共找到 20 条结果
This study compares machine learning and deep learning approaches for cyberbullying detection in Indonesian-language Instagram comments. Using a balanced dataset of 650 comments labeled as Bullying and Non-Bullying, the study evaluates Naive Bayes, Logistic Regression, and Support Vector Machine with TF-IDF features, as well as BiLSTM and BiLSTM with Bahdanau Attention. A preprocessing pipeline tailored to informal Indonesian text is applied, including slang normalization, stopword removal, and stemming. The results show that Logistic Regression performs best among the machine learning models, while BiLSTM with Attention achieves the strongest overall deep learning performance. The findings highlight the value of domain-specific preprocessing and show that although deep learning captures contextual patterns more effectively, machine learning remains a competitive option for resource-constrained deployments.
The two Terrestrial Planet Finder (TPF) missions aim to perform spectroscopy on extrasolar Earths; TPF-C will operate in visible light, and TPF-I will operate in the mid-infrared. Extrasolar Earths are assumed to be roughly 26 magnitude in V band, roughly 0.3 microJy in the mid-IR, and located as close as roughly 30 milliarcseconds from a reasonable set of target stars, demanding high sensitivity, angular resolution and dynamic range to study. With capabilities matched to this task, the TPF missions could easily undertake a broad range of further scientific investigations. This document discusses the potential of TPF for general astrophysics and comparative planetology beyond its base mission, focusing on science obtainable with no or minimal modifications to the mission design, but also exploring possible modifications to TPF with high scientific merit and no impact on the basic search for extrasolar Earth analogs. It addresses both TPF-C and TPF-I, but emphasizes TPF-C, because its launch is planned for 2015, while TPF-I s nominal launch date is in 2019. This document does not attempt to describe all of the astrophysics TPF will be capable of, only to present some highlights of a
While machine learning has revolutionized many fields such as natural language processing (NLP) and computer vision, its impact on time-series forecasting is still widely disputed, especially in the finance domain. This paper compares forecasting performance on U.S. Treasury yield curve data across econometrics/time-series analysis, classical machine learning, and deep learning methods, using daily data over 47 years. The Treasury yield curve is important because it is widely used by every participant in the bond markets, which are larger than equity markets. We examine a variety of methods that have not been tested on yield curve forecasting, especially deep learning algorithms. The algorithms include the Autoregressive Integrated Moving Average (ARIMA) model and its extensions, naive benchmarks, ensemble methods, Recurrent Neural Networks (RNNs), and multiple transformers built for forecasting. ARIMA and naive econometric models outperform other models overall, except in one time block. Of the machine learning methods, TimeGPT, LGBM and RNNs perform the best. Furthermore, the paper explores whether stationary or nonstationary data are more appropriate as input to deep learning mo
Background: Large language models have demonstrated strong performance on general medical examinations, but subspecialty clinical reasoning remains challenging due to rapidly evolving guidelines and nuanced evidence hierarchies. Methods: We evaluated January Mirror, an evidence-grounded clinical reasoning system, against frontier LLMs (GPT-5, GPT-5.2, Gemini-3-Pro) on a 120-question endocrinology board-style examination. Mirror integrates a curated endocrinology and cardiometabolic evidence corpus with a structured reasoning architecture to generate evidence-linked outputs. Mirror operated under a closed-evidence constraint without external retrieval. Comparator LLMs had real-time web access to guidelines and primary literature. Results: Mirror achieved 87.5% accuracy (105/120; 95% CI: 80.4-92.3%), exceeding a human reference of 62.3% and frontier LLMs including GPT-5.2 (74.6%), GPT-5 (74.0%), and Gemini-3-Pro (69.8%). On the 30 most difficult questions (human accuracy less than 50%), Mirror achieved 76.7% accuracy. Top-2 accuracy was 92.5% for Mirror versus 85.25% for GPT-5.2. Conclusions: Mirror provided evidence traceability: 74.2% of outputs cited at least one guideline-tier so
