Studiul investighează tranziția de la alfabetul chirilic la alfabetul latin în documentația oficială a instituțiilor de învățământ superior din Republica Moldova, în contextul aplicării Legii privind funcționarea limbilor (1989). Analiza proceselor-verbale universitare, a ordinelor rectorale și a dosarelor studențești evidențiază caracterul complex al transformării, cu dimensiuni tehnico-administrative și identitare. Ritmul implementării a variat în funcție de profilul instituțiilor: facultățile umaniste (Institutul Pedagogic „A. Russo”, Universitatea de Stat din Chișinău) au adoptat rapid grafia latină și glotonimul „limba română”, în timp ce unitățile tehnice și reale (Institutul Politehnic „S. Lazo”, unele catedre științifice) au menținut o perioadă bilingvismul. La Institutul Agricol „M. V. Frunze”, schimbarea s-a produs gradual. Concluziile subliniază rolul universităților ca spații de legitimare a alfabetului latin și a redefinirii
glotonimice de la „limba moldovenească” la „limba română”.
Cuvinte-cheie: alfabet latin, procese-verbale universitare, legislație lingvistică, identitate națională, învățământ superior, limba română, glotonim, tranziție lingvistică
DOI: https://doi.org/10.59295/sum10(220)2025_18
History of scholarship and learning. The humanities
Purpose: The objective of this study is to explore and comprehend the factors from the perceived environment that impact travellers' attitudes and trust in agritourism at farms integrated with aquaculture, which have been creatively adapted for tourism purposes. These findings contribute to understanding how agritourism fosters rural innovation and sustainable development by transforming traditional agricultural practices into diversified tourism experiences. The findings of the study could demonstrate that certain outcomes play a crucial role in the successful innovation of rural areas. Methodology/design/approach: The study extensively utilized the Theory of Planned Behavior framework to develop its measurement constructs. Data collection occurred in regions where tourists frequented farms that combine aquaculture with traditional farming practices, yielding a total sample size of 332 respondents. The data were analyzed using the Partial Least Squares Structural Equation Modeling (PLS-SEM) method, employing the SmartPLS software version 4.0.9.2. Results: The results identified factors perceived environmentally positive influence on personal perception. Attitude and trust were found to mediate the relationships between perceived environment and revisit intention, with the mediating effect of attitude being stronger than that of trust. Originality of the research: Visitor attitudes significantly determine the innovation from making farms, orchards, aquaculture areas to the experiential tourism business. Successful innovations, such as enhancing rural incomes and sustaining agricultural livelihoods through agritourism transformation, are significantly driven by positive visitor perceptions and trust.
History of scholarship and learning. The humanities, Social sciences (General)
Felix Olajide Talabi, Christiana Shade Ade-johnson, Joseph Moyinoluwa Talabi
et al.
Abstract The wave of artificial intelligence (AI) is transforming all spheres of human life. AI is continuously expanding, shaping the future of humanity and raising important ethical and societal implications. Hence, this study explored the bandwagon effect of AI and its use among media houses in Oyo State, Nigeria. The study adopted the ethnographic qualitative design, chiefly utilising focus group discussion (FGD to gain rich empirical insight into the phenomenon. Twelve media professionals were purposively sampled for the FGD. The study found that AI is becoming more prevalent in Oyo State, Nigerian media houses for tasks like generating content, analysing data, verifying facts, and managing social media. The study concluded that AI is revolutionising the media industry and can serve as a competitive edge for media houses that embrace it, bearing in mind that responsible use, ethical considerations, and technical challenges are crucial for harnessing AI’s potential.
History of scholarship and learning. The humanities, Social Sciences
As digital transformation (Digital) accelerates globally, conventional enterprise production models are proving increasingly insufficient to meet the demands of today’s dynamic market landscape. China has innovated the concept of New Quality Productivity (NQPF), and exploring its functioning is critical to promoting high-quality enterprise development. This study examines the impact mechanism of Digital on NQPF in manufacturing firms by applying spatial econometric models—including the spatial Durbin model, spatial mediation model, and spatial threshold model—to panel data from A-share listed manufacturers (2013–2022). The results indicate that digital transformation significantly influences the level of NQPF, exhibiting spatial spillover effects and spatial attenuation boundaries. This influence initially promotes and subsequently inhibits productivity. The analysis of the spatial mediation effect reveals that Digital affects enterprise productivity levels by influencing total factor productivity. Furthermore, the spatial threshold effect analysis indicates that higher total enterprise assets enhance the positive impact of Digital on NQPF. These results provide robust micro-level empirical evidence to inform manufacturing enterprise development strategies.
