Exploring the relations between teachers’ cue-utilization, monitoring and regulation of students’ text learning

This study investigated teachers’ monitoring and regulation of students’ learning from texts. According to the cue-utilization framework (Koriat, in Journal of Experimental Psychology, 126, 349–370, 1997), monitoring accuracy depends on how predictive the information (or cues) that teachers use to make monitoring judgments actually is for students’ performance. Accurate monitoring of students’ comprehension is considered a precondition for adaptive regulation of students’ learning. However, these assumptions have not yet been directly investigated. We therefore examined teachers’ cue-utilization and how it affects their monitoring and regulation accuracy. In a within-subjects design, 21 secondary education teachers made monitoring judgments and regulation decisions for fifteen students under three cue-availability conditions: 1) only student cues (i.e., student’s name), 2) only performance cues (i.e., diagrams students completed about texts they had read), and 3) both student and performance cues (i.e., student’s name and completed diagram). Teachers’ absolute and relative monitoring accuracy was higher when having student cues available in addition to diagram cues. Teachers’ relative regulation accuracy was higher when having only performance cues available instead of only student cues (as indicated by a direct effect). Monitoring accuracy predicted regulation accuracy and in addition to a direct effect, we also found and indirect effect of cue-availability on regulation accuracy (via monitoring accuracy). These results suggest that accurate regulation can be brought about both indirectly by having accurate monitoring judgments and directly by cue-utilization. The findings of this study can help to refine models of teacher monitoring and regulation and can be useful in designing effective interventions to promote teachers’ monitoring and regulation.

autoBOT: evolving neuro-symbolic representations for explainable low resource text classification

Learning from texts has been widely adopted throughout industry and science. While state-of-the-art neural language models have shown very promising results for text classification, they are expensive to (pre-)train, require large amounts of data and tuning of hundreds of millions or more parameters. This paper explores how automatically evolved text representations can serve as a basis for explainable, low-resource branch of models with competitive performance that are subject to automated hyperparameter tuning. We present autoBOT (automatic Bags-Of-Tokens), an autoML approach suitable for low resource learning scenarios, where both the hardware and the amount of data required for training are limited. The proposed approach consists of an evolutionary algorithm that jointly optimizes various sparse representations of a given text (including word, subword, POS tag, keyword-based, knowledge graph-based and relational features) and two types of document embeddings (non-sparse representations). The key idea of autoBOT is that, instead of evolving at the learner level, evolution is conducted at the representation level. The proposed method offers competitive classification performance on fourteen real-world classification tasks when compared against a competitive autoML approach that evolves ensemble models, as well as state-of-the-art neural language models such as BERT and RoBERTa. Moreover, the approach is explainable, as the importance of the parts of the input space is part of the final solution yielded by the proposed optimization procedure, offering potential for meta-transfer learning.

A modular framework for ontology learning from text in Portuguese

Research on ontology learning has been carried out in many knowledge areas, especially in Artificial Intelligence. Semi-automatic or automatic ontology learning can contribute to the field of knowledge representation. Many semi-automatic approaches to ontology learning from texts have been proposed. Most of these proposals use natural language processing techniques. This paper describes a computational framework construction for semi-automated ontology learning from texts in Portuguese. Axioms are not treated in this paper. The work described here originated from the Philipp Cimiano’s proposal along with text standardization mechanisms, natural language processing, identification of taxonomic relations and techniques for structuring ontologies. In this work, a case study on public security domain was also done, showing the benefits of the developed computational framework. The result of this case study is an ontology for this area.

To What Extent Do Situation-Model-Approach Interventions Improve Relative Metacomprehension Accuracy? Meta-Analytic Insights

This meta-analysis investigated the extent to which relative metacomprehension accuracy can be increased by interventions that aim to support learners’ use of situation-model cues as a basis for judging their text comprehension. These interventions were delayed-summary writing, delayed-keywords listing, delayed-diagram completion, self-explaining, concept mapping, rereading, and setting a comprehension-test expectancy. First, the general effectiveness of situation-model-approach interventions was examined. The results revealed that, across 28 effect sizes (comprising a total of 2,236 participants), situation-model-approach interventions exerted a medium positive effect (g = 0.46) on relative metacomprehension accuracy. Second, the interventions were examined individually. The results showed that, with the exception of self-explaining, each intervention had a significant positive effect on relative metacomprehension accuracy. Yet, there was a tendency for setting a comprehension-test expectancy to be particularly effective. A further meta-analysis on comprehension in the selected studies revealed that, overall, the situation-model-approach interventions were also beneficial for directly improving comprehension, albeit the effect was small. Taken together, the findings demonstrate the utility of situation-model-approach interventions for supporting self-regulated learning from texts.

Mapping and Drawing to Improve Students’ and Teachers’ Monitoring and Regulation of Students’ Learning from Text: Current Findings and Future Directions

For (facilitating) effective learning from texts, students and teachers need to accurately monitor students’ comprehension. Monitoring judgments are accurate when they correspond to students’ actual comprehension. Accurate monitoring enables accurate (self-)regulation of the learning process, i.e., making study decisions that are in line with monitoring judgments and/or students’ comprehension. Yet, (self-)monitoring accuracy is often poor as the information or cues used are not always diagnostic (i.e., predictive) for students’ actual comprehension. Having students engage in generative activities making diagnostic cues available improves monitoring and regulation accuracy. In this review, we focus on generative activities in which text is transformed into visual representations using mapping and drawing (i.e., making diagrams, concept maps, or drawings). This has been shown to improve monitoring and regulation accuracy and is suited for studying cue diagnosticity and cue utilization. First, we review and synthesize findings of studies regarding (1) students’ monitoring accuracy, regulation accuracy, learning, cue diagnosticity, and cue utilization; (2) teachers’ monitoring and regulation accuracy and cue utilization; and (3) how mapping and drawing affect using effort as a cue during monitoring and regulation, and how this affects monitoring and regulation accuracy. Then, we show how this research offers unique opportunities for future research on advancing measurements of cue diagnosticity and cue utilization and on how effort is used as a cue during monitoring and regulation. Improving measures of cue diagnosticity and cue utilization can provide us with more insight into how students and teachers monitor and regulate students’ learning, to help design effective interventions to foster these important skills.