Abstract: We present a formal and constructive theory showing that probabilistic finite automata (PFAs) can be exactly simulated using symbolic feedforward neural networks. Our architecture represents state distributions as vectors and transitions as stochastic matrices, enabling probabilistic state propagation via matrix-vector products. This yields a parallel, interpretable, and differentiable simulation of PFA dynamics using soft updates-without recurrence. We formally characterize probabilistic subset construction, $\varepsilon$-closure, and exact simulation via layered symbolic computation, and prove equivalence between PFAs and specific classes of neural networks. We further show that these symbolic simulators are not only expressive but learnable: trained with standard gradient descent-based optimization on labeled sequence data, they recover the exact behavior of ground-truth PFAs. This learnability, formalized in Proposition 5.1, is the crux of this work. Our results unify probabilistic automata theory with neural architectures under a rigorous algebraic framework, bridging the gap between symbolic computation and deep learning.

Abstract: Hybrid optical neural networks (HONNs) offload some electronic computation to optical preprocessors to achieve low-power and fast training and inference phases in machine learning tasks. Our contribution to the development of HONNs is a spectral-methods paradigm for building synthetic training data for machine-learned models. Here, our synthetic training image data does not resemble the image test data. As a result, the neural network focuses on learning specific features parameterized by the synthetic training data. Within this paradigm, a dataset's singular value decomposition entropy indicates learnability, i.e., how rapidly a model converges. Subsequently, we train a neural network model to rapidly learn specific features for further downstream analyses.

Abstract: We study the PAC property of scenario decision-making algorithms, that is, the ability to make a decision that has an arbitrarily low risk of violating an unknown safety constraint, provided sufficiently many realizations (called scenarios) of the safety constraint are sampled. Sufficient conditions for scenario decision-making algorithms to be PAC are available in the literature, such as finiteness of the VC dimension of its associated classifier and existence of a compression scheme. We study the question of whether these sufficient conditions are also necessary. We show with counterexamples that this is not the case in general. This contrasts with binary classification learning, for which the analogous conditions are sufficient and necessary. Popular scenario decision-making algorithms, such as scenario optimization, enjoy additional properties, such as stability and consistency. We show that even under these additional assumptions the above conclusions hold. Finally, we derive a necessary condition for scenario decision-making algorithms to be PAC, inspired by the VC dimension and the so-called no-free-lunch theorem.

Abstract: This study aims to evaluate the usability level of the Livin' by Mandiri mobile banking application using five dimensions developed by Jakob Nielsen, namely learnability, efficiency, memorability, errors, and user satisfaction. The research method used is a quantitative approach with Partial Least Square-Structural Equation Modeling (PLS-SEM) analysis techniques, and refers to the international standard ISO 9241-11 which focuses on user experience and perception. Data were collected through an online questionnaire distributed to 70 respondents who are active users of the Livin' by Mandiri application. This study emphasizes the importance of understanding the overall user experience in the context of digital banking applications. The results of the analysis show that of the five dimensions analyzed, only learnability has a significant influence on the perception of the application's overall usability. This indicates that the ease of learning and using the application from the beginning is a key factor in determining how well the application is accepted and utilized by users. Meanwhile, the other dimensions, although still relevant in the evaluation context, did not show a significant influence on the perception of usability. These findings reinforce the view that each usability dimension can have different impacts depending on user characteristics and the context in which the app is used.

Abstract: Machine learning algorithms often encounter different or "out-of-distribution" (OOD) data at deployment time, and OOD detection is frequently employed to detect these examples. While it works reasonably well in practice, existing theoretical results on OOD detection are highly pessimistic. In this work, we take a closer look at this problem, and make a distinction between uniform and non-uniform learnability, following PAC learning theory. We characterize under what conditions OOD detection is uniformly and non-uniformly learnable, and we show that in several cases, non-uniform learnability turns a number of negative results into positive. In all cases where OOD detection is learnable, we provide concrete learning algorithms and a sample-complexity analysis.