History of scholarship and learning. The humanities, Social Sciences
Off-dynamics reinforcement learning (RL), where training and deployment transition dynamics are different, can be formulated as learning in a robust Markov decision process (RMDP) where uncertainties in transition dynamics are imposed. Existing literature mostly assumes access to generative models allowing arbitrary state-action queries or pre-collected datasets with a good state coverage of the deployment environment, bypassing the challenge of exploration. In this work, we study a more realistic and challenging setting where the agent is limited to online interaction with the training environment. To capture the intrinsic difficulty of exploration in online RMDPs, we introduce the supremal visitation ratio, a novel quantity that measures the mismatch between the training dynamics and the deployment dynamics. We show that if this ratio is unbounded, online learning becomes exponentially hard. We propose the first computationally efficient algorithm that achieves sublinear regret in online RMDPs with $f$-divergence based transition uncertainties. We also establish matching regret lower bounds, demonstrating that our algorithm achieves optimal dependence on both the supremal visitation ratio and the number of interaction episodes. Finally, we validate our theoretical results through comprehensive numerical experiments.
The accelerated evolution of digital infrastructures and algorithmic systems is reshaping how the humanities engage with knowledge and culture. Rooted in the traditions of Digital Humanities and Digital Humanism, the concept of "Cyber Humanities" proposes a critical reconfiguration of humanistic inquiry for the post-digital era. This Manifesto introduces a flexible framework that integrates ethical design, sustainable digital practices, and participatory knowledge systems grounded in human-centered approaches. By means of a Decalogue of foundational principles, the Manifesto invites the scientific community to critically examine and reimagine the algorithmic infrastructures that influence culture, creativity, and collective memory. Rather than being a simple extension of existing practices, "Cyber Humanities" should be understood as a foundational paradigm for humanistic inquiry in a computationally mediated world. Keywords: Cyber Humanities, Digital Humanities, Transdisciplinary Epistemology, Algorithmic Reflexivity, Human-centered AI, Ethics-by-Design, Knowledge Ecosystems, Digital Sovereignty, Cognitive Infrastructures
Francisco Mena, Dino Ienco, Cassio F. Dantas
et al.
Multi-modal co-learning is emerging as an effective paradigm in machine learning, enabling models to collaboratively learn from different modalities to enhance single-modality predictions. Earth Observation (EO) represents a quintessential domain for multi-modal data analysis, wherein diverse remote sensors collect data to sense our planet. This unprecedented volume of data introduces novel challenges. Specifically, the access to the same sensor modalities at both training and inference stages becomes increasingly complex based on real-world constraints affecting remote sensing platforms. In this context, multi-modal co-learning presents a promising strategy to leverage the vast amount of sensor-derived data available at the training stage to improve single-modality models for inference-time deployment. Most current research efforts focus on designing customized solutions for either particular downstream tasks or specific modalities available at the inference stage. To address this, we propose a novel multi-modal co-learning framework capable of generalizing across various tasks without targeting a specific modality for inference. Our approach combines contrastive and modality discriminative learning together to guide single-modality models to structure the internal model manifold into modality-shared and modality-specific information. We evaluate our framework on four EO benchmarks spanning classification and regression tasks across different sensor modalities, where only one of the modalities available during training is accessible at inference time. Our results demonstrate consistent predictive improvements over state-of-the-art approaches from the recent machine learning and computer vision literature, as well as EO-specific methods. The obtained findings validate our framework in the single-modality inference scenarios across a diverse range of EO applications.
Alakh Sharma, Gaurish Trivedi, Kartikey Singh Bhandari
et al.
