We present an exactly soluble electron trajectory that permits an analysis of the soft (deep infrared) radiation emitted, the existence of which has been experimentally observed during beta decay via lowest order inner bremsstrahlung. Our treatment