Abstract: Website Daftarmenu.com adalah platform digital yang digunakan oleh restoran untuk menampilkan daftar menu secara daring melalui pemindaian barcode di meja makan. Penelitian ini bertujuan untuk mengevaluasi tingkat usability situs tersebut berdasarkan lima prinsip usability menurut Jakob Nielsen, yaitu Learnability (X1), Efficiency (X2), Memorability (X3), Error (X4), dan Satisfaction (X5) terhadap variabel Usability (Y). Metode yang digunakan adalah survei kuantitatif dengan penyebaran kuesioner secara daring kepada 72 responden. Analisis data dilakukan menggunakan perangkat lunak SmartPLS 3.0. Hasil penelitian menunjukkan bahwa hanya variabel Efficiency (t = 3,860; p = 0,000) yang berpengaruh signifikan terhadap Usability pada tingkat signifikansi 5%. Sementara itu, variabel Learnability (p = 0,086), Memorability (p = 0,504), Error (p = 0,661), dan Satisfaction (p = 0,064) tidak memiliki pengaruh yang signifikan. Kesimpulannya, pengembang Daftarmenu.com disarankan untuk memprioritaskan peningkatan aspek efisiensi dalam desain dan pengalaman pengguna, karena aspek tersebut terbukti paling berkontribusi terhadap persepsi usability secara keseluruhan.

Abstract: <p>We present a unified framework bridging the information-theoretic foundations of symbolic systems and the cognitive realities of language acquisition. By modeling grammars as information channels, we utilize the Generalized Entropy-Complexity Correspondence, which decomposes the expected Kolmogorov complexity of syntax into semantic entropy $\HH(X)$, syntactic redundancy (Degeneracy, $\HH(Y|X)$), and semantic uncertainty (Ambiguity, $\HH(X|Y)$). We provide a novel resolution to Gold's theorem by defining a learnable class of grammars $(\mathcal{G}_{K})$ constrained by bounds on both Kolmogorov complexity and structural degeneracy. We formally prove that this class possesses finite elasticity, guaranteeing identification in the limit. We then introduce the Dynamic Entropy-Complexity Correspondence, yielding the Complexity Dynamics Equation (CDE), an equation of motion for the learning process. The CDE defines a phase space of adaptation (Plasticity, Efficiency, Robustness) that formally models how a learner navigates trade-offs over time. We establish the NP-completeness of the Linguistic Semantic Inference Problem (LSIP), providing a complexity-theoretic foundation for Universal Grammar. Finally, we provide a concrete empirical validation of the CDE framework by analyzing the training dynamics of a Large Language Model (LLM), demonstrating that its learning trajectory through the adaptive phase space aligns with theoretical predictions.</p>

Abstract: As machine learning systems increasingly train on self-annotated data, they risk reinforcing errors and becoming echo chambers of their own beliefs. We model this phenomenon by introducing a learning-theoretic framework: Online Learning in the Replay Setting. In round $t$, the learner outputs a hypothesis $\hat{h}_t$; the adversary then reveals either the true label $f^\ast(x_t)$ or a replayed label $\hat{h}_i(x_t)$ from an earlier round $i < t$. A mistake is counted only when the true label is shown, yet classical algorithms such as the SOA or the halving algorithm are easily misled by the replayed errors. We introduce the Extended Threshold dimension, $\mathrm{ExThD}(\mathcal{H})$, and prove matching upper and lower bounds that make $\mathrm{ExThD}(\mathcal{H})$ the exact measure of learnability in this model. A closure-based learner makes at most $\mathrm{ExThD}(\mathcal{H})$ mistakes against any adaptive adversary, and no algorithm can perform better. For stochastic adversaries, we prove a similar bound for every intersection-closed class. The replay setting is provably harder than the classical mistake bound setting: some classes have constant Littlestone dimension but arbitrarily large $\mathrm{ExThD}(\mathcal{H})$. Proper learning exhibits an even sharper separation: a class is properly learnable under replay if and only if it is (almost) intersection-closed. Otherwise, every proper learner suffers $Ω(T)$ errors, whereas our improper algorithm still achieves the $\mathrm{ExThD}(\mathcal{H})$ bound. These results give the first tight analysis of learning against replay adversaries, based on new results for closure-type algorithms.