Scalable multi-agent reinforcement learning (MARL) remains a central challenge for AI. Existing population-based methods, like Policy-Space Response Oracles, PSRO, require storing explicit policy populations and constructing full payoff matrices, incurring quadratic computation and linear memory costs. We present Generative Evolutionary Meta-Solver (GEMS), a surrogate-free framework that replaces explicit populations with a compact set of latent anchors and a single amortized generator. Instead of exhaustively constructing the payoff matrix, GEMS relies on unbiased Monte Carlo rollouts, multiplicative-weights meta-dynamics, and a model-free empirical-Bernstein UCB oracle to adaptively expand the policy set. Best responses are trained within the generator using an advantage-based trust-region objective, eliminating the need to store and train separate actors. We evaluated GEMS in a variety of Two-player and Multi-Player games such as the Deceptive Messages Game, Kuhn Poker and Multi-Particle environment. We find that GEMS is up to ~$\mathbf{6\times}$ faster, has $\mathbf{1.3\times}$ less memory usage than PSRO, while also reaps higher rewards simultaneously. These results demonstrate that GEMS retains the game theoretic guarantees of PSRO, while overcoming its fundamental inefficiencies, hence enabling scalable multi-agent learning in multiple domains.
Labeling is required by the interpretive system. When a head merges with a phrase, the head provides the label. However, lexical heads and T with poor inflectional features are too weak to be labels. Although insightful, this theory leaves at least one problem that needs prompt solutions: are there other kinds of weak heads? In this paper, we address this issue by proposing that phonological features play a crucial role in the labeling algorithm and by putting forward an additional version of weak heads. That is, a head that loses phonological features in the syntax is also weak. This approach to weak heads, together with the constraint that a structure must be labeled for interpretation, can capture the distribution of empty categories in topicalization, relativization, ellipsis, and other phenomena, some of which have not received enough scholarly attention. Therefore, our syntactic-phonological approach to labeling can open up new possibilities to account for the distribution of empty categories in a principled manner.
History of scholarship and learning. The humanities, Social Sciences
Abstract While many organizations have successfully leveraged big data analytics capabilities to improve their performance, our understanding is limited on whether and how big data analytics capabilities affect social innovation in organizations. Based on the organizational information processing theory and the organizational learning theory, this study aims to investigate how big data analytics capabilities support social innovation, and how knowledge ambidexterity mediates this relationship. A total of 354 high-tech companies in China, this study shows that big data analytics management, big data analytics technology, and big data analytics personnel capabilities all have positive effects on social innovation. In addition, both knowledge exploration and knowledge exploitation play a mediating role in this process. Furthermore, a polynomial regression and response surface analysis shows that social innovation increases when knowledge exploration and knowledge exploitation are highly consistent but declines when knowledge exploration and knowledge exploitation are inconsistent. This study not only provides new perspectives for understanding how big data analytics capabilities contribute to social innovation, complementing the existing literature on big data analytics capabilities and social innovation, but also provides important practical guidance on how organizations can develop big data analytics capabilities to improve social innovation and solve social problems in the digital age.
History of scholarship and learning. The humanities, Social Sciences
Predicting the biophysical and functional properties of proteins is essential for in silico protein design. Machine learning has emerged as a promising technique for such prediction tasks. However, the relative scarcity of in vitro annotations means that these models often have little, or no, specific data on the desired fitness prediction task. As a result of limited data, protein language models (PLMs) are typically trained on general protein sequence modeling tasks, and then fine-tuned, or applied zero-shot, to protein fitness prediction. When no task data is available, the models make strong assumptions about the correlation between the protein sequence likelihood and fitness scores. In contrast, we propose meta-learning over a distribution of standard fitness prediction tasks, and demonstrate positive transfer to unseen fitness prediction tasks. Our method, called Metalic (Meta-Learning In-Context), uses in-context learning and fine-tuning, when data is available, to adapt to new tasks. Crucially, fine-tuning enables considerable generalization, even though it is not accounted for during meta-training. Our fine-tuned models achieve strong results with 18 times fewer parameters than state-of-the-art models. Moreover, our method sets a new state-of-the-art in low-data settings on ProteinGym, an established fitness-prediction benchmark. Due to data scarcity, we believe meta-learning will play a pivotal role in advancing protein engineering.
