Search

This study evaluates fine-tuning strategies for text classification using the DistilBERT model, specifically the distilbert-base-uncased-finetuned-sst-2-english variant. Through structured experiments, we examine the influence of hyperparameters such as learning rate, batch size, and epochs on accuracy, F1-score, and loss. Polynomial regression analyses capture foundational and incremental impacts of these hyperparameters, focusing on fine-tuning adjustments relative to a baseline model. Results reveal variability in metrics due to hyperparameter configurations, showing trade-offs among performance metrics. For example, a higher learning rate reduces loss in relative analysis (p=0.027) but challenges accuracy improvements. Meanwhile, batch size significantly impacts accuracy and F1-score in absolute regression (p=0.028 and p=0.005) but has limited influence on loss optimization (p=0.170). The interaction between epochs and batch size maximizes F1-score (p=0.001), underscoring the importance of hyperparameter interplay. These findings highlight the need for fine-tuning strategies addressing non-linear hyperparameter interactions to balance performance across metrics. Such variability and metric trade-offs are relevant for tasks beyond text classification, including NLP and computer vision. This analysis informs fine-tuning strategies for large language models and promotes adaptive designs for broader model applicability.

Using the hybrid formalism with manifest $N=1$ $d=4$ spacetime supersymmetry, we construct the quadratic term in the heterotic superstring field theory action. As in open superstring field theory using the hybrid formalism, the heterotic string field theory action is constructed with three string fields and the massless sector describes $N=1$ $d=10$ supergravity in terms of $N=1$ $d=4$ superfields., Comment: 10 pages, 0 figures

Since classical machine learning has become a powerful tool for developing data-driven algorithms, quantum machine learning is expected to similarly impact the development of quantum algorithms. The literature reflects a mutually beneficial relationship between machine learning and quantum computing, where progress in one field frequently drives improvements in the other. Motivated by the fertile connection between machine learning and quantum computing enabled by parameterized quantum circuits, we use a resource-efficient and scalable Single-Qubit Quantum Neural Network (SQQNN) for both regression and classification tasks. The SQQNN leverages parameterized single-qubit unitary operators and quantum measurements to achieve efficient learning. To train the model, we use gradient descent for regression tasks. For classification, we introduce a novel training method inspired by the Taylor series, which can efficiently find a global minimum in a single step. This approach significantly accelerates training compared to iterative methods. Evaluated across various applications, the SQQNN exhibits virtually error-free and strong performance in regression and classification tasks, including the MNIST dataset. These results demonstrate the versatility, scalability, and suitability of the SQQNN for deployment on near-term quantum devices., Comment: 21 pages, 7 figures, 6 tables

We explicitly compute the effective action from Open Superstring Field Theory in the hybrid formalism to quartic order in the $\alpha'\rightarrow 0$ limit, and show that it reproduces ten-dimensional Super Yang-Mills in terms of four-dimensional superfields. We also show that in this limit the gauge transformations coincide with SYM to all orders, which means that the effective action should reproduce SYM to all orders., Comment: 10 pages

Many quantum technologies, including quantum computers, quantum heat engines, and quantum sensors, rely on operating conditions in the subkelvin regime. It is therefore desirable to develop practical tools and methods for the precise control of the temperature in nanoscale quantum systems. Here, we present a proposal for fast, accurate, and local temperature control using qubits, which regulate the flow of heat between a quantum system and its thermal environment. The qubits are kept in a thermal state with a temperature that is controlled in an interplay between work done on the qubits by changing their energy splittings and the flow of heat between the qubits and the environment. Using only a few qubits, it is possible to control the thermal environment of another quantum system, which can be heated or cooled by the qubits. As an example, we show how a quantum system at subkelvin temperatures can be significantly and accurately cooled on a nanosecond timescale. Our proposal can potentially be realized with superconducting flux qubits, charge qubits, or spin qubits, which can now be fabricated and manipulated with exquisite control., Comment: 12 pages, 7 figures

The statistics of transmitted photons in microwave cavities play a foundational role in microwave quantum optics and its technological applications. By utilizing quantum mechanical phase-space methods, we here develop a general theory of the photon counting statistics in Gaussian bosonic networks consisting of driven cavities with beamsplitter interactions and two-mode-squeezing. The dynamics of the network can be captured by a Lyapunov equation for the covariance matrix of the cavity fields, which generalizes to a Riccati equation, when counting fields are included. By solving the Riccati equation, we obtain the statistics of emitted and absorbed photons as well as the time-dependent correlations encoded in waiting time distributions and second-order coherence functions. To illustrate our theoretical framework, we first apply it to a simple linear network consisting of two coupled cavities, for which we evaluate the photon cross-correlations and discuss connections between the photon emission statistics and the entanglement between the cavities. We then consider a bosonic circulator consisting of three coupled cavities, for which we investigate how a synthetic flux may affect the direction of the photon flow, similarly to recent experiments. Our general framework paves the way for systematic investigations of the photon counting statistics in Gaussian bosonic networks., Comment: 37 pages, 7 figures

