Andreas Varvarigos, Ali Maatouk, Jiasheng Zhang
et al.
While large language models (LLMs) have become the de facto framework for literature-related tasks, they still struggle to function as domain-specific literature agents due to their inability to connect pieces of knowledge and reason across domain-specific contexts, terminologies, and nomenclatures. This challenge underscores the need for a tool that facilitates such domain-specific adaptation and enables rigorous benchmarking across literature tasks. To that end, we introduce LitBench, a benchmarking tool designed to enable the development and evaluation of domain-specific LLMs tailored to literature-related tasks. At its core, LitBench uses a data curation process that generates domain-specific literature sub-graphs and constructs training and evaluation datasets based on the textual attributes of the resulting nodes and edges. The tool is designed for flexibility, supporting the curation of literature graphs across any domain chosen by the user, whether high-level fields or specialized interdisciplinary areas. In addition to dataset curation, LitBench defines a comprehensive suite of literature tasks, ranging from node and edge level analyses to advanced applications such as related work generation. These tasks enable LLMs to internalize domain-specific knowledge and relationships embedded in the curated graph during training, while also supporting rigorous evaluation of model performance. Our results show that small domain-specific LLMs trained and evaluated on LitBench datasets achieve competitive performance compared to state-of-the-art models like GPT-4o and DeepSeek-R1. To enhance accessibility and ease of use, we open-source the tool along with an AI agent tool that streamlines data curation, model training, and evaluation.
Research has demonstrated that policing is a stressful occupation and that this stress has a negative impact on police officers’ mental and physical health, performance, and interactions with citizens. Mental health at the workplace has become a concern due to the costs of depression, anxiety, burnout, and even suicide, which is high among police officers. To ameliorate occupational health, it is therefore crucial to identify stress and burnout levels on a regular basis. However, the instruments frequently used to measure stress have not valorized the specificity of policing tasks. This study aims to: (i) conduct a literature review to identify questionnaires used to assess occupational stress and burnout among police officers; (ii) analyze the psychometric characteristics of a Portuguese version of Operational Police Stress Questionnaire (PSQ-Op); and, using the PSQ-Op and other questionnaires, (iii) to identify operational stress, burnout, and distress levels among Portuguese police officers. The literature review identified 108 studies which use a multiplicity of questionnaires to measure burnout or occupational stress among police officers, but few studies use specific police stress questionnaires. Sample sizes were mostly below 500 participants and studies were mainly developed in the last decade in the USA and Brazil, but also in another 24 countries, showing the extent of the interest in this topic. This study applied to 2057 police officers from the National Portuguese Police, a force policing urban centers, and used the PSQ-Op, as well the Spanish Burnout Inventory and the Kessler Psychological Distress Scale. The results show that the psychometric properties of the Portuguese version of PSQ-Op are adequate. Factorial analysis revealed two dimensions defined as social and work issues, which were associated with measures of distress and burnout. Fit indices suggested a second-order solution called operational police stress. Overall, and considering the scale range of each questionnaire, the results showed moderate values of operational stress, distress, and burnout. However, considering their cut-off points, 85% of the sample presented high operational stress levels, 11% critical values for burnout, and 28% high distress levels, with 55% of the sample at risk of a psychological disorder. These results reinforce the need to prevent stress and to invest in police officers’ occupational health.
Cliffed (and rocky) coasts are geomorphic features occurring in about 80% of the coastline of the world and are strongly influenced by a broad range of both natural and anthropogenic processes that may cause serious erosion problems. Since the sea wave motion is a fundamental driver of cliff erosion, the cliffs become sensitive to increasing of global sea levels and to extreme weather events, which are both associated with global warming. Because of its importance, a considerable amount of investigations on coastal cliff erosion (CCE) were reported during the last decades. A bibliometric analysis is an useful tool to identify patterns of a given theme from a large body of academic literature. There is no previous evidence of a global bibliometric analysis in the literature in English on themes of CCE. Therefore, the aim of this article was to carry out a bibliometric analysis from Scopus database of CCE for the period 2000–2023. Once obtained, two filtering steps for selection of documents consisting of a custom R script implementation and a careful reading of the remaining documents were applied. During the search, a dynamic approach that puts emphasis on the processes operating on rocky coasts was selected instead of an evolutionary geological perspective. The final list reached 583 documents. A second aim was to discuss the research trends and challenges based on the latest highly-cited documents. As main result, the trend of the scientific production in the theme of CCE had an increasing interest over the last years, with an average compound annual growth rate of 15.6%. On the other side, the results demonstrated that even though the USA took the second place, European countries (United Kingdom, Italy, France, Portugal, Spain and Poland) lead the ranking; therefore, there is a scarcity of knowledge about the theme in large regions such as South America and Africa where seacliffs are dominants.
