Knowledge Of World Research Articles

Graph Reasoning Transformers for Knowledge-Aware Question Answering

Augmenting Language Models (LMs) with structured knowledge graphs (KGs) aims to leverage structured world knowledge to enhance the capability of LMs to complete knowledge-intensive tasks. However, existing methods are unable to effectively utilize the structured knowledge in a KG due to their inability to capture the rich relational semantics of knowledge triplets. Moreover, the modality gap between natural language text and KGs has become a challenging obstacle when aligning and fusing cross-modal information. To address these challenges, we propose a novel knowledge-augmented question answering (QA) model, namely, Graph Reasoning Transformers (GRT). Different from conventional node-level methods, the GRT serves knowledge triplets as atomic knowledge and utilize a triplet-level graph encoder to capture triplet-level graph features. Furthermore, to alleviate the negative effect of the modality gap on joint reasoning, we propose a representation alignment pretraining to align the cross-modal representations and introduce a cross-modal information fusion module with attention bias to enable fine-grained information fusion. Extensive experiments conducted on three knowledge-intensive QA benchmarks show that the GRT outperforms the state-of-the-art KG-augmented QA systems, demonstrating the effectiveness and adaptation of our proposed model.

Visual Chain-of-Thought Prompting for Knowledge-Based Visual Reasoning

Knowledge-based visual reasoning remains a daunting task since it not only requires machines to interpret the concepts and relationships from visual scenes but also associate them with external world knowledge to conduct a chain of reasoning on open-world questions. Previous works, however, treat visual perception and language-based reasoning as two independent modules, failing to attend to both modules throughout all stages of reasoning. To this end, we propose Visual Chain-of-thought Prompting (VCTP) for knowledge-based reasoning, which involves the interaction between visual content and natural language in an iterative step-by-step reasoning manner. VCTP contains three stages, see, think, and confirm. The see stage scans the image and grounds the visual concept candidates with a visual perception model. The think stage adopts a pre-trained large language model (LLM) to attend to key visual concepts from natural language questions adaptively. It then transforms key visual context into text context for prompting with a visual captioning model, and adopts the LLM to generate the answer. The confirm stage further uses the LLM to generate the supporting rationale to the answer, which is then passed through a cross-modality classifier to verify that it’s consistent with the visual context. We iterate through the think-confirm stages to ensure the verified rationale is consistent with the answer. We conduct experiments on a range of knowledge-based visual reasoning datasets. We found our VCTP enjoys several benefits, 1). it achieves better performance than the previous few-shot learning baselines; 2). it enjoys the total transparency and trustworthiness of the whole reasoning process by providing rationales for each reasoning step; 3). it is computation-efficient compared with other fine-tuning baselines. Our code is available at https://github.com/UMass-Foundation-Model/VisualCoT.git

A Comparative Study of Relation Classification Approaches for Japanese Discourse Relation Analysis

This paper presents a comparative analysis of classification approaches in the Japanese discourse relation analysis (DRA) task. In the Japanese DRA task, it is difficult to resolve implicit relations where explicit discourse phrases do not appear. To understand implicit relations further, we compared the four approaches by incorporating a special token to encode the relations of the given discourses. Our four approaches included inserting a special token at the beginning of a sentence, end of a sentence, conjunctive position, and random position to classify the relation between the two discourses into one of the following categories: CAUSE/REASON, CONCESSION, CONDITION, PURPOSE, GROUND, CONTRAST, and NONE. Our experimental results revealed that special tokens are available to encode the relations of given discourses more effectively than pooling-based approaches. In particular, the random insertion of a special token outperforms other approaches, including pooling-based approaches, in the most numerous CAUSE/REASON category in implicit relations and categories with few instances. Moreover, we classified the errors in the relation analysis into three categories: confounded phrases, ambiguous relations, and requiring world knowledge for further improvements.