Determining the carrying capacity of urban tourism destinations and public spaces is essential for sustainable management. This paper presents an online tool that calculates pedestrian carrying capacities for user-defined areas based on OpenStreetMap (OSM) data. The tool considers physical, real, and effective carrying capacities by incorporating parameters such as area per pedestrian, rotation factor, corrective factors, and management capacity. The carrying capacity calculator aids in balancing environmental, economic, social, and experiential factors to prevent overcrowding and preserve the quality of life for residents and visitors. This tool is particularly useful for tourism destination management, urban planning, and event management, ensuring positive visitor experiences and sustainable infrastructure development. We detail the implementation of the calculator, its underlying algorithm, and its application to the Santa Maria Maior parish in Lisbon, highlighting its effectiveness in managing urban tourism and public spaces.

The Hiperwalk package is designed to facilitate the simulation of quantum walks using heterogeneous high-performance computing, taking advantage of the parallel processing power of diverse processors such as CPUs, GPUs, and acceleration cards. This package enables the simulation of both the continuous-time and discrete-time quantum walk models, effectively modeling the behavior of quantum systems on large graphs. Hiperwalk features a user-friendly Python package frontend with comprehensive documentation, as well as a high-performance C-based inner core that leverages parallel computing for efficient linear algebra calculations. This versatile tool empowers researchers to better understand quantum walk behavior, optimize implementation, and explore a wide range of potential applications, including spatial search algorithms., Comment: 16 pages, 6 figures

We introduce a two parameter family of string field theory vertices, which we refer to as hyperbolic Kaku vertices. It is defined in terms of hyperbolic metrics on the Riemann surface, but the geometry is allowed to depend on inputs of the states. The vertices are defined for both open and closed strings. In either case, the family contains the hyperbolic vertices. Then we show that the open string lightcone vertex is obtained as the flat limit of the hyperbolic Kaku vertices. The open string Kaku vertices, which interpolate between the Witten vertex and the open string lightcone vertex, is also obtained as a flat limit. We use the same limit on the case of closed strings to define the closed string Kaku vertices: a one parameter family of vertices that interpolates between the polyhedral vertices - which are covariant, but not cubic - and the closed string lightcone vertex - which is cubic, but not Lorentz covariant., Comment: 27 pages, 20 figures

As with most other data domains, EEG data analysis relies on rich domain-specific preprocessing. Beyond such preprocessing, machine learners would hope to deal with such data as with any other time series data. For EEG classification many models have been developed with layer types and architectures we typically do not see in time series classification. Furthermore, typically separate models for each individual subject are learned, not one model for all of them. In this paper, we systematically study the differences between EEG classification models and generic time series classification models. We describe three different model setups to deal with EEG data from different subjects, subject-specific models (most EEG literature), subject-agnostic models and subject-conditional models. In experiments on three datasets, we demonstrate that off-the-shelf time series classification models trained per subject perform close to EEG classification models, but that do not quite reach the performance of domain-specific modeling. Additionally, we combine time-series models with subject embeddings to train one joint subject-conditional classifier on all subjects. The resulting models are competitive with dedicated EEG models in 2 out of 3 datasets, even outperforming all EEG methods on one of them.

In this paper, we analyze the dynamics of quantum walks on a graph structure resulting from the integration of a main connected graph $G$ and a secondary connected graph $G'$. This composite graph is formed by a disjoint union of $G$ and $G'$, followed by the contraction of a selected pair of vertices creating a cut vertex $v^*$ and leading to a unique form of geometric perturbation. Our study focuses on instances where $G$ is a complete graph $K_N$ and $G'$ is a star graph $S_m$. The core of our analysis lies in exploring the impact of this geometric perturbation on the success probability of quantum walk-based search algorithms, particularly in an oracle-free context. Despite initial findings suggesting a low probability of locating the perturbed vertex $v^*$, we demonstrate that introducing a phase reversal to the system significantly enhances the success rate. Our results reveal that with an optimal running time and specific parameter conditions, the success probability can be substantially increased. The paper is structured to first define the theoretical framework, followed by the presentation of our main results, detailed proofs, and concluding with a summary of our findings and potential future research directions.