We give new differentially private algorithms for the classic problems of learning decision lists and large-margin halfspaces in the PAC and online models. In the PAC model, we give a computationally efficient algorithm for learning decision lists with minimal sample overhead over the best non-private algorithms. In the online model, we give a private analog of the influential Winnow algorithm for learning halfspaces with mistake bound polylogarithmic in the dimension and inverse polynomial in the margin. As an application, we describe how to privately learn decision lists in the online model, qualitatively matching state-of-the art non-private guarantees.
Abstract The transition to electric vehicles is reshaping industrial structures and regional economies, particularly in countries with strong automotive sectors. This study investigates how regional network centrality within the automotive industry affects regional economic growth, focusing on the moderating roles of industrial specialization and diversity. While regions with higher centrality are expected to benefit from denser supply-chain linkages and knowledge flows, the magnitude of this effect depends on the composition of their industrial base. Using inter-firm transaction data from 15 South Korean provinces between 2008 and 2021, regional centrality is derived from social network analysis using degree centrality. Industrial specialization and diversity are measured to reflect the structural composition of regional industries. Panel regressions are employed to assess both the direct and interactive effects of these variables on regional economies. The results show that regional centrality has a significant positive effect on economic growth, confirming that regions occupying central positions in industrial networks achieve stronger economic performance. However, specialization and diversity weaken this positive effect, whereas a balanced interaction among the three factors enhances economic outcomes. These findings highlight the importance of network structures and industrial composition in shaping regional development dynamics. The study provides empirical evidence and policy insights for governments seeking to strengthen regional competitiveness and foster sustainable industrial transformation during technological transitions.
History of scholarship and learning. The humanities, Social Sciences
Proponemos un estudio del filme Hojas de otoño (2023) de Aki Kaurismäki en términos de una reflexión cinemática acerca de cómo es posible perseverar afectivamente en el mundo contemporáneo dominado por la lógica de la extracción de plusvalía por medio del trabajo asalariado. Luego de una contextualización de este filme dentro la obra de Kaurismäki, donde destaca una reflexión sobre el mundo laboral contemporáneo, argumentamos que en su cine se da una revalorización del amor romántico. Aunque ciertamente escenificado por motivos heterosexuales tradicionales, se trata de una forma de amor romántico que implica un intento rupturista por encontrar una suerte de refugio afectivo en medio de una estructura de la sociedad y el mundo que tiende a la aniquilación de la vitalidad. De ahí que propongamos la idea de un “conservadurismo paradojal” en los motivos del amor romántico en Kaurismäki: paradojal, puesto que las formas tradicionales del amor con las cuales piensa Hojas de otoño implican formas de potenciamiento afectivo y no simplemente la mantención del statu quo.
History of scholarship and learning. The humanities, Philology. Linguistics
This paper presents a comprehensive comparative analysis of Natural Language Processing (NLP) methods for automated toxicity detection in online gaming chats. Traditional machine learning models with embeddings, large language models (LLMs) with zero-shot and few-shot prompting, fine-tuned transformer models, and retrieval-augmented generation (RAG) approaches are evaluated. The evaluation framework assesses three critical dimensions: classification accuracy, processing speed, and computational costs. A hybrid moderation system architecture is proposed that optimizes human moderator workload through automated detection and incorporates continuous learning mechanisms. The experimental results demonstrate significant performance variations across methods, with fine-tuned DistilBERT achieving optimal accuracy-cost trade-offs. The findings provide empirical evidence for deploying cost-effective, efficient content moderation systems in dynamic online gaming environments.
Yaswanth Chittepu, Blossom Metevier, Will Schwarzer
et al.
Existing approaches to language model alignment often treat safety as a tradeoff against helpfulness, which can lead to unacceptable responses in sensitive domains. To ensure reliable performance in such settings, we propose High-Confidence Safe Reinforcement Learning from Human Feedback (HC-RLHF), a method that provides high-confidence safety guarantees while maximizing helpfulness. Similar to previous methods, HC-RLHF explicitly decouples human preferences into helpfulness and harmlessness (safety), which are learned by training a reward model and a cost model, respectively. It then employs a two-step process to find safe solutions. In the first step, it optimizes the reward function under an intentionally pessimistic version of the cost constraint. In the second step, the trained model undergoes a safety test to verify whether its performance stays within an upper-confidence bound of the actual cost constraint. We provide a theoretical analysis of HC-RLHF, including proof that it will not return an unsafe solution with a probability greater than a user-specified threshold. For our empirical analysis, we apply HC-RLHF to align three different language models (Qwen2-1.5B, Qwen2.5-3B, and LLaMa3.2-3B) with human preferences. Our results demonstrate that HC-RLHF produces safe models with high probability and can improve harmlessness and helpfulness compared to previous methods.