Konsep Ilmu dalam Perspektif Islam

The study of knowledge in Islam is very important. Knowledge in modern times is considered lacking in explaining phenomena in this era. The purpose of writing this article is to analyze the Islamic perspective on science, including the concepts, classifications, and roots of science. This research method is qualitative with a literature study approach originating from e-journals, books and trusted media. The results of this study state that first, the concept of science according to Islam does not only discuss physical aspects, but also metaphysical aspects. Science must also have objectivity, meaning that it can be accepted by many people and not be influenced by individual thoughts. Second, Islam broadly classifies knowledge into two, namely fardu ain knowledge and fardu kifayah knowledge. Figures of Muslim scientists such as Al-Farabi and Ibn Nadhim divided world knowledge into more detail. Third, Knowledge in Islam comes from the Al-Qur'an, Hadith, intellect, and intuition. How to obtain this knowledge can be through revelation, five senses, reason, and heart or intuition. Fourth, Modern science in the West began to develop, and Western society relinquished the influence of the church to adopt a modern way of thinking that emphasized rationality and empirical evidence. Western thought considers religion and science to be separate, viewing this division as a hindrance to progress. However, in the Islamic tradition, religious and secular powers are not divided

Children and adults privilege linguistic over visual information when creating comparison classes for prenominal gradable adjectives

Mastering the semantics of gradable adjectives is a complex acquisition task, since their interpretation requires the calculation of a threshold relative to a comparison class. Context plays a role in determining the comparison class in multiple ways; empirical research has focused on non-linguistic cues, such as world knowledge and visual context, and less on linguistic cues, such as the modified head noun. The present study is the first to examine how preschoolers and adults use the taxonomic category encoded by the modified head noun as well as visual cues to interpret relative (big, small) and absolute (clean, dirty) gradable adjectives. Using existing objects from different taxonomic categories (basic-level and superordinate-level), we developed a picture-choice task that allowed us to infer the comparison class and the threshold from participants’ choices. Forty-three children (aged 3 to 5) and 26 adults, all monolingual German-speaking, were tested. Confirming previous research, children, like adults, showed a context-sensitive interpretation of relative but not of absolute gradable adjectives. The taxonomic label was preferred to visual cues in creating the comparison class already by age three; this advantage of linguistic over non-linguistic information was also present in the adults. We explain this finding with the privileged role of information provided by the taxonomic label: the modified head noun encodes the cognitive representation of an object category and the corresponding taxonomic hierarchy. Accordingly, this representation is easily accessible while creating the comparison class, which provides further evidence that language is a powerful tool to modulate our mental computations.

Indigenous Shamanic Worldviews as Eco-cosmologies and Indigenous Knowledge Systems of Sustainability

Abstract This article discusses shamanic worldviews theoretically and empirically based on the author’s long-term comparative studies in rural regions of Odisha, India. It deals with local eco-cosmological worldviews expressed in trance traditions, interpreted by the author as expressions of contemporary shamanic worldviews and an indication of ways of transmission of indigenous knowledge. In this sense the comparative examples of shamanic traditions of nag bacca (snake children) and alekh gurumai (ritual specialists) in Odisha are examples of transformative healing through trance rituals based both on concepts of holy craziness and sacred play (baaya/kheelo) and on spirit possession rituals (boil), widely spread in cultures of orality. In shamanic worldviews, therianthropic transformations of animal human and ecological encounters are transmitted in a rich culture of orality expressed in songs, performances, and trance dances. In this indigenous knowledge transmission, visions and dreams are the most important expressions of shamanic imaginaries, realities, epistemologies, and ontologies, revealing imagined, dreamt, and lived experiences of local shamanic societies. In this way, the visual mental imagery experiences construe the inner and outer knowledge of shamanic life worlds and worldviews.