Large language models (LLMs) have grown in their usage to provide support for question answering across numerous disciplines. The models on their own have already shown promise for answering basic questions, however fail quickly where expert domain knowledge is required or the question is nuanced. Scientific research often involves searching for relevant literature, distilling pertinent information from that literature and analysing how the findings support or contradict one another. The information is often encapsulated in the full text body of research articles, rather than just in the abstracts. Statements within these articles frequently require the wider article context to be fully understood. We have built an LLM-based system that performs such search and distillation of information encapsulated in scientific literature, and we evaluate our keyword based search and information distillation system against a set of biology related questions from previously released literature benchmarks. We demonstrate sparse retrieval methods exhibit results close to state of the art without the need for dense retrieval, with its associated infrastructure and complexity overhead. We also show how to increase the coverage of relevant documents for literature review generation.
The digital transformation of our society is a constant challenge, as data is generated in almost every digital interaction. To use data effectively, it must be of high quality. This raises the question: what exactly is data quality? A systematic literature review of the existing literature shows that data quality is a multifaceted concept, characterized by a number of quality dimensions. However, the definitions of data quality vary widely. We used feature-oriented domain analysis to specify a taxonomy of data quality definitions and to classify the existing definitions. This allows us to identify research gaps and future topics.
Context: Dynamic production environments make it challenging to maintain reliable machine learning (ML) systems. Runtime issues, such as changes in data patterns or operating contexts, that degrade model performance are a common occurrence in production settings. Monitoring enables early detection and mitigation of these runtime issues, helping maintain users' trust and prevent unwanted consequences for organizations. Aim: This study aims to provide a comprehensive overview of the ML monitoring literature. Method: We conducted a multivocal literature review (MLR) following the well established guidelines by Garousi to investigate various aspects of ML monitoring approaches in 136 papers. Results: We analyzed selected studies based on four key areas: (1) the motivations, goals, and context; (2) the monitored aspects, specific techniques, metrics, and tools; (3) the contributions and benefits; and (4) the current limitations. We also discuss several insights found in the studies, their implications, and recommendations for future research and practice. Conclusion: Our MLR identifies and summarizes ML monitoring practices and gaps, emphasizing similarities and disconnects between formal and gray literature. Our study is valuable for both academics and practitioners, as it helps select appropriate solutions, highlights limitations in current approaches, and provides future directions for research and tool development.
Maria Diniz Nogueira, Ana Maria Marinho Diniz, Cláudia Tartaglia Reis
et al.
Abstract Introduction: Patient safety (PS) is a fundamental pillar of healthcare quality. The effective dissemination of information and knowledge management regarding PS is essential to foster safe practices. Despite significant progress in recent years, gaps remain in how knowledge and information on PS are managed and disseminated across healthcare organizations. Objective: The aim of this study was to map the scientific evidence concerning knowledge management strategies and the dissemination of information related to PS, as implemented by healthcare organizations for healthcare professionals. Methods: A scoping review was conducted in accordance with the Joanna Briggs Institute (JBI) methodology. Searches were performed in Portuguese, English, and Spanish, with no time limitations, across databases including PubMed, Scopus, and Web of Science, as well as grey literature sources such as NOVA Discovery (EBSCO) and selected institutional websites. The review was reported in line with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. Results: Of the 247 publications identified, 16 were included. Three principal strategies for PS-related knowledge management emerged: acquisition, dissemination, and mediation. Continuous training and the adoption of innovative educational methodologies were found to enhance knowledge acquisition. Digital platforms and health marketing tools facilitated dissemination. Knowledge mediation was supported by strategic leadership, informal networks, and interdisciplinary partnerships involving risk managers, clinical leaders, and subject matter experts. Discussion: Knowledge management strategies demonstrated potential to strengthen PS, with continuous training, organizational culture, and innovation standing out in knowledge acquisition. Dissemination was effective through digital platforms and marketing, while mediation relied on leaders and managers. Challenges remain, such as validating the impact and updating the content. Future research should assess the impact of these strategies on clinical practice, including the perspectives of patients and carers.