A people-centred approach placing value on humanity is at the core of numerous democratic-led governments, and educational institutions to solve socio-economic and environmental challenges facing civilisation. This paper explored the contribution of Scholarship of Teaching and Learning (SoTL) in emphasising reflection and theories to enable people to improve teaching and learning and to reinforce the application of logic when addressing barriers confronting humanity. The purpose of this study was to examine the relevance of SoTL platforms using reflection and theories to improve the quality of teaching and learning by promoting intellectual and ethical virtues in human interactions. Qualitative data was generated from literature and a purposively sampled respondent group of twelve lecturers in one of the South African universities. Interviews were conducted by using an interview schedule questionnaire whilst data collected was categorised and analysed into themes. The findings revealed that SoTL encourages knowledge sharing and inspires academics to reflect on their practice by answering some of the questions that hinder effective teaching and learning and its snowball effect on improved human interactions in general. Answering complex questions relating to teaching and learning requires all parties beyond borders to share ideas, reflect, think critically, engage in research, and apply relevant theories that embed values that inform human interactions. The implication is that SoTL discourse can eliminate the silo and irrational ways of solving problems by employing inquiry and critical reflective strategies to stimulate reasoning and restore values that position humanity at the centre of governments and all human interactions. Therefore, SoTL approaches can foster intellectual decisions that can sustain the upholding of ethics and intellectual virtues and advance human dignity. Keywords: Humanity, Reflection, Scholarship of Teaching and Learning, Theories
This article focuses on che raised. fourth or Lydian scale degree of the scale in Chopin's mazurkas from the viewpoint of the stylised mazurka's fusion of folkloric elements and nineteenth-century tonal language. A brief exposition of the functions of the raised fourth and its integradon into the tonal structures of the mazurkas is given. An investigation of the raised fourth in three selected mazurkas· reveals a significant expansion of its role musi~al meaning, function and effect, As an authentic, modal, structural, motivic and form-generating device, it is in terms of integral co tonal elaboration, development, unification, colouration and contrast.
History of scholarship and learning. The humanities, Political science
Md. Touhidul Islam, Md. Mahadi Hasan, Md. Redwanuzzaman
et al.
This research paper explores the practices of artificial intelligence (AI) and its impact on improving businesses in Bangladesh through the review of existing literature and primary data analysis. Additionally, a conceptual analysis has been conducted on key AI concepts and their potential applications in the Bangladeshi business context. The study's objectives include providing a detailed understanding of specific AI technologies, measuring the benefits of AI, assessing the challenges faced by the industry in implementing AI and finding probable solutions to overcome these challenges. Data were collected from 120 respondents from 10 types of businesses in Bangladesh and analyzed using MS Excel and SPSS (V. 25). According to the results, using AI in enterprises may have a significant positive impact on efficiency, decision-making, production, cost, fraud detection, and supply chain optimization. However, obstacles to AI deployment include a lack of qualified personnel, poor data quality, money, infrastructure, and legal frameworks. Businesses should raise employee understanding of AI, look for diversified financing and qualified personnel, work with the government on infrastructure support and legislation, address concerns about job displacement through training, and encourage employee acceptance of change in order to overcome these challenges. Businesses in Bangladesh may improve operations and competitiveness by using these techniques. Business executives, decision-makers, and academics interested in maximizing the potential of AI and enhancing business outcomes in Bangladesh might learn from this study.
History of scholarship and learning. The humanities, Social sciences (General)
Abstract The purpose of this study is to examine how economic development and stability are affected by carbon trading pilot programs. Using panel data from 31 Chinese provinces and autonomous regions between 2005 and 2021, two hypotheses were tested using the DID model. The findings indicate that (1) the carbon trading market pilot program has an immediate effect on economic stability. (2) High-quality economic development is positively and significantly impacted by the carbon trading market pilot program. Regional heterogeneity exists in the effects of carbon trading pilot programs on high-quality development and economic stability. The relationship between carbon trading pilot programs and economic development is not conclusive, despite the growing number of these policies. Given this, additional investigation into this connection is required. Understanding the results of carbon trading pilot programs can be used to gauge how successful these initiatives are. This research adds to the body of knowledge regarding the impact of the carbon trading pilot programs. It then makes policy recommendations that may serve as a guide for future “double carbon” research.
History of scholarship and learning. The humanities, Social Sciences
Liqun Zhao, Keyan Miao, Konstantinos Gatsis
et al.