Large language models (LLMs) achieve strong performance across many natural language processing tasks, yet their decision processes remain difficult to interpret. This lack of transparency creates challenges for trust, debugging, and deployment in real-world systems. This paper presents an applied comparative study of three explainability techniques: Integrated Gradients, Attention Rollout, and SHAP, on a fine-tuned DistilBERT model for SST-2 sentiment classification. Rather than proposing new methods, the focus is on evaluating the practical behavior of existing approaches under a consistent and reproducible setup. The results show that gradient-based attribution provides more stable and intuitive explanations, while attention-based methods are computationally efficient but less aligned with prediction-relevant features. Model-agnostic approaches offer flexibility but introduce higher computational cost and variability. This work highlights key trade-offs between explainability methods and emphasizes their role as diagnostic tools rather than definitive explanations. The findings provide practical insights for researchers and engineers working with transformer-based NLP systems. T
This empirical study investigates the impact of the Hofstede cultural dimensions (HCD) on the Global Innovation Index (GII) scores in four different years (2007, 2009, 2019 and 2021) to compare the impacts during the pre- and post-crisis (financial and COVID-19) period by employing ordinary least square (OLS) and robust least square (Robust) analyses. The purpose of this study is to identify the impact of cultural factors on the innovation development for different income groups during the pre- and post-crisis period. We found that, in general, the same cultural properties were required for countries to enhance innovation inputs and outputs regardless of pre- and post-crisis periods and time variances. The significant cultural factors (driving forces) of the innovation performance do not change over time. However, our empirical results revealed that not the crisis itself but the income group (either developed or developing) is the factor that influences the relationship between cultural properties and innovation. It is also worth noting that cultural properties have lost much of their impact on innovation, particularly in developing countries, during recent periods. It is highly li
Small open-source medical large language models (LLMs) offer promising opportunities for low-resource deployment and broader accessibility. However, their evaluation is often limited to accuracy on medical multiple choice question (MCQ) benchmarks, and lacks evaluation of consistency, robustness, or reasoning behavior. We use MCQ coupled to human evaluation and clinical review to assess six small open-source medical LLMs (HuatuoGPT-o1 (Chen 2024), Diabetica-7B, Diabetica-o1 (Wei 2024), Meditron3-8B (Sallinen2025), MedFound-7B (Liu 2025), and ClinicaGPT-base-zh (Wang 2023)) in pediatric endocrinology. In deterministic settings, we examine the effect of prompt variation on models' output and self-assessment bias. In stochastic settings, we evaluate output variability and investigate the relationship between consistency and correctness. HuatuoGPT-o1-8B achieved the highest performance. The results show that high consistency across the model response is not an indicator of correctness, although HuatuoGPT-o1-8B showed the highest consistency rate. When tasked with selecting correct reasoning, both HuatuoGPT-o1-8B and Diabetica-o1 exhibit self-assessment bias and dependency on the order
We present Six Llamas, a comparative study examining whether large language models fine-tuned on distinct religious corpora encode systematically different patterns of ethical reasoning. Six variants of Meta-Llama-3.1-8B are constructed: one unmodified control and five LoRA-adapted models trained exclusively on the sacred and theological texts of Christianity, Islam, Judaism, Hinduism, or Buddhism. All six models are probed with an identical battery of 17 standardized ethical prompts spanning moral dilemmas, game-theoretic scenarios, public policy questions, and moral-psychological self-assessments. To assess robustness and reproducibility, we implement a multi-temperature sampling design spanning ten temperature settings. We compute response consistency metrics, pairwise inter-model agreement rates, temperature sensitivity coefficients across four prompt domains, and run-to-run stability analyses. Findings show that LoRA-adapted models produce ethical reasoning patterns that are (a) systematically differentiated from the base model, (b) consistent with the moral logics of their training traditions, (c) structured along interpretable dimensions in moral-philosophical space, (d) cor
