Forecasting accuracy is bounded by the information available about the future. This paper makes that statement precise using information-theoretic tools. Under logarithmic loss, the expected performance of any probabilistic forecast decomposes into two parts: an irreducible component and an approximation component. The irreducible term is the conditional entropy of the future given the available information, while the approximation term is the divergence between the true conditional distribution and the forecasting method. The gap between this conditional-entropy limit and an unconditional baseline is exactly the mutual information between the future observation and the declared information set. This leads to a definition of forecastability as the maximum achievable reduction in expected log loss. Evaluated across horizons, forecastability forms a profile that describes how predictive information varies with lead time. This profile reflects the dependence structure of the process and need not be monotone: predictive information may be concentrated at particular lags, including seasonal horizons, even when intermediate horizons contain little useful signal. From this profile, the paper defines the informative horizon set: the horizons at which forecastability exceeds a practical threshold. At horizons not in this set, the achievable gain over the unconditional baseline is necessarily small, regardless of the forecasting method used. The framework therefore separates what is learnable from what is not, and distinguishes limits imposed by the data from errors introduced by modelling. The result is a pre-modelling diagnostic that identifies where meaningful prediction is feasible before any model is chosen, providing a principled basis for allocating modelling effort across forecast horizons.
This study explores the awareness and usage patterns of library resources and services among users of the Maulana Azad Library at Aligarh Muslim University (AMU). A survey methodology was employed, with 220 questionnaires distributed to library users, including undergraduate, postgraduate, research scholars, and faculty members. A total of 180 responses were received, resulting in an 81.81% response rate. The study found that a significant number of users visit the library primarily to borrow or return books, with books and journals being the most frequently utilized information sources. Most users are aware of copyright laws and regulations. However, many users lack familiarity with information literacy skills and techniques, leading to varied levels of awareness regarding library resources and services. While users expressed general satisfaction with the library's offerings, they suggested improvements, such as increasing access to electronic resources and extending library operating hours. The study further identifies factors influencing user awareness and usage, including academic discipline, research interests, and study habits. Based on these findings, the study recommends strategies to enhance library services, including improving user awareness, offering more user-friendly services, and providing continuous training and support to better meet the needs of library users.
As the complexity and scale of modern parallel machines continue to grow, programmers increasingly rely on composition of software libraries to encapsulate and exploit parallelism. However, many libraries are not designed with composition in mind and assume they have exclusive access to all resources. Using such libraries concurrently can result in contention and degraded performance. Prior solutions involve modifying the libraries or the OS, which is often infeasible. We propose Virtual Library Contexts (VLCs), which are process subunits that encapsulate sets of libraries and associated resource allocations. VLCs control the resource utilization of these libraries without modifying library code. This enables the user to partition resources between libraries to prevent contention, or load multiple copies of the same library to allow parallel execution of otherwise thread-unsafe code within the same process. In this paper, we describe and evaluate C++ and Python prototypes of VLCs. Experiments show VLCs enable a speedup up to 2.85x on benchmarks including applications using OpenMP, OpenBLAS, and LibTorch.
Thousands of Lithuanians, Latvians, and Estonians left their native lands during World War II. They escaped from the second Soviet occupation and spent the post-war years in displaced persons (DP) camps in various West European countries. The living conditions in the camps were very difficult, but the social and cultural activity of the DPs was vibrant. Book publishing was one expression of this activity. Since DPs from the Baltic States lived in the same camps, they had book connections. Several areas of their activity can be discussed.
One of these was the publishing activity of Baltic DP joint institutions such as the Baltic University, the Lithuanian-Latvian Unity, the Baltic Society of Philately "Baltia," and others. A few books were published as a result of joint Baltic efforts, such as art exhibitions, poetry competitions, or individual initiatives (e.g., Salomėja Narkėliūnaitė, Aleks Rannit, V. K. Jonynas, and others).
Many publishing ideas were not realized (e.g., a book about the Meerback Baltic DP camp, the Vorarlberg DP camp Baltic almanac "Lootus-Ceritas-Viltis," a collection of Baltic writers' prose in German, a map of the three Baltic states, and others). Some Latvian and Estonian DP books caused dissatisfaction among Lithuanian DPs as they allocated Vilnius and the Vilnius area to Poland.
The book connections between Lithuanian DPs and Latvian and Estonian DPs were sporadic, but this tradition continued after 1952 in various countries where Baltic emigrants lived.