Biased inferences about gender from names

How do alternative forms of reference to individuals—first, last, and full names—guide inferences about the gender of the referent? Given distributional correspondences between English first names and gender, first names provide probabilistic information about an individual's gender. While English last names do not vary with gender, men are more likely to be referred to by last name alone. Across four experiments, we demonstrate that inferences about gender are shaped by a persistent bias to infer that people are male, along with probabilistic information carried by the first name. When an individual was introduced by last name alone, participants overwhelmingly used he to subsequently refer to the person, suggesting that participants inferred that the person was male. This bias was still present when the individual was introduced using a first or full name, with participants less likely to use she than the distributional characteristics of the first names would predict. When explicitly asked to recall an individual’s gender who was introduced by last name alone, participants preferentially responded that the person was male. This bias persisted even when the person was introduced using a first or full name. Repeated reference attenuated, but did not eliminate, this bias. We discuss implications for models of how world knowledge is linked to language use.

A Prospective Metaverse Paradigm Based on the Reality-Virtuality Continuum and Digital Twins

Abstract: After decades of introducing the concept of virtual reality, the expansion, and significant advances of technologies and innovations, such as 6g, edge computing, the internet of things, robotics, artificial intelligence, blockchain, quantum computing, and digital twins, the world is on the cusp of a new revolution. By moving through the three stages of the digital twin, digital native, and finally surrealist, the metaverse has created a new vision of the future of human and societal life so that we are likely to face the next generation of societies (perhaps society 6) in the not too distant future. However, until then, the reality has been that the metaverse is still in its infancy, perhaps where the internet was in 1990. There is still no single definition, few studies have been conducted, there is no comprehensive and complete paradigm or clear framework, and due to the high financial volume of technology giants, most of these studies have focused on profitable areas such as gaming and entertainment. The motivation and purpose of this article are to introduce a prospective metaverse paradigm based on the revised reality-virtuality continuum and provide a new supporting taxonomy with the three dimensions of interaction, immersion, and extent of world knowledge to develop and strengthen the theoretical foundations of the metaverse and help researchers. Furthermore, there is still no comprehensive and agreed-upon conceptual framework for the metaverse. To this end, by reviewing the research literature, discovering the important components of technological building blocks, especially digital twins, and presenting a new concept called meta-twins, a prospective conceptual framework based on the revised reality-virtuality continuum with a new supporting taxonomy was presented.

A semi-supervised learning approach for semantic parsing boosted by BERT word embedding

In the modern world, structured and semi-structured knowledge bases hold a considerable amount of data. There-fore, people who are familiar with formal query languages should not be the only ones who can efficiently and clearly query them. Semantic Parsing (SP) is converting natural language utterances into formal meaning representations. The paper suggests a model for SP that uses a novel method of utilizing the Semi-Supervised Generative Adversarial Network (SS-GAN) to enhance the classifier performance. The proposed SS-GAN extends the fine-tuning of word embedding architectures using unlabeled examples in a generative adversarial environment. We provide a regularization strategy for addressing the mode missing problem and unstable training in SS-GAN. The main viewpoint is to use the extracted feature vectors from the discriminator. Hence, the generator produces outputs by aiding the discriminator’s learned features. A reconstruction loss is added to the loss function of the SS-GAN to drive the genera-tor to reconstruct outputs from the discriminator’s features, hence steering the generator toward actual data configurations. The proposed reconstruction loss improves the performance of SS-GAN, produces high-quality outputs, and may be combined with other regularization loss functions to improve the performance of diverse GANs. We employ BERT word embedding for our model, which can be included in a downstream task and fine-tuned as a model, while the pre-trained BERT model can capture various linguistic properties. We examine the suggested model using the WikiSQL and SparC datasets, and the analysis findings reveal our model outperforms its rivals. The findings from our experiments indicate that the need for labeled samples can be minimized, down to as few as 100 instances, while still achieving commendable classification outcomes.