We investigate whether black holes can persist through the bounce with a minimal scale factor in a non-singular cosmology, whereby black holes from a previous contracting phase survive into the current expanding one. We do so by studying a generalized McVittie spacetime which embeds a spherically symmetric black hole in a positive spatial curvature bouncing FLRW cosmological background within the modified theory of teleparallel new general relativity. There are no further assumptions on the spacetime (e.g., on the form of the scale factor) initially, and the local evolution is derived from the field equations of the theory, utilizing a perturbative scheme which is valid ``near the bounce". To leading order we obtain a simple bounce solution similar to that in general relativity for a closed FLRW model with a positive cosmological constant, but in which the curvature term in the Friedmann equation is re-normalized within new general relativity. Qualitatively the minimum of the bounce at $t=0$ changes, but near the bounce the evolution remains symmetric. The central inhomogeneity evolves at higher perturbative orders, where the details depend on the arbitrary constants of the perturb
In anticipation of the completion of the High-Luminosity Large Hadron Collider (HL-LHC) programme by the end of 2041, CERN is preparing to launch a new major facility in the mid-2040s. According to the 2020 update of the European Strategy for Particle Physics (ESPP), the highest-priority next collider is an electron-positron Higgs factory, followed in the longer term by a hadron-hadron collider at the highest achievable energy. The CERN directorate established a Future Colliders Comparative Evaluation working group in June 2023. This group brings together project leaders and domain experts to conduct a consistent evaluation of the Future Circular Collider (FCC) and alternative scenarios based on shared assumptions and standardized criteria. This report presents a comparative evaluation of proposed future collider projects submitted as input for the Update of the European Strategy for Particle Physics. These proposals are compared considering main performance parameters, environmental impact and sustainability, technical maturity, cost of construction and operation, required human resources, and realistic implementation timelines. An overview of the international collider projects w
Object detection in remotely sensed satellite pictures is fundamental in many fields such as biophysical, and environmental monitoring. While deep learning algorithms are constantly evolving, they have been mostly implemented and tested on popular ground-based taken photos. This paper critically evaluates and compares a suite of advanced object detection algorithms customized for the task of identifying aircraft within satellite imagery. Using the large HRPlanesV2 dataset, together with a rigorous validation with the GDIT dataset, this research encompasses an array of methodologies including YOLO versions 5 and 8, Faster RCNN, CenterNet, RetinaNet, RTMDet, and DETR, all trained from scratch. This exhaustive training and validation study reveal YOLOv5 as the preeminent model for the specific case of identifying airplanes from remote sensing data, showcasing high precision and adaptability across diverse imaging conditions. This research highlight the nuanced performance landscapes of these algorithms, with YOLOv5 emerging as a robust solution for aerial object detection, underlining its importance through superior mean average precision, Recall, and Intersection over Union scores. T
Satellite dynamics and tracking remain important challenges in the context of space exploration and communication systems. Accurate state estimation is essential to maintain reliable orbital motion and system performance. This paper presents a mathematical framework for satellite state estimation based on a linearized model described by radial and angular states. The model incorporates two types of measurement noise corresponding to range and scaled angular deviations, which are assumed to be mutually independent with known covariance structures. The estimation problem is formulated using the Kalman filter, together with the associated Algebraic Riccati Equation (ARE), leading to both time-varying and steady-state solutions. In addition, a micro-Kalman filter ($μ$KF) formulation is considered and compared with the classical Kalman filter, as well as with the extended Kalman filter (EKF), unscented Kalman filter (UKF), and an adaptive Kalman filter under a unified simulation setup. The results demonstrate that the proposed $μ$KF achieves estimation performance nearly identical to that of the classical Kalman filter and its variants, with small and bounded estimation errors. The mean
The zeroth-order general Randić index $R^{0}_{a+1}$ of an $n$-vertices oriented graph $D$ is equal to the sum of $(d^{+}_{u_i})^{a}+(d^{-}_{u_j})^{a}$ over all arcs $u_iu_j$ of $D$, where we denote by $d^{+}_{u_i}$ the out-degree of the vertex $u_i$ and $d^{-}_{u_j}$ the in-degree of the vertex $u_j$, $a$ is an arbitrary real number. In the paper, we determine the orientations of cacti with the maximum value of the zeroth-order general Randić index for $a\geq 1$.