Abstract: This study analyses the effects of hard and soft skills on teachers' innovative capabilities and performance. A random sample of 300 teachers from the Higher Education Institutes of the NCR has been taken, out of which 211 valid replies for the study has been considered. Hard and soft skills positively and considerably affect teachers' innovation capabilities, both directly and indirectly. Furthermore, teachers' innovative abilities improved their performance significantly. The study advocates using organizational learning and hard and soft skills to improve teachers' innovation and performance. This methodology encourages adaptability and innovation to prepare instructors for Education.

Abstract: We introduce a psychologically grounded and artist-informed framework for modeling visual creativity across four domains: Inner, Outer, Imaginative, and Moral Worlds. Drawing on interviews with practicing artists and theories from psychology, we define 12 traits that capture affective, symbolic, cultural, and ethical dimensions of creativity.Using 20k artworks from the SemArt dataset, we annotate images with GPT 4.1 using detailed, theory-aligned prompts, and evaluate the learnability of these traits from CLIP image embeddings. Traits such as Environmental Dialogicity and Redemptive Arc are predicted with high reliability ($R^2 \approx 0.64 - 0.68$), while others like Memory Imprint remain challenging, highlighting the limits of purely visual encoding. Beyond technical metrics, we visualize a "creativity trait-space" and illustrate how it can support interpretable, trait-aware co-creation - e.g., sliding along a Redemptive Arc axis to explore works of adversity and renewal. By linking cultural-aesthetic insights with computational modeling, our work aims not to reduce creativity to numbers, but to offer shared language and interpretable tools for artists, researchers, and AI systems to collaborate meaningfully.

Abstract: This study aims to evaluate the usability level of the Livin' by Mandiri mobile banking application using five dimensions developed by Jakob Nielsen, namely learnability, efficiency, memorability, errors, and user satisfaction. The research method used is a quantitative approach with Partial Least Square-Structural Equation Modeling (PLS-SEM) analysis techniques, and refers to the international standard ISO 9241-11 which focuses on user experience and perception. Data were collected through an online questionnaire distributed to 70 respondents who are active users of the Livin' by Mandiri application. This study emphasizes the importance of understanding the overall user experience in the context of digital banking applications. The results of the analysis show that of the five dimensions analyzed, only learnability has a significant influence on the perception of the application's overall usability. This indicates that the ease of learning and using the application from the beginning is a key factor in determining how well the application is accepted and utilized by users. Meanwhile, the other dimensions, although still relevant in the evaluation context, did not show a significant influence on the perception of usability. These findings reinforce the view that each usability dimension can have different impacts depending on user characteristics and the context in which the app is used.

Abstract: This analysis evaluates the impact of usability components on the Tahu Sumedang public service application. Based on survey data from 68 users and analyzed using PLS-SEM, the results show that memorability, error handling, and learnability significantly influence usability. The model demonstrates strong predictive power with an R² of 84.5%.

Abstract: his study evaluates the usability of the Google Forms application through a framework derived from Nielsen’s heuristics. This framework encompasses five key dimensions: Learnability, Efficiency, Memorability, Errors, and Satisfaction. A quantitative approach was implemented using the Partial Least Squares Structural Equation Modeling (PLS-SEM) method, with data analysis performed using SmartPLS software. One hundred and one respondents participated, completing a questionnaire based on a 5-point Likert scale. The analysis results indicate that Efficiency, Satisfaction, and Memorability significantly influence Usability, with Satisfaction emerging as the most dominant factor. Conversely, Errors and Learnability do not exhibit significant effects. These findings suggest Google Forms is perceived as efficient, satisfying, and relatively easy to remember. However, there is room for improvement in terms of initial learnability and error handling to further enhance the overall user experience.