Social Theory and Navigating Indeterminacy: A Configurational Analysis of Iranian Youth’s Identity Construction in Contemporary Iran

An emerging disparity within contemporary social science highlights a disconnection between the world in the process of metamorphosing and cosmopolitanization and the knowledge of the social world that is still trapped in the cognitive assumptions of modern episteme, which provided the conditions for the emergence of modern social sciences a century ago. This divide inhibits the efficacy of social analysis in comprehending and elucidating contemporary phenomena. This article advocates for a shift in the ontology of social theory and science towards the cosmopolitanization of the world, characterized by the prioritization of indeterminacy and fluidity in the construction of social phenomena. It investigates the epistemological implications and prerequisites of this ontological transformation, favoring a post-foundationalist approach as the most suitable epistemological framework. In response to the challenges posed by the uncertainty and indeterminacy of cosmopolitanization, after reviewing some of the existing theoretical efforts to address and provide alternatives to this challenge, the article proposes the examination of social configurations as the most fitting subjects for study. This approach necessitates the suspension of conventional, given, regulated categories, and trans-historical theories. It underscores the importance of recognizing configurations as incomplete, contingent units shaped within specific historical contexts and moments. The fluidity, relationality, and indeterminacy of configurations situated between the universal and the singular make them suitable for analysis at the level of particular. After elaborating on the most important features of social configurations, finally, by employing the proposed theoretical framework, this article aims to investigate its effectiveness in analyzing the process of identity construction among Iranian youth in Tehran in the context of the cosmopolitanization of reality, particularly in the face of the Islamist regime of Iran’s official politics of identity. Through a review and revision of selected empirical studies on youth identity construction in the consumer spaces of Tehran, based on the idea of social configurations within the framework of cosmopolitanization, it is argued that the genuine understanding of identity politics in contemporary Iran is not rooted in conventional analytical norms and categories but rather in a comprehensible conceptual apparatus characterized by fluidity and indeterminacy, capable of effectively making sense of the conflict between the politics of determinacy and indeterminacy in Iranian everyday life.

Elementary teachers as collaborators: developing educative support materials for citizen science projects

ABSTRACT Science education is an important component of a full education beginning in primary grades. In recent decades, research has identified young learners’ rich knowledge of the natural world and their potential to connect with sophisticated science ideas. Elementary teachers face many challenges to implementing reform-based science instruction in their classrooms. Some teachers may choose to enhance their students’ science experiences by introducing them to citizen science (CS) projects. Unfortunately, few CS projects offer substantial guidance for teachers seeking to implement the projects for instructional purposes, placing a heavy burdenon teachers. To address these burdens, our research team collaborated with Teacher Advisory Group (TAG teachers) during the development and revision of educative support materials for two CS projects. We present data about how the TAG teachers informed our CS support materials’ revisions, how they implemented the two CS projects with and without educative support materials, and how they perceived their students’ classroom and outdoor experiences with the CS projects. These data demonstrate the importance of including teachers’ voices and experiences in reform efforts, particularly when trying to incorporate instructional elements that teachers may perceive as deviations from what they are expected to teach.

Does word knowledge account for the effect of world knowledge on pronoun interpretation?

Abstract To what extent can statistical language knowledge account for the effects of world knowledge in language comprehension? We address this question by focusing on a core aspect of language understanding: pronoun resolution. While existing studies suggest that comprehenders use world knowledge to resolve pronouns, the distributional hypothesis and its operationalization in large language models (LLMs) provide an alternative account of how purely linguistic information could drive apparent world knowledge effects. We addressed these confounds in two experiments. In Experiment 1, we found a strong effect of world knowledge plausibility (measured using a norming study) on responses to comprehension questions that probed pronoun interpretation. In experiment 2, participants were slower to read continuations that contradicted world knowledge-consistent interpretations of a pronoun, implying that comprehenders deploy world knowledge spontaneously. Both effects persisted when controlling for the predictions of GPT-3, an LLM, suggesting that pronoun interpretation is at least partly driven by knowledge about the world and not the word. We propose two potential mechanisms by which knowledge-driven pronoun resolution occurs, based on validation- and expectation-driven discourse processes. The results suggest that while distributional information may capture some aspects of world knowledge, human comprehenders likely draw on other sources unavailable to LLMs.