The high costs and risks involved in extensive environment interactions hinder the practical application of current online safe reinforcement learning (RL) methods. While offline safe RL addresses this by learning policies from static datasets, the performance therein is usually limited due to reliance on data quality and challenges with out-of-distribution (OOD) actions. Inspired by recent successes in offline-to-online (O2O) RL, it is crucial to explore whether offline safe RL can be leveraged to facilitate faster and safer online policy learning, a direction that has yet to be fully investigated. To fill this gap, we first demonstrate that naively applying existing O2O algorithms from standard RL would not work well in the safe RL setting due to two unique challenges: \emph{erroneous Q-estimations}, resulted from offline-online objective mismatch and offline cost sparsity, and \emph{Lagrangian mismatch}, resulted from difficulties in aligning Lagrange multipliers between offline and online policies. To address these challenges, we introduce \textbf{Marvel}, a novel framework for O2O safe RL, comprising two key components that work in concert: \emph{Value Pre-Alignment} to align the Q-functions with the underlying truth before online learning, and \emph{Adaptive PID Control} to effectively adjust the Lagrange multipliers during online finetuning. Extensive experiments demonstrate that Marvel significantly outperforms existing baselines in both reward maximization and safety constraint satisfaction. By introducing the first policy-finetuning based framework for O2O safe RL, which is compatible with many offline and online safe RL methods, our work has the great potential to advance the field towards more efficient and practical safe RL solutions.
We present Gradient Boosting Reinforcement Learning (GBRL), a framework that adapts the strengths of gradient boosting trees (GBT) to reinforcement learning (RL) tasks. While neural networks (NNs) have become the de facto choice for RL, they face significant challenges with structured and categorical features and tend to generalize poorly to out-of-distribution samples. These are challenges for which GBTs have traditionally excelled in supervised learning. However, GBT's application in RL has been limited. The design of traditional GBT libraries is optimized for static datasets with fixed labels, making them incompatible with RL's dynamic nature, where both state distributions and reward signals evolve during training. GBRL overcomes this limitation by continuously interleaving tree construction with environment interaction. Through extensive experiments, we demonstrate that GBRL outperforms NNs in domains with structured observations and categorical features while maintaining competitive performance on standard continuous control benchmarks. Like its supervised learning counterpart, GBRL demonstrates superior robustness to out-of-distribution samples and better handles irregular state-action relationships.
Learning outcomes have an essential role as a measuring tool for achieving the goals of learning activities. However, there is still material below the average value, one of which is Islamic history material, because the learning seems monotonous, so some students are not focused and lack attention, making it difficult to understand the lesson. This study aimed to increase students' activity and learning outcomes in Islamic history material in class 2 I KMI at Darussalam Gontor Islamic Boarding School. The research method used classroom action research with the model Kemmis and McTaggart. Applying Magic Box learning media in class 2I in Islamic history subjects at Darussalam Gontor Boarding school produces the following research results: 1) student activeness In cycle 1 reached a percentage of 75.2%; in cycle 2, the increase in student activity reached 83.5%. The comparison of increased active learning cycles 1 and 2 is 30%. 2). Students who got scores above the average, In cycle 1, students who got scores above the average, namely 23 out of 40, reached 57.5% of all students with an average of 45.75. In cycle 2, students got scores above the average, namely 36 out of 40, achieving 90% of all students with an average of 74.25. The comparison of increased student learning outcomes in cycles 1 and 2 is 35%.
Abstract This paper used two waves (2016 and 2018) of longitudinal data from the China Families Panel Survey (CFPS) to analyze the economic impact of Non-communicable chronic diseases (NCDs) on individual earned income using propensity score matching and difference in difference (PSM-DID) methods to control for potential confounding. The occurrence of a NCDs was associated with a significant decrease in earned income by 19.2% (P = 0.002, t = 3.75). The reasons for this decrease include: a lower labour force participation rate; lower weekly hours worked; and a lower average hourly wage. After holding labour market behaviours constant, different types of NCDs have different impacts on earned income. Musculoskeletal diseases have the greatest negative impact, accounting for a 21.5% decrease in individual earned income (p < 0.0001, t = −7.84), while digestive system diseases have the smallest impact accounting for a 6.9% decrease in earned income (p = 0.012, t = −2.52).
History of scholarship and learning. The humanities, Social Sciences
Andrii Voshchepynets, Oleksiy Agapitov, Lynn Wilson
et al.