Bibliography. Library science. Information resources
Using White Rose University Press (WRUP) as an example, this paper looks at the contribution of library publishing to the changing publishing landscape. It discusses the key drivers that have led the growth of library and institutional publishing, the service-led ethos that underpins much of this activity, and the future – both for WRUP and the growing community of open institutional publishers in the UK.
Bibliography. Library science. Information resources
Mostafa Pahlevanzadeh, Nadjla Hariri, Dariush Matlabi
et al.
Introduction The purpose of the current research is to design a knowledge management system performance evaluation model in the software industry using a neural network. Based on the collected data, a quantitative study was conducted to confirm the findings obtained from the qualitative stage. For exploratory study and extraction of categories related to evaluation factors, the meta-combination method (Sandelowski and Barroso model) was used. The research method in the quantitative part is descriptive survey. The statistical population of the research was made up of all software developers and software experts in universities and companies. Findings: 7 main categories including individual factors, organizational factors, technology and infrastructure, functional factors, knowledge management tools, economic factors, knowledge management tools, and 29 sub-categories were identified. The innovation of the research is building a model using neural network algorithms that have the ability to predict the performance evaluation index of the knowledge management system and the impact of each of the indicators using a neural network in the field of software. Conclusion: The results obtained from the questionnaire have been used for the input of the network model, the results showed that components such as technology infrastructure factors and functional factors have a greater impact on the evaluation of knowledge management performance in software development.Literature ReviewIn a research, they evaluated the performance of the knowledge management system in Iranian software companies. The results showed that the knowledge management system consists of 4 processes of identifying and creating, recording and maintaining, sharing and applying and internalizing knowledge. In a research, they designed a fuzzy inference system to evaluate the performance of the knowledge management system in the software development industry. The use of neural networks in evaluating the key factors of the knowledge management system in Iranian companies based in Alborz province was investigated. A research modeled an organizational knowledge management system based on artificial intelligence. Fuzzy theory was used to create knowledge extraction mechanism and reference model library from project model to dedicated reference model.MethodologyThe method used in this research is a mixed research method of exploratory type with a qualitative approach and meta-composite and Delphi methods. In the first stage, the meta-composite method was used to identify the main and sub-categories of the indicators, and then the validation and presentation of the final indicators were done with the fuzzy Delphi method. The current research method is practical in terms of purpose. The sample size was selected by simple random sampling method with Cochran's formula of 186 people. In the meta-combination method of the research, library sources and documents including articles, reliable and referable internet sources, as well as domestic and foreign scientific reports were used. For exploratory study and extraction of categories related to evaluation factors, the meta-combination method (Sandelowski and Barroso model) was used. Factors and dimensions of knowledge management system evaluation for which indicators are considered were provided to 20 members and experts. The implementation of the Delphi panel was carried out in two periods. Fuzzy Delphi method was used to screen and identify the final indicators and to answer the first and second questions of the research regarding the agreement of the experts of the research community regarding the obtained components, which includes software experts and knowledge management experts. 7 main categories including individual factors, organizational factors, technology and infrastructure, functional factors, knowledge management tools, economic factors, and 29 sub-categories were identified. In order to collect quantitative data, a researcher-made questionnaire (40 items) was used, the items of which were taken from the results of the meta-composite analysis in the first stage. In this research, in order to check the reliability of the research questionnaire, Cronbach's alpha coefficient was estimated at 0.89 for infrastructure factors and 0.88 for functional factors, respectively.Results In this research, the performance of the knowledge management system was evaluated with a neural network approach. Examining the results showed that the following components affect the evaluation of knowledge management performance in the software development industry. 1. Individual factors 2. Economic factors 3. Organizational factors 4. Knowledge management processes 5. Functional factors 6. Technological infrastructure factors 7. Knowledge management toolsDiscussionSolutions to improve the performance of knowledge management in the software development industry were presented: • Adjust the strategies in such a way that the creation of new knowledge, the application of new knowledge, its dissemination and sharing, and the storage and documentation of knowledge are explicitly considered. • Identifying influential people in the process of implementation and establishment of knowledge management, to improve the effective factors in the effective establishment of knowledge management more than in the past. • Developing procedures for documenting the experiences of experts in the software development industry on a continuous basis. • Managers and practitioners of the software industry should also consider parameters such as available budget, organizational culture, infrastructure, etc. • To provide the relevant managers and practitioners with a criterion for reviewing future policies and investments and help them make more appropriate decisions.ConclusionIn this research, 29 primary indicators have been identified based on the research literature, which include: • Organizational culture for sharing and using knowledge • Organizational Structure • The physical environment • Organization strategy • Support of senior managers such as motivation and commitment • Supporting innovations and digital technologies • Specialized knowledge of software development • General knowledge in software development • Involvement of developers • Education • Being up-to-date in the fields of specialized software • Knowledge and awareness of the knowledge management system • Correct understanding of system design requirements • Portals and portals of knowledge such as the Internet and email and social networks • MIS, Expert, DSS systems • Data warehouse - knowledge warehouse • Search and recovery tools and dashboard • Data security • The degree of integration of organizational systems • Quality of knowledge • Document management • Data management and workflow • Process Management • Creation and acquisition of knowledge, transfer and sharing of knowledge • Acquisition and use of knowledge • Operating cost of the software • Cost of software support.