MGPT: Few-Shot Learners Go Multilingual

Abstract This paper introduces mGPT, a multilingual variant of GPT-3, pretrained on 61 languages from 25 linguistically diverse language families using Wikipedia and the C4 Corpus. We detail the design and pretraining procedure. The models undergo an intrinsic and extrinsic evaluation: language modeling in all languages, downstream evaluation on cross-lingual NLU datasets and benchmarks in 33 languages, and world knowledge probing in 23 languages. The in-context learning abilities are on par with the contemporaneous language models while covering a larger number of languages, including underrepresented and low-resource languages of the Commonwealth of Independent States and the indigenous peoples in Russia. The source code and the language models are publicly available under the MIT license.

On Integrating Generative Models into Cognitive Architectures for Improved Computational Sociocultural Representations

What might the integration of cognitive architectures and generative models mean for sociocultural representations within both systems? Beyond just integration, we see this question as paramount to understanding the potential wider impact of integrations between these two types of computational systems. Generative models, though an imperfect representation of the world and various con-texts, nonetheless may be useful as general world knowledge with careful considerations of sociocultural representations provided therein, including the represented sociocultural systems or, as we explain, genres of the Human. Thus, such an integration gives an opportunity to develop cognitive models that represent from the physiological/biological time scale to the social timescale and that more accurately represent the effects of ongoing sociocultural systems and structures on behavior. In addition, integrating these systems should prove useful to audit and test many generative models under more realistic cognitive uses and conditions. That is, we can ask what it means that people will likely be using knowledge from such models as knowledge for their own behavior and actions. We further discuss these perspectives and focus these perspectives using ongoing and potential work with (primarily) the ACT-R cognitive architecture. We also discuss issues with using generative models as a system for integration.

Methodological Impasses: Facing Interrogation and Silence While Gathering Data on Sexual Violence in India

Feminist standpoint theorists highlight how relations of power and inequality impact our knowledge of the social world (Smith 1974). The hierarchical positioning of different social groups creates a world in which the experiences and perspectives of certain people are acknowledged while others are silenced (Hartsock 1998; Hekman 1997). Moreover, a researcher’s personal background—her race, gender, class, and sexuality, among other factors—condition what she is able to learn and how (Collins 2000). Together, this literature underscores how the social world—and who we are within it—shapes knowledge production.

‘People don’t buy art, they buy artists’: Robot artists – work, identity, and expertise

This article critically examines the construction of the artistic identity and career of Ai-Da, ‘the world's first ultra-realistic humanoid robot artist’. Engaging with scholarship on posthumanism and creative assemblages, and creative work, identity and expertise, this article conceptualises Ai-Da's distinctive positioning and focuses on the practices used to construct a creative worker identity and career. The article uses qualitative content analysis to examine journalistic coverage, promotional and presentation activities, exhibitions and performances, and social media postings over a four-year period from Ai-Da’s first launch to international visibility. The analysis shows how Ai-Da is positioned as a high-profile, border crossing artist, engaging in debates about Artificial Intelligence (AI), art, and the environment. It considers the creative assemblage of Ai-Da as a humanoid robot artist, the creator Aidan Meller and the team working with him, and the wider contextual factors of aesthetic expertise, networks and knowledge of art worlds which have shaped Ai-Da's artistic identity and career trajectory. The focus on how Ai-Da signals expertise on social media helps to frame the specific techniques used to speak about and for Ai-Da on social media platforms and wider media coverage. This includes articulating inspiration, showcasing artistic processes and cultivating audience relationships. In concluding, the implications of connecting critical perspectives on creative work with developments in art, AI and robot artists are explored: firstly, for understanding how the practices for constructing an artistic identity shape the development of robot artists; secondly, for understanding how developments in art and AI can frame reflections on artistic identity and careers.

