learning environments beyond the classroom. The chapter argues that support should focus on the development of two vital learning skills, namely being able to self-regulate and to collaborate effectively in the learning process. We base our argumentation on the theoretical lenses of self-regulated learning (SRL) and collaborative learning in the context of mobile situated learning that can take place in a variety of settings. The chapter examines a sample of selected empirical studies within the field of mobile-assisted language learning with a twofold aim. Firstly, the studies are analyzed in order to understand the role of learner self-regulation and collaboration while acquiring a new language beyond the classroom. Secondly, we aim to provide a deeper understanding of any mechanisms provided to develop or support language learners' self-regulated and collaborative learning skills. Finally, we propose that fostering SRL and collaborative learning skills and strategies will benefit from recent advances in the fields of learning analytics and artificial intelligence, coupled with the use of mobile technologies and self-monitoring mechanisms. The ultimate aim is to enable the provision of individual adaptive learning paths to facilitate language learning beyond the classroom.

While generative artificial intelligence (Gen AI) increasingly transforms academic environments, a critical gap exists in understanding and mitigating human biases in AI interactions, such as anchoring and confirmation bias. This position paper advocates for metacognitive AI literacy interventions to help university students critically engage with AI and address biases across the Human-AI interaction workflows. The paper presents the importance of considering (1) metacognitive support with deliberate friction focusing on human bias; (2) bi-directional Human-AI interaction intervention addressing both input formulation and output interpretation; and (3) adaptive scaffolding that responds to diverse user engagement patterns. These frameworks are illustrated through ongoing work on "DeBiasMe," AIED (AI in Education) interventions designed to enhance awareness of cognitive biases while empowering user agency in AI interactions. The paper invites multiple stakeholders to engage in discussions on design and evaluation methods for scaffolding mechanisms, bias visualization, and analysis frameworks. This position contributes to the emerging field of AI-augmented learning by emphasizing the critical role of metacognition in helping students navigate the complex interaction between human, statistical, and systemic biases in AI use while highlighting how cognitive adaptation to AI systems must be explicitly integrated into comprehensive AI literacy frameworks.

Reciprocal questioning is essential for effective teaching and learning, fostering active engagement and deeper understanding through collaborative interactions, especially in large classrooms. Can large language model (LLM), such as OpenAI's GPT (Generative Pre-trained Transformer) series, assist in this? This paper investigates a pedagogical approach of classroom flipping based on flipped interaction in LLMs. Flipped interaction involves using language models to prioritize generating questions instead of answers to prompts. We demonstrate how traditional classroom flipping techniques, including Peer Instruction and Just-in-Time Teaching (JiTT), can be enhanced through flipped interaction techniques, creating student-centric questions for hybrid teaching. In particular, we propose a workflow to integrate prompt engineering with clicker and JiTT quizzes by a poll-prompt-quiz routine and a quiz-prompt-discuss routine to empower students to self-regulate their learning capacity and enable teachers to swiftly personalize training pathways. We develop an LLM-driven chatbot software that digitizes various elements of classroom flipping and facilitates the assessment of students using these routines to deliver peer-generated questions. We have applied our LLM-driven chatbot software for teaching both undergraduate and graduate students from 2020 to 2022, effectively useful for bridging the gap between teachers and students in remote teaching during the COVID-19 pandemic years. In particular, LLM-driven classroom flipping can be particularly beneficial in large class settings to optimize teaching pace and enable engaging classroom experiences.

Content assessment has broadly improved in e-learning scenarios in recent decades. However, the eLearning process can give rise to a spatial and temporal gap that poses interesting challenges for assessment of not only content, but also students' acquisition of core skills such as self-regulated learning. Our objective was to discover students' self-regulated learning processes during an eLearning course by using Process Mining Techniques. We applied a new algorithm in the educational domain called Inductive Miner over the interaction traces from 101 university students in a course given over one semester on the Moodle 2.0 platform. Data was extracted from the platform's event logs with 21629 traces in order to discover students' self-regulation models that contribute to improving the instructional process. The Inductive Miner algorithm discovered optimal models in terms of fitness for both Pass and Fail students in this dataset, as well as models at a certain level of granularity that can be interpreted in educational terms, which are the most important achievement in model discovery. We can conclude that although students who passed did not follow the instructors' suggestions exactly, they did follow the logic of a successful self-regulated learning process as opposed to their failing classmates. The Process Mining models also allow us to examine which specific actions the students performed, and it was particularly interesting to see a high presence of actions related to forum-supported collaborative learning in the Pass group and an absence of those in the Fail group.