We construct a model unifying general relativity and quantum mechanics in a broader structure of noncommutative geometry. The geometry in question is that of a transformation groupoid given by the action of a finite group G on a space E. We define the algebra of smooth complex valued functions on the groupoid, with convolution as multiplication, in terms of which the groupoid geometry is developed. Owing to the fact that the group G is finite the model can be computed in full details. We show that by suitable averaging of noncommutative geometric quantities one recovers the standard space-time geometry. The quantum sector of the model is explored in terms of the regular representation of the groupoid algebra, and its correspondence with the standard quantum mechanics is established.
In this paper, by proposing a generalized $specific~volume$, we restudy the $P-V$ criticality of charged AdS black holes in the extended phase space. The results show that most of the previous conclusions can be generalized without change, but the ratio $\tildeρ_c$ should be $3 \tildeα/16$ in general case. Further research on the thermodynamical phase transition of black hole leads us to a natural interpretation of our assumption, and more black hole properties can be generalized. Finally, we study the number density for charged AdS black hole in higher dimensions, the results show the necessity of our assumption.
**PLEASE FIND THE FULL EXTENDED ARTICLE "From OTFS to AFDM: A Comparative Study of Next-Generation Waveforms for ISAC in Doubly-Dispersive Channels" (Accepted for publication at the IEEE Signal Processing Magazine - Special Issue on Signal Processing for the Integrated Sensing and Communications Revolution)** This white paper aims to briefly describe a proposed article that will provide a thorough comparative study of waveforms designed to exploit the features of doubly-dispersive channels arising in heterogeneous high-mobility scenarios as expected in the beyond fifth generation (B5G) and sixth generation (6G), in relation to their suitability to integrated sensing and communications (ISAC) systems. In particular, the full article will compare the well-established delay-Doppler domain-based orthognal time frequency space (OTFS) and the recently proposed chirp domain-based affine frequency division multiplexing (AFDM) waveforms. Both these waveforms are designed based on a full delay- Doppler representation of the time variant (TV) multipath channel, yielding not only robustness and orthogonality of information symbols in high-mobility scenarios, but also a beneficial implication f
A systematic study of the Weyl-type / Yang-Mills-type action possessing local conformal invariance and quadratic curvature is undertaken. The dynamical breaking of this conformal invariance / scale invariance induces general relativity (GR) as an effective long distance limit of the theory. We prove that the corresponding field equations of the theory have the linearly rising potential, which naturally possesses asymptotic freedom and color confinement properties of quantum chromodynamics (QCD). Solutions to the neutrino mass and dark energy problems come as free gifts of this formulation. This approach provides a strong gravity basis for the unification of quantum Yang-Mills theory (QYMT) with Einstein GR.
This paper surveys aspects of the convergence and degeneration of Riemannian metrics on a given manifold M - the Cheeger-Gromov theory - and extensions thereof to Ricci curvature in place of full curvature. This theory is then applied to study a collection of different issues in mathematical aapects of General Relativity.
In this work we investigate the structure of white dwarfs using the Tolman-Oppenheimer-Volkoff equations and compare our results with those obtained from Newtonian equations of gravitation in order to put in evidence the importance of General Relativity (GR) for the structure of such stars. We consider in this work for the matter inside white dwarfs two equations of state, frequently found in the literature, namely, the Chandrasekhar and Salpeter equations of state. We find that using Newtonian equilibrium equations, the radii of massive white dwarfs ($M>1.3M_{\odot}$) are overestimated in comparison with GR outcomes. For a mass of $1.415M_{\odot}$ the white dwarf radius predicted by GR is about 33\% smaller than the Newtonian one. Hence, in this case, for the surface gravity the difference between the general relativistic and Newtonian outcomes is about 65\%. We depict the general relativistic mass-radius diagrams as $M/M_{\odot}=R/(a+bR+cR^2+dR^3+kR^4)$, where $a$, $b$, $c$ and $d$ are parameters obtained from a fitting procedure of the numerical results and $k=(2.08\times 10^{-6}R_{\odot})^{-1}$, being $R_{\odot}$ the radius of the Sun in km. Lastly, we point out that GR play
We attempt to define what is necessary to construct an Artificial Scientist, explore and evaluate several approaches to artificial general intelligence (AGI) which may facilitate this, conclude that a unified or hybrid approach is necessary and explore two theories that satisfy this requirement to some degree.