At the University of New Mexico (UNM), about 19.2% of undergraduate courses are taught by graduate teaching assistants (UGW 2021). Within UNM’s Department of Spanish and Portuguese, the majority of enrolled Spanish graduate students are instructors of Spanish as a Heritage Language (SHL) undergraduate courses—a language retrieval course designed for students with cultural connections to the Spanish language. Oftentimes these graduate students may step into these roles with minimal teaching experience, and specifically, a lack of training in Culturally Responsive Teaching (CRT) methodologies. While there is extant literature on the CRT within classrooms, little research reflects on the unique experiences of graduate students who teach (Haynie & Spong, 2022). As instructors in introductory courses that can significantly influence undergraduate students’ lives and academic trajectories, learning to incorporate culturally responsive practices within their teaching is critical. We reflect on our time as UNM graduate students as we entered the roles of SHL instructors and the lack of pedagogical and methodological preparation that would have better equipped us for teaching in SHL classrooms. This work centers our own testimonios, having gone through the process of self-learning, as a means of grounding our enseñanzas, or teaching moments, and consejos, or pieces of advice, so that we may provide a) a point of reference for graduate students in need of clarity surrounding how to incorporate CRT in their classrooms, and b) a call to action towards the incorporation of CRT methods in the initial and continuous training and development of SHL instructors.
Diabetes is associated with a number of complications, particularly if glycaemic targets are not achieved. Glycaemic control is highly linked to treatment persistence and adherence. To understand the burden of poor persistence and adherence, this systematic literature review identified existing evidence regarding basal insulin adherence/non-adherence and persistence/non-persistence among people with diabetes in Western Europe (defined as the UK, France, Spain, Switzerland, the Netherlands, Ireland, Austria, Portugal, Denmark, Norway, Sweden, Finland, Italy, Germany, Iceland and Belgium). Eligible studies were systematically identified from two databases, Medline and Embase (published between 2012 and June 2022). Conference abstracts from ISPOR and EASD were manually included. Identified studies were screened by two independent reviewers in a two-step blinded process. The eligibility of studies was decided on the basis of pre-established criteria. A proportional meta-analysis and comparative narrative analyses were conducted to analyse the included studies. Twelve studies were identified. Proportions of adherence/non-adherence and persistence/non-persistence varied across studies. Pooled rates of non-persistence at 6, 12 and 18 months were 20.3% (95% CI 13.8; 27.8), 33.8% (95% CI 24.1; 44.3) and 36.5% (95% CI 33.6; 39.4), respectively. In the literature, the proportion of adherent people ranged from 41% to 64% (using the outcome measure medication possession ratio (MPR) > 80%), with a pooled rate of 55.6% (95% CI 45.3; 65.6), suggesting that approximately 44% of people with type 2 diabetes (T2D) are non-adherent. The results highlight that almost half of patients with T2D in Western Europe have poor adherence to insulin therapy and, at 18 months, one in three patients do not persist on treatment. These findings call for new basal insulin therapies and diabetes management strategies that can improve treatment persistence and adherence among people with T2D.