Abstract: This paper explores the potential of video games as an effective medium for teaching programming concepts to children leveraging gamification principles to enhance engagement. Through the development and testing of an educational game called “Codonia” implementing challenge-based learning and interactive feedback, this research examines how game mechanics support problem-solving and logical thinking. After a qualitative and quantitative study, the results indicate that gamification fosters increased motivation and confidence, suggesting that well-structured educational games can complement or surpass traditional teaching methods. By demonstrating how digital learning environments encourage experimentation and sustained engagement, this study contributes to the broader discourse on innovative educational tools for programming instruction.

Abstract: This contribution focuses on the clausal architecture of Heritage German in theUnited States. In particular, we identify structural properties of German clauseswhich prove to be robustly canonical even though there is inter- and intra-individualvariation on other levels of the linguistic system. With respect to specificclausal features, group comparisons of heritage speakers in both their languagesand with monolingually-raised speakers of German have been conducted. We alsocontribute to current discussions of the heterogeneity of heritage speaker competence and performance and the relevance of individual speaker profiles by qualitative analyses based on data from what we consider Tiny Language Islands, i.e.,acquisition scenarios with the heritage language only spoken within the immediate family. These findings matter because they speak to hypotheses concerninginterfaces between subsystems of our overall linguistic knowledge and, moreover,to discussions on language learnability under conditions of reduced first languageexposure.

Abstract: In this paper, we examine the hypothesis that the interactions recorded in many Recommendation Systems datasets are distributed according to a low-rank distribution, i.e. a mixture of factorizable distributions. Surprisingly, we find that on several popular datasets, a simple non-negative matrix factorization method equals or outperforms more modern methods such as LightGCN, which indicates that the sampling distribution over interactions is indeed low-rank. Furthermore, we mathematically prove that low-rank distributions are learnable with a sparse number \(\widetilde{O}((m+n)r)\) of observations (where m/n and r refer to the number of users/items and the non-negative rank respectively) both in terms of the total variation norm and in terms of the expected recall at k, arguably providing some of the first generalization bounds for recommender systems in the implicit feedback setting. We also provide a modified version of the NMF algorithm which provides further performance improvements compared to the standard NMF baseline on the smaller datasets considered. Finally, we propose the theoretically grounded concept of empirical expected recall as an uncertainty estimate for probabilistic models of the recommendation task, and demonstrate its success in a setting where user-wise abstentions are allowed.

Abstract: We present a new freely available software package `SoftStress' that learns and solves weighted constraint grammars with hidden structure. The package is equipped with a Maximum Entropy Grammar learner that gradually updates the constraint weights based on the probability distribution over all possible hidden representations, and a Linear Programming solver that can check all possible analyses involving hidden representations given a constraint set, without having to generate the factorial typology for the constraint set. We demonstrate the learner and the solver can be used to compare the representational capacity and the learnability of foot-based and grid-based theories of word stress, for a set of attested stress patterns.

Abstract: This study aims to analyze user interaction with the by.U application using the PLS-SEM method. Data was obtained through a questionnaire distributed online and filled out by 65 respondents, but only 52 data were used because 13 of them were invalid. The analysis was conducted using SmartPLS by testing validity, reliability, and hypotheses between variables. The validity test results showed that 17 out of 30 indicators met the loading factor value > 0.70. The reliability test states that 5 out of 6 variables meet the Cronbach's Alpha and Composite Reliability criteria > 0.70. In hypothesis testing, the learnability variable has a significant effect on usability with a p-value of 0.001. This shows that the ease with which users can learn the application has a real impact on the perceived usability of the application. This study concludes that most of the user interaction variables on the by.U application have a significant influence and can be used as a reference for improving application quality.