Information technology, Bibliography. Library science. Information resources
The first manga café was established in Japan in the late 1970s. The original concept of a small library collection in a café, where visitors could stay longer than just the time it takes to drink one beverage, has been developed over several decades. Today, there are thousands of such cafés across the country, many of which are open 24 hours a day. This paper aims to highlight the significance of these alternative library spaces, which further promote reading practices. The phenomenon of manga cafés is viewed from a cultural perspective as a specialized public space that, in addition to providing a reading space, also functions as a potential sanctuary to those less socially secure. The trend of manga cafés has spread beyond the borders of Japan, with such spaces opened in East Asia, North America, and Europe. However, the concepts of manga cafés outside of Japan vary and are adapted to the logic of sustainability and profitability to the local culture.
Bibliography. Library science. Information resources
C.V. Raman (1888 - 1970) was a creative scientist, enthusiastic teacher and a science celebrity in India. In all these roles, he communicated science effectively. In this essay, I ask how and why did he communicate science. I take a few examples from his research writings and show his ability to explain science lucidly. By looking into his thoughts on teaching and those of his students, I explore Raman, the teacher. Finally, I discuss a few aspects of his methods to communicate science to the public. I emphasize his exposition and reveal a dichotomy.
The information bottleneck (IB) approach is popular to improve the generalization, robustness and explainability of deep neural networks. Essentially, it aims to find a minimum sufficient representation $\mathbf{t}$ by striking a trade-off between a compression term $I(\mathbf{x};\mathbf{t})$ and a prediction term $I(y;\mathbf{t})$, where $I(\cdot;\cdot)$ refers to the mutual information (MI). MI is for the IB for the most part expressed in terms of the Kullback-Leibler (KL) divergence, which in the regression case corresponds to prediction based on mean squared error (MSE) loss with Gaussian assumption and compression approximated by variational inference. In this paper, we study the IB principle for the regression problem and develop a new way to parameterize the IB with deep neural networks by exploiting favorable properties of the Cauchy-Schwarz (CS) divergence. By doing so, we move away from MSE-based regression and ease estimation by avoiding variational approximations or distributional assumptions. We investigate the improved generalization ability of our proposed CS-IB and demonstrate strong adversarial robustness guarantees. We demonstrate its superior performance on six real-world regression tasks over other popular deep IB approaches. We additionally observe that the solutions discovered by CS-IB always achieve the best trade-off between prediction accuracy and compression ratio in the information plane. The code is available at \url{https://github.com/SJYuCNEL/Cauchy-Schwarz-Information-Bottleneck}.
Robert Litschko, Oliver Kraus, Verena Blaschke
et al.
A large amount of local and culture-specific knowledge (e.g., people, traditions, food) can only be found in documents written in dialects. While there has been extensive research conducted on cross-lingual information retrieval (CLIR), the field of cross-dialect retrieval (CDIR) has received limited attention. Dialect retrieval poses unique challenges due to the limited availability of resources to train retrieval models and the high variability in non-standardized languages. We study these challenges on the example of German dialects and introduce the first German dialect retrieval dataset, dubbed WikiDIR, which consists of seven German dialects extracted from Wikipedia. Using WikiDIR, we demonstrate the weakness of lexical methods in dealing with high lexical variation in dialects. We further show that commonly used zero-shot cross-lingual transfer approach with multilingual encoders do not transfer well to extremely low-resource setups, motivating the need for resource-lean and dialect-specific retrieval models. We finally demonstrate that (document) translation is an effective way to reduce the dialect gap in CDIR.