HIGHLIGHTS Participatory forest management follows diverse stages and methods according to the desired level of stakeholder engagement. The use of participatory approaches is advocated to improve forest governance, by integrating different stakeholders' voices and views. Participatory methods in forest management seldom assure stakeholders' empowerment. Social benefits of participatory methods in forest management appear more prominently than economic and environmental ones. Social benefits are often intertwined with the economic and environmental value of forests. SUMMARY Understanding participatory processes and identifying successful implementation methods is key to effective bottom-up sustainable forest management strategies. This paper aims to contribute to that understanding by systematically reviewing the literature dealing with participatory methods to forest management in five European Mediterranean countries (France, Greece, Italy, Portugal and Spain), specifically identifying the relationship between the level of stakeholder involvement, type of stakeholders and the methods applied, as well as the results obtained, and the main challenges identified. Our findings show that stakeholders commonly involved are NGOs, landowners, companies, and local government bodies. A strong correspondence between the methods used and the purpose and level of stakeholders' involvement is identified. Social benefits stand out as the most relevant result of this engagement somehow overshadowing economic and environmental ones. Results show that participatory forest management methods are simple and do not depend on sophisticated methods and techniques, so that their wider application depends above all on the will to do so. Comprendre les processus participatifs et identifier les méthodes de mise en œuvre efficaces est essentiel pour élaborer des stratégies ‘bottom-up’ effectives pour la gestion durable des forêts. Cet article vise à contribuer à cette compréhension en examinant systématiquement la littérature traitant des méthodes participatives de gestion forestière dans cinq pays méditerranéens européens (France, Grèce, Italie, Portugal et Espagne), en identifiant spécifiquement la relation entre le niveau d’engagement des acteurs concernés, le type d’acteurs et les méthodes appliquées, ainsi que les résultats obtenus et les principaux défis identifiés. Nos résultats montrent que les acteurs concernés généralement engagés sont les ONG, les propriétaires fonciers, les entreprises et les autorités locales. Une forte correspondance entre les méthodes utilisées et l’objectif et le niveau d’engagement des acteurs est identifiée. Les bénéfices sociaux ressortent comme le résultat le plus pertinent de cet engagement, éclipsant d'une certaine manière les avantages économiques et environnementaux. Les résultats montrent que les méthodes de gestion forestière participative sont simples et ne dépendent pas de méthodes et de techniques sophistiquées, de façon que leur application plus large dépend avant tout de la volonté de le faire. La comprensión de los procesos participativos y la identificación de métodos de aplicación eficaces son fundamentales para la eficacia de las estrategias de gestión forestal sostenible desde la base. Este trabajo tiene como objetivo contribuir a esa comprensión mediante una revisión sistemática de la literatura sobre los métodos participativos para la gestión forestal en cinco países mediterráneos europeos (Francia, Grecia, Italia, Portugal y España), identificando específicamente la relación entre el nivel de participación de los agentes interesados, el tipo de agentes y los métodos aplicados, así como los resultados obtenidos y los principales desafíos identificados. Nuestros resultados muestran que las partes interesadas que suelen participar son ONG, propietarios de terrenos, empresas y organismos de la administración local. Se observa una fuerte correspondencia entre los métodos utilizados y el propósito con el nivel de participación de los agentes. Los beneficios sociales se destacan como el resultado más relevante de esta participación, eclipsando de alguna manera a los beneficios económicos y medioambientales. Los resultados muestran que los métodos de gestión forestal participativa son sencillos y no dependen de métodos y técnicas sofisticados, por lo que su aplicación más amplia depende sobre todo de la voluntad de hacerlo. A compreensão dos processos participativos e a identificação de métodos de implementação bem sucedidos são fundamentais para estratégias eficazes de gestão florestal sustentável de tipo ‘bottom-up’. O presente trabalho visa contribuir para essa compreensão através de uma revisão sistemática da literatura sobre métodos participativos de gestão florestal em cinco países europeus mediterrânicos (França, Grécia, Itália, Portugal e Espanha), identificando especificamente a relação entre o nível de envolvimento dos agentes interessados, o tipo de agentes e os métodos aplicados, bem como os resultados obtidos e os principais desafios identificados. Os resultados mostram que os agentes interessados habitualmente envolvidos são as ONG, os proprietários florestais, as empresas e os organismos da administração local. Verifica-se ainda uma forte correspondência entre os métodos utilizados e o objetivo com o nível de envolvimento dos agentes. Os benefícios sociais destacam-se como o resultado mais relevante deste envolvimento, ultrapassando em certa medida os benefícios económicos e ambientais. Os resultados demonstram que os métodos de gestão florestal participativa são simples e não dependem de métodos e técnicas sofisticados, pelo que a sua aplicação mais alargada dependerá sobretudo da vontade de o fazer.