Abstract: While museums traditionally cater to visual perception, some are adopting novel approaches to aid visitors with impairments. In recent years, robot-based tour guides have become a popular museum attraction. Instead of traditional guides, special-purpose robots assist people and allow for more engaging multimodal touring. While promising, such robotics guides tend to prioritize efficiency and safety over inclusivity, focusing on collision avoidance and movement. At the same time, accessibility challenges magnified by different approaches to learning and disabilities remain underexplored. To tackle this issue, we propose to enhance robotic touring with augmented reality (AR) to overlay additional information onto a visitor’s view. Previous research suggests that AR can improve learning, knowledge retention, and accessibility, particularly when traditional approaches are impractical. Consequently, we integrated AR with a Lindsey robotic guiding system specifically designed to work in an open environment. Next, we tested the hypothesis that AR-enhanced robot guides improve learning and memory retention. A user study with 21 participants found that augmenting the robotic tour guide with AR can improve the learning outcomes of museum visitors.

Abstract: Abstrak. Pemerintah menyediakan layanan publik berbasis digital yang efektif dan efisien. Berupa website resmi pemerintah daerah yang berfungsi sebagai media informasi. Penelitian ini bertujuan untuk menganalisis pengaruh usability terhadap kemudahan penggunaan website Pemerintah Kota Tomohon (tomohon.go.id) menggunakan lima indikator Jakob Nielsen: learnability, efficiency, memorability, error, dan satisfaction. Melalui pendekatan kuantitatif dengan sampel 100 responden, hasil menunjukkan usability berpengaruh signifikan terhadap kemudahan penggunaan dengan korelasi sangat kuat (73,9%). Seluruh indikator usability berada pada kategori baik, dengan learnability dan memorability memperoleh nilai tertinggi (4,47 dan 4,40 dari skala 5). Namun masih terdapat kekurangan seperti kecepatan akses lambat dan halaman yang tidak dapat diakses. Disarankan agar Diskominfo Kota Tomohon meningkatkan pengelolaan website, memperbaiki desain antarmuka, dan menambah fitur interaktif untuk meningkatkan kepuasan pengguna serta mendukung keterbukaan informasi publik dan pelayanan berbasis elektronik.Abstract. The government provides efficient and effective digital-based public services, including official local government websites that function as information media. This research aims to analyze the influence of usability on ease of use of the Tomohon City Government website (tomohon.go.id) using Jakob Nielsen's five indicators: learnability, efficiency, memorability, error, and satisfaction. Through a quantitative approach with a sample of 100 respondents, results show that usability significantly affects ease of use with a very strong correlation (73.9%). All usability indicators fall into the good category, with learnability and memorability obtaining the highest scores (4.47 and 4.40 on a scale of 5). However, shortcomings remain, such as slow access speeds and inaccessible pages. It is recommended that the Communication and Information Department of Tomohon City improve website management, enhance interface design, and add interactive features to increase user satisfaction and support public information disclosure and electronic-based services.

Abstract: We provide a full characterization of the concept classes that are optimistically universally online learnable with $\{0, 1\}$ labels. The notion of optimistically universal online learning was defined in [Hanneke, 2021] in order to understand learnability under minimal assumptions. In this paper, following the philosophy behind that work, we investigate two questions, namely, for every concept class: (1) What are the minimal assumptions on the data process admitting online learnability? (2) Is there a learning algorithm which succeeds under every data process satisfying the minimal assumptions? Such an algorithm is said to be optimistically universal for the given concept class. We resolve both of these questions for all concept classes, and moreover, as part of our solution, we design general learning algorithms for each case. Finally, we extend these algorithms and results to the agnostic case, showing an equivalence between the minimal assumptions on the data process for learnability in the agnostic and realizable cases, for every concept class, as well as the equivalence of optimistically universal learnability.

Abstract: This study analyses the effects of hard and soft skills on teachers' innovative capabilities and performance. A random sample of 300 teachers from the Higher Education Institutes of the NCR has been taken, out of which 211 valid replies for the study has been considered. Hard and soft skills positively and considerably affect teachers' innovation capabilities, both directly and indirectly. Furthermore, teachers' innovative abilities improved their performance significantly. The study advocates using organizational learning and hard and soft skills to improve teachers' innovation and performance. This methodology encourages adaptability and innovation to prepare instructors for Education.