We present the results of processing the effects of the powerful Gamma Ray Burst GRB221009A captured by the charged particle detectors (electrostatic analyzers and solid-state detectors) onboard spacecraft at different points in the heliosphere on October 9, 2022. To follow the GRB221009A propagation through the heliosphere we used the electron and proton flux measurements from solar missions Solar Orbiter and STEREO-A; Earth magnetosphere and the solar wind missions THEMIS and Wind; meteorological satellites POES15, POES19, MetOp3; and MAVEN - a NASA mission orbiting Mars. GRB221009A had a structure of four bursts: less intense Pulse 1 - the triggering impulse - was detected by gamma-ray observatories at 131659 UT (near the Earth); the most intense Pulses 2 and 3 were detected on board all the spacecraft from the list, and Pulse 4 detected in more than 500 s after Pulse 1. Due to their different scientific objectives, the spacecraft, which data was used in this study, were separated by more than 1 AU (Solar Orbiter and MAVEN). This enabled tracking GRB221009A as it was propagating across the heliosphere. STEREO-A was the first to register Pulse 2 and 3 of the GRB, almost 100 seconds before their detection by spacecraft in the vicinity of Earth. MAVEN detected GRB221009A Pulses 2, 3, and 4 at the orbit of Mars about 237 seconds after their detection near Earth. By processing the time delays observed we show that the source location of the GRB221009A was at RA 288.5 degrees, Dec 18.5 degrees (J2000) with an error cone of 2 degrees
The domain of machine learning is confronted with a crucial research area known as class imbalance learning, which presents considerable hurdles in precise classification of minority classes. This issue can result in biased models where the majority class takes precedence in the training process, leading to the underrepresentation of the minority class. The random vector functional link (RVFL) network is a widely used and effective learning model for classification due to its good generalization performance and efficiency. However, it suffers when dealing with imbalanced datasets. To overcome this limitation, we propose a novel graph embedded intuitionistic fuzzy RVFL for class imbalance learning (GE-IFRVFL-CIL) model incorporating a weighting mechanism to handle imbalanced datasets. The proposed GE-IFRVFL-CIL model offers plethora of benefits: $(i)$ leveraging graph embedding to preserve the inherent topological structure of the datasets, $(ii)$ employing intuitionistic fuzzy theory to handle uncertainty and imprecision in the data, $(iii)$ and the most important, it tackles class imbalance learning. The amalgamation of a weighting scheme, graph embedding, and intuitionistic fuzzy sets leads to the superior performance of the proposed models on KEEL benchmark imbalanced datasets with and without Gaussian noise. Furthermore, we implemented the proposed GE-IFRVFL-CIL on the ADNI dataset and achieved promising results, demonstrating the model's effectiveness in real-world applications. The proposed GE-IFRVFL-CIL model offers a promising solution to address the class imbalance issue, mitigates the detrimental effect of noise and outliers, and preserves the inherent geometrical structures of the dataset.
The ability to explain decisions made by machine learning models remains one of the most significant hurdles towards widespread adoption of AI in highly sensitive areas such as medicine, cybersecurity or autonomous driving. Great interest exists in understanding which features of the input data prompt model decision making. In this contribution, we propose a novel approach to identify relevant features of the input data, inspired by methods from the energy landscapes field, developed in the physical sciences. By identifying conserved weights within groups of minima of the loss landscapes, we can identify the drivers of model decision making. Analogues to this idea exist in the molecular sciences, where coordinate invariants or order parameters are employed to identify critical features of a molecule. However, no such approach exists for machine learning loss landscapes. We will demonstrate the applicability of energy landscape methods to machine learning models and give examples, both synthetic and from the real world, for how these methods can help to make models more interpretable.
AbstractGiven a set of target language documents and their translators, the translator attribution task aims at identifying which translator translated which documents. The attribution and the identification of the translator’s style could contribute to fields including translation studies, digital humanities, and forensic linguistics. To conduct this investigation, firstly, we develop a new corpus containing the translations of world-famous books into Arabic. We then pre-process the books in our corpus which mainly involves cleaning irrelevant material, morphological segmentation analysis of words, and devocalization. After pre-processing the books, we propose to use 100 most frequent words and/or morphologically segmented function words as writing style markers of the translators (i.e. stylometric features) to differentiate between translations of different translators. After the completion of features extraction process, we applied several supervised and unsupervised machine-learning algorithms along with our novel cluster-to-author index to perform this task. We found that the translators are not invisible, and morphological analysis may not be more useful than just using the 100 most frequent words as features. The support vector machine linear kernel algorithm reported 99% classification accuracy. Similar findings were reported by the unsupervised machine-learning methods, namely, K-mean clustering and hierarchical clustering.