Algorithm design is a vital skill developed in most undergraduate Computer Science (CS) programs, but few research studies focus on pedagogy related to algorithms coursework. To understand the work that has been done in the area, we present a systematic survey and literature review of CS Education studies. We search for research that is both related to algorithm design and evaluated on undergraduate-level students. Across all papers in the ACM Digital Library prior to August 2023, we only find 94 such papers. We first classify these papers by topic, evaluation metric, evaluation methods, and intervention target. Through our classification, we find a broad sparsity of papers which indicates that many open questions remain about teaching algorithm design, with each algorithm topic only being discussed in between 0 and 10 papers. We also note the need for papers using rigorous research methods, as only 38 out of 88 papers presenting quantitative data use statistical tests, and only 15 out of 45 papers presenting qualitative data use a coding scheme. Only 17 papers report controlled trials. We then synthesize the results of the existing literature to give insights into what the corpus reveals about how we should teach algorithms. Much of the literature explores implementing well-established practices, such as active learning or automated assessment, in the algorithms classroom. However, there are algorithms-specific results as well: a number of papers find that students may under-utilize certain algorithmic design techniques, and studies describe a variety of ways to select algorithms problems that increase student engagement and learning. The results we present, along with the publicly available set of papers collected, provide a detailed representation of the current corpus of CS Education work related to algorithm design and can orient further research in the area.
Abstract. Automatically generating scientific literature surveys is a valuable task that can significantly enhance research efficiency. However, the diverse and complex nature of information within a literature survey poses substantial challenges for generative models. In this paper, we design a series of prompts to systematically leverage large language models (LLMs), enabling the creation of comprehensive literature surveys through a step-by-step approach. Specifically, we design prompts to guide LLMs to sequentially generate the title, abstract, hierarchical headings, and the main content of the literature survey. We argue that this design enables the generation of the headings from a high-level perspective. During the content generation process, this design effectively harnesses relevant information while minimizing costs by restricting the length of both input and output content in LLM queries. Our implementation with Qwen-long achieved third place in the NLPCC 2024 Scientific Literature Survey Generation evaluation task, with an overall score only 0.03% lower than the second-place team. Additionally, our soft heading recall is 95.84%, the second best among the submissions. Thanks to the efficient prompt design and the low cost of the Qwen-long API, our method reduces the expense for generating each literature survey to 0.1 RMB, enhancing the practical value of our method.
Alexander Naumann, Felix Hertlein, Laura Dörr
et al.
Computer vision applications in transportation logistics and warehousing have a huge potential for process automation. We present a structured literature review on research in the field to help leverage this potential. The literature is categorized w.r.t. the application, i.e. the task it tackles and w.r.t. the computer vision techniques that are used. Regarding applications, we subdivide the literature in two areas: Monitoring, i.e. observing and retrieving relevant information from the environment, and manipulation, where approaches are used to analyze and interact with the environment. Additionally, we point out directions for future research and link to recent developments in computer vision that are suitable for application in logistics. Finally, we present an overview of existing datasets and industrial solutions. The results of our analysis are also available online at https://a-nau.github.io/cv-in-logistics.
This article conducts a literature review on the topic of monetary policy in developing countries and focuses on the effectiveness of monetary policy in promoting economic growth and the relationship between monetary policy and economic growth. The literature review finds that the activities of central banks in developing countries are often overlooked by economic models, but recent studies have shown that there are many factors that can affect the effectiveness of monetary policy in these countries. These factors include the profitability of central banks and monetary unions, the independence of central banks in their operations, and lags, rigidities, and disequilibrium analysis. The literature review also finds that studies on the topic have produced mixed results, with some studies finding that monetary policy has a limited or non-existent impact on economic growth and others finding that it plays a crucial role. The article aims to provide a comprehensive understanding of the current state of research in this field and to identify areas for future study.