Inductive Metaphysics Versus Logical Construction—Russell’s Methods and Realisms in 1912 and 1914

In his 1912 book The Problems of Philosophy, Bertrand Russell advocates an indirect realism with regard to physical objects. Only two years later, in his book Our Knowledge of the External World and the paper “The Relation of Sense-Data to Physics”, he changes his method in philosophy. Instead of inferring the existence of physical objects, he now sets out to construct them out of sense-data. As I will argue in this article, the main argument from The Problems of Philosophy can be rationally reconstructed as an inference to the best explanation which infers to unobservable objects. The main motivation for the new approach in Our Knowledge of the External World, on the other hand, is to establish a more direct variant of realism, in particular because Russell became skeptical with regard to inferences to unobservable objects. As I will argue, the resulting theory of the physical world loses so much in simplicity that it becomes an unattractive alternative to his earlier position, and Russell’s reason for rejecting simplicity as a criterion of theory choice turns out to be inconsequential.

Prekäres Weltwissen und kulturelles Erbe online. Onlineplattform vernetzt Lehre, Forschung und Wissenstransfer zum östlichen Christentum

The “Eastern Christian Studies Online Campus” (www.ku.de/ecsonca) is a multimedia online platform on the cultural heritage of Eastern Christianity with innovative teaching formats, international research projects and creative transfer services. The concept was developed by Prof. Dr. Thomas Kremer and Dipl.-Theol. Joachim Braun from the Chair of Theology of the Christian East at the Faculty of Theology of the Catholic University of Eichstätt-Ingolstadt. The Volkswagen Foundation is funding the project with 970 000 euros as part of the “World Knowledge - Structural Strengthening of Small Subjects” program. An initial workshop report provides an insight into the background to the work and the content of the online campus.

Modern finance through quantum computing-A systematic literature review.

Human intellectual restlessness originates from the need for knowledge of the modern world. The financial world is struggling to prototype accurate and fast data at low risk. The quantum approach to finance can support this desire. The goal of this paper is to provide a comprehensive review of the literature on how quantum computing can be used in finance. This research aims to expose an architecture of the state of the art in quantum finance. In terms of methodology, the PSALSAR framework was used to conduct this systematic literature review. The selection procedure followed the PRISMA guidelines and was applied in two databases (Web of Science and Scopus) without time limit. In total, 94 out of 1646 articles were included for data extraction and assessment of content evaluation covering the period 2001-2023. The current review of quantum finance literature is structured around the following themes: journals, research methods, tested data series, research topics in quantum finance, and future research directions. Within the financial sector, quantum computing is used in three main areas: simulation, optimization, and machine learning. These areas are supported by algorithms that have been created in recent years. Finally, we propose to highlight the benefits and the applications of quantum finance and to stimulate the interest in the future prospects of the debates.

Transformer networks of human conceptual knowledge.

We present a computational model capable of simulating aspects of human knowledge for thousands of real-world concepts. Our approach involves a pretrained transformer network that is further fine-tuned on large data sets of participant-generated feature norms. We show that such a model can successfully extrapolate from its training data, and predict human knowledge for new concepts and features. We apply our model to stimuli from 25 previous experiments in semantic cognition research and show that it reproduces many findings on semantic verification, concept typicality, feature distribution, and semantic similarity. We also compare our model against several variants, and by doing so, establish the model properties that are necessary for good prediction. The success of our approach shows how a combination of language data and (laboratory-based) psychological data can be used to build models with rich world knowledge. Such models can be used in the service of new psychological applications, such as the modeling of naturalistic semantic verification and knowledge retrieval, as well as the modeling of real-world categorization, decision-making, and reasoning. (PsycInfo Database Record (c) 2024 APA, all rights reserved).