Reinforcement learning (RL) excels in applications such as video games, but ensuring safety as well as the ability to achieve the specified goals remains challenging when using RL for real-world problems, such as human-aligned tasks where human safety is paramount. This paper provides safety and stability definitions for such human-aligned tasks, and then proposes an algorithm that leverages neural ordinary differential equations (NODEs) to predict human and robot movements and integrates the control barrier function (CBF) and control Lyapunov function (CLF) with the actor-critic method to help to maintain the safety and stability for human-aligned tasks. Simulation results show that the algorithm helps the controlled robot to reach the desired goal state with fewer safety violations and better sample efficiency compared to other methods in a human-aligned task.
In the evolving field of machine learning, ensuring group fairness has become a critical concern, prompting the development of algorithms designed to mitigate bias in decision-making processes. Group fairness refers to the principle that a model's decisions should be equitable across different groups defined by sensitive attributes such as gender or race, ensuring that individuals from privileged groups and unprivileged groups are treated fairly and receive similar outcomes. However, achieving fairness in the presence of group-specific concept drift remains an unexplored frontier, and our research represents pioneering efforts in this regard. Group-specific concept drift refers to situations where one group experiences concept drift over time while another does not, leading to a decrease in fairness even if accuracy remains fairly stable. Within the framework of Federated Learning, where clients collaboratively train models, its distributed nature further amplifies these challenges since each client can experience group-specific concept drift independently while still sharing the same underlying concept, creating a complex and dynamic environment for maintaining fairness. The most significant contribution of our research is the formalization and introduction of the problem of group-specific concept drift and its distributed counterpart, shedding light on its critical importance in the field of fairness. Additionally, leveraging insights from prior research, we adapt an existing distributed concept drift adaptation algorithm to tackle group-specific distributed concept drift which uses a multi-model approach, a local group-specific drift detection mechanism, and continuous clustering of models over time. The findings from our experiments highlight the importance of addressing group-specific concept drift and its distributed counterpart to advance fairness in machine learning.
Moritz Lampert, Christopher Blöcker, Ingo Scholtes
Dynamic link prediction is an important problem considered in many recent works that propose approaches for learning temporal edge patterns. To assess their efficacy, models are evaluated on continuous-time and discrete-time temporal graph datasets, typically using a traditional batch-oriented evaluation setup. However, as we show in this work, a batch-oriented evaluation is often unsuitable and can cause several issues. Grouping edges into fixed-sized batches regardless of their occurrence time leads to information loss or leakage, depending on the temporal granularity of the data. Furthermore, fixed-size batches create time windows with different durations, resulting in an inconsistent dynamic link prediction task. In this work, we empirically show how traditional batch-based evaluation leads to skewed model performance and hinders the fair comparison of methods. We mitigate this problem by reformulating dynamic link prediction as a link forecasting task that better accounts for temporal information present in the data.
This paper presents a novel technique for incorporating user input when learning and inferring user preferences. When trying to provide users of black-box machine learning models with actionable recourse, we often wish to incorporate their personal preferences about the ease of modifying each individual feature. These recourse finding algorithms usually require an exhaustive set of tuples associating each feature to its cost of modification. Since it is hard to obtain such costs by directly surveying humans, in this paper, we propose the use of the Bradley-Terry model to automatically infer feature-wise costs using non-exhaustive human comparison surveys. We propose that users only provide inputs comparing entire recourses, with all candidate feature modifications, determining which recourses are easier to implement relative to others, without explicit quantification of their costs. We demonstrate the efficient learning of individual feature costs using MAP estimates, and show that these non-exhaustive human surveys, which do not necessarily contain data for each feature pair comparison, are sufficient to learn an exhaustive set of feature costs, where each feature is associated with a modification cost.
Agne Brandisauskiene, Loreta Buksnyte-Marmiene, Jurate Cesnaviciene
et al.
This study aimed to investigate the role of teacher support and equity in the relationship between teacher’s autonomy-promoting behavior and students’ learning strategies. The approach examines the direct relationship between teacher’s autonomy-supportive behavior and students’ learning strategies and via perceived teacher support. It also discusses the effect of perceived equity for the relationship between autonomy-supportive behavior in teacher’s and teacher support. Data were obtained from 24 secondary schools in nine Lithuanian municipalities with poor socioeconomic level contexts ( N = 632 pupils). The findings revealed that teacher’s autonomy-supportive behavior is directly associated to student’s greater use of learning strategies, as well as through a mediator—student’s perceptions of teacher support. The association between teacher’s autonomy-supportive behavior and teacher support is moderated by students’ perceived equity so that the positive relationship is stronger for students with a higher than with a lower perceived equity. This study adds to the understanding of the importance of teacher behavior for students’ learning by concentrating on equity, which is especially essential for students from low-income families. Teachers may foster equity by providing more possibilities for autonomy for all students, creating a supportive classroom environment and inviting students to be active participants in the learning process.