This article highlights the creation of an automated 3D printed system created at a health sciences library at a large research university. As COVID-19 limited in-person interaction with 3D printers, a group of library staff came together to code a form that took users’ 3D printed files and connected them to machines automatically. A ticketing system and payment form was also automated via this system. The only in-person interactions are dedicated staff members that unload the prints. This article will describe the journey in getting to an automated system and share code and strategies so others can try it for themselves.
Bibliography. Library science. Information resources
In March and April of 2021, my co-investigators and I conducted semi-structured interviews with academic librarians across Canada about their work during the COVID-19 pandemic, which included their thoughts about going “back to normal.” Most participants were resistant to returning to the “old normal” without myriad changes inspired by the COVID-necessitated adaptations. However, there were concerns raised about whether or not their ideas would be implemented or even heard by their administrations. Additionally, many participants felt caught between proving their value through productive (and measurable) labour and the care-work that felt necessary and pressing but was not externally validated. This paper highlights the need for refocusing on building library collegial governance structures that include all library workers. As well, there is indication that the COVID-19 pandemic presents a unique opportunity to do so, as, removed from the “sacred space” (Ettarh 2018) of the library building, participants showed resistance to the austerity narratives typically invoked during a crisis. Embodying our values starts with establishing and building on shared library governance structures. If the changes inspired by COVID are to come to pass, then our vision of care and relationship-building must be inclusive to our own workers, to harness our collective power to build a future that works for everyone.
Bibliography. Library science. Information resources
Although there are various ways to represent data patterns and models, visualization has been primarily taught in many data science courses for its efficiency. Such vision-dependent output may cause critical barriers against those who are blind and visually impaired and people with learning disabilities. We argue that instructors need to teach multiple data representation methods so that all students can produce data products that are more accessible. In this paper, we argue that accessibility should be taught as early as the introductory course as part of the data science curriculum so that regardless of whether learners major in data science or not, they can have foundational exposure to accessibility. As data science educators who teach accessibility as part of our lower-division courses in two different institutions, we share specific examples that can be utilized by other data science instructors.
Information Extraction (IE) seeks to derive structured information from unstructured texts, often facing challenges in low-resource scenarios due to data scarcity and unseen classes. This paper presents a review of neural approaches to low-resource IE from \emph{traditional} and \emph{LLM-based} perspectives, systematically categorizing them into a fine-grained taxonomy. Then we conduct empirical study on LLM-based methods compared with previous state-of-the-art models, and discover that (1) well-tuned LMs are still predominant; (2) tuning open-resource LLMs and ICL with GPT family is promising in general; (3) the optimal LLM-based technical solution for low-resource IE can be task-dependent. In addition, we discuss low-resource IE with LLMs, highlight promising applications, and outline potential research directions. This survey aims to foster understanding of this field, inspire new ideas, and encourage widespread applications in both academia and industry.
Introduction. This study aims to explore effects of bubble filter and echo chamber on information searching behaviour on the internet, how it affects the scope of the internet ecosystem and users, and what we can see from the behaviour.
Data Collection Method. The paper used a mixed-methods approach with surveys, online discussions, and literature research. Twenty respondents between 19-21 years old participated in this exploratory study.
Analysis Data. Data was obtained and calculations have automatically accumulated through Google Form. The user data analysis and discussion were conducted manually by considering aspects of rationality concerning the literature and previous research.
Results and Discussions. The analysis results obtained were bubble filter and echo chambers were proven to affect internet users both in positive and negative ways.
Conclusions. Bubble filter and echo chamber may have both positive and negative effects simultaneously, however, it depends on the user because the system has provided control features to reduce both features. Using internet wisely is also parts of the important aspect.
Bibliography. Library science. Information resources
We consider a generalization of an important class of high-dimensional inference problems, namely spiked symmetric matrix models, often used as probabilistic models for principal component analysis. Such paradigmatic models have recently attracted a lot of attention from a number of communities due to their phenomenological richness with statistical-to-computational gaps, while remaining tractable. We rigorously establish the information-theoretic limits through the proof of single-letter formulas for the mutual information and minimum mean-square error. On a technical side we improve the recently introduced adaptive interpolation method, so that it can be used to study low-rank models (i.e., estimation problems of "tall matrices") in full generality, an important step towards the rigorous analysis of more complicated inference and learning models.