Interactive Development of Adaptive Learning and Memory

Across development, interactions between value-based learning and memory processes promote the formation of mental models that enable flexible goal pursuit. Value cues in the environment signal information that may be useful to prioritize in memory; these prioritized memories in turn form the foundation of structured knowledge representations that guide subsequent learning. Critically, neural and cognitive component processes of learning and memory undergo marked shifts from infancy to adulthood, leading to developmental change in the construction of mental models and how they are used to guide goal-directed behavior. This review explores how changes in reciprocal interactions between value-based learning and memory influence adaptive behavior across development and highlights avenues for future research.

Scaling Up Blended Learning on Innovation and Entrepreneurship: a Proposal to Increase Value in a Semi Presential Module

Education on Innovation and Entrepreneurship (I&E) has increased in the last two decades, specially, through MOOCs. Lately, these reusable online alternatives have tended to be revalorized by HEIs into blended learning activities, posing new challenges for instructors, specially, on how to bridge prior knowledge with in-class activities. Adopting a discursive approach to knowledge, our proposal aims to meet this challenge by identifying student’s ‘representations’, i.e., patterned constructions on disciplinary knowledge. Representations can be found across different cohorts and thus further complemented by instructors. To test this assumption and build our proposal, we analysed student’s representations in two observations. We mapped students’ representations over key I&E definitions (e.g., ‘start-up’) and, to know how prior knowledge may be complemented by instructors, we identified students’ alignment with expert disciplinary knowledge. Firstly, we found that the two cohorts tended to express representations by turning attention to several dimensions, e.g., referring to different types of features or finalities associated with concepts. Secondly, the disciplinary alignment description revealed that students tended to focus on the same components present in experts’ definitions, but with a greater level of generality. Our results have been packaged into a proposal that aims to help instructors scale their blended activities.

Identification of Bacteriophages Using Deep Representation Model with Pre-training

Bacteriophages/Phages are viruses that infect and replicate within bacteria and archaea. Antibiotic resistance is one of the biggest threats to global health. The therapeutic use of bacteriophages provides another potential solution for solving antibiotic resistance. To develop phage therapies, the identification of phages from metagenome sequences is the fundamental step. Currently, several methods have been developed for identifying phages. These methods can be categorized into two types: database-based methods and alignment-free methods. The database-based approach, such as VIBRANT, utilizes existing databases and compares sequence similarity between candidates and those in the databases. The alignment-free method, such as Seeker and DeepVirFinder, uses deep learning models to directly predict phages based on nucleotide sequences. Both approaches have their advantages and disadvantages. In this work, we propose using a deep representation learning model with pre-training to integrate the database-based and non-alignment-based methods (we call it INHERIT). The pre-training is used as an alternative way for acquiring knowledge representations from existing databases, while the BERT-style deep learning framework retains the advantage of alignment-free methods. We compared the proposed method with VIBRANT and Seeker on a third-party benchmark dataset. Our experiments show that INHERIT achieves better performance than the database-based approach and the alignment-free method, with the best F1-score of 0.9868. Meanwhile, we demonstrated that using pre-trained models helps to improve the non-alignment deep learning model further.

The language network supports both lexical access and sentence generation during language production

A network of left frontal and temporal brain regions has long been implicated in language comprehension and production. However, because of relatively fewer investigations of language production, the precise role of this ‘language network’ in production-related cognitive processes remains debated. Across four fMRI experiments that use picture naming/description to mimic the translation of conceptual representations into words and sentences, we characterize the response of the language regions to production demands. In line with prior studies, sentence production elicited strong responses throughout the language network. Further, we report three novel results. First, we demonstrate that production-related responses in the language network are robust to output modality (speaking vs. typing). Second, the language regions respond to both lexical access and sentence-generation demands. This pattern implies strong integration between lexico-semantic and combinatorial processes, mirroring the picture that has emerged in language comprehension. Finally, some have previously hypothesized the existence of production-selective mechanisms given that syntactic encoding is a critical part of sentence production, whereas comprehension is possible even when syntactic cues are degraded or absent. Contrary to this hypothesis, we find no evidence of brain regions that selectively support sentence generation. Instead, language regions respond overall more strongly during production than during comprehension, which suggests that production incurs a greater cost for the language network. Together, these results align with the idea that language comprehension and production draw on the same knowledge representations, which are stored in the language-selective network and are used both to interpret linguistic input and generate linguistic output.

Comparison of Different Knowledge Representations for Complex Structured Objects in Solving AI Problems

The problem of knowledge representation for a complex structured object is one of the actual problems of AI. This is due to the fact that many of the objects under study are not a single indivisible object characterized by its properties, but complex structures whose elements have some known properties and are in some, often multiplace, relations with each other. An approach to the representation of such knowledge based on first-order logic (predicate calculus formulas) is compared in this paper with two currently widespread approaches based on the representation of data information with the use of finite-valued strings or graphs. It is shown that the use of predicate calculus formulas for description of a complex structured object, despite the NP-difficulty of the solved problems arising after formalization, actually have no greater computational complexity than the other two approaches, what is usually not mentioned by their supporters. An algorithm for constructing an ontology is proposed that does not depend on the methodof desc ribing an object, and is based on the selection of the maximum common property of objects from a given set.

A Comprehensive Survey of Cognitive Graphs: Techniques, Applications, Challenges

The realization of the third-generation artificial intelligence (AI) requires the evolution from perceptual intelligence to cognitive intelligence, where knowledge graphs may not meet the practical needs anymore. Based on the dual channel theory, cognitive graphs are established and developed through coordinating the implicit extraction module and the explicit reasoning module as well as integrating knowledge graphs, cognitive reasoning and logical expressions, which have achieved successes in multi-hop question answering. It is desired for cognitive graphs to be widely used in advanced AI applications such as large-scale knowledge representations and intelligent responses, promoting the development of Al dramatically. This review discusses cognitive graphs systematically and elaborately, including basic concepts, generations, theories and technologies. Moreover, we try to predict the development of cognitive intelligence in the short-term future and further enlighten more researches and studies.

Knowledge Representation and Management: Interest in New Solutions for Ontology Curation

Summary Objective: To select, present and summarize some of the best papers in the field of Knowledge Representation and Management (KRM) published in 2020. Methods: A comprehensive and standardized review of the medical informatics literature was performed to select the most interesting papers of KRM published in 2020, based on PubMed queries. This review was conducted according to the IMIA Yearbook guidelines. Results: Four best papers were selected among 1,175 publications. In contrast with the papers selected last year, the four best papers of 2020 demonstrated a significant focus on methods and tools for ontology curation and design. The usual KRM application domains (bioinformatics, machine learning, and electronic health records) were also represented. Conclusion: In 2020, ontology curation emerges as a significant topic of research interest. Bioinformatics, machine learning, and electronics health records remain significant research areas in the KRM community with various applications. Knowledge representations are key to advance machine learning by providing context and to develop novel bioinformatics metrics. As in 2019, representations serve a great variety of applications across many medical domains, with actionable results and now with growing adhesion to the open science initiative.