History of scholarship and learning. The humanities, Social Sciences
Jupyter Notebook fornece um ambiente onde você pode trabalhar com facilidade o seu código na linguagem Python. Esta lição descreve como instalar o software Jupyter Notebook, como executar e criar ficheiros para o Jupyter Notebook.
History of scholarship and learning. The humanities, Computer software
AbstractThe use of computing tools and methods has irreversibly impacted the creation, use and communication of research. As a result, a still divisive movement of digital humanities (DH) has emerged over the last few decades. This article attempts to provide a theoretical contribution to the discussion of the core fundamentals of the field. In order to do so, it takes a sample of papers indexed under Library and Information Science, in the Web of Knowledge database, and studies them using a quantitative data analysis and a qualitative literature review combined with the author’s personal reflection to illustrate the main research topics. The notion of scholarly primitives, initially formulated by John Unsworth, provides the background for the theoretical analysis. It is concluded that DH embody a community patterned by collaborative and shared networks of communication, where digital tools amplify research possibilities without changing the humanistic values of its practitioners.
Souradeep Dutta, Kaustubh Sridhar, Osbert Bastani
et al.
Long horizon robot learning tasks with sparse rewards pose a significant challenge for current reinforcement learning algorithms. A key feature enabling humans to learn challenging control tasks is that they often receive expert intervention that enables them to understand the high-level structure of the task before mastering low-level control actions. We propose a framework for leveraging expert intervention to solve long-horizon reinforcement learning tasks. We consider \emph{option templates}, which are specifications encoding a potential option that can be trained using reinforcement learning. We formulate expert intervention as allowing the agent to execute option templates before learning an implementation. This enables them to use an option, before committing costly resources to learning it. We evaluate our approach on three challenging reinforcement learning problems, showing that it outperforms state-of-the-art approaches by two orders of magnitude. Videos of trained agents and our code can be found at: https://sites.google.com/view/stickymittens
Se trata de una inmersión en el MAR - Museo de Arte de Rio, Río de Janeiro, Brasil. El propósito de esta inmersión era entender cómo un museo permite conexiones entre educación y arte a través de prácticas educativas inventivas. Educación, mediación y prácticas educativas inventivas se convirtieron en las nociones para operar esta investigación. Los resultados de la investigación han ayudado a comprender que las prácticas educativas inventivas, construidas por educadores/mediadores de museos, pueden producir nuevas experiencias de enseñanza y aprendizaje. Además, posibilitaron pensar el museo como un espacio susceptible de producir experiencias de mediación y aprendizaje con sus públicos. Finalmente, cabe señalar que ésta investigación se llevó a cabo en consonancia con el Grupo de Investigación de Currículo, Espaço, Movimento (CEM); y fue apoyada por la Coordinación de Mejoramiento del Personal de Educación Superior (CAPES, por sus siglas en portugués).
Education (General), History of scholarship and learning. The humanities
Raimunda Nonata Monteiro, Enaile do Espírito Santo Iadanza, Helena Maria Martins Lastres
El presente Dossier titulado Amazonia: Cultura, Educación y Memoria, es la segunda parte de la compilación de textos iniciada con el Dossier Amazonia Brasileña: ocupación y políticas socioambientales en la transición de los siglos XX al XXI. Esta edición reúne cinco artículos que discuten vivencias, imaginarios y expresiones culturales de pueblos tradicionales e indígenas. Es posible percibir visiones del mundo que expresan las especificidades y la diversidad de saberes, haceres y resistencias de grupos sociales representativos de las poblaciones amazónicas. Se describen cosmovisiones, pertenencias y representaciones sociales, así como las rupturas ocasionadas por las transformaciones impuestas por el "progreso". Los autores de este Dossier nos sumergen en la vida de comunidades y pueblos y nos proporcionan una pequeña inmersión en concepciones del mundo que forman los multiversos culturales de la Amazonia. También muestran que la Historia de la región se enseña mediante narrativas que ignoran las existencias y lugares (lugar sin lugar), generalizan y homogeneizan la Amazonia, reducida a las percepciones de los colonizadores.
History of scholarship and learning. The humanities, Law in general. Comparative and uniform law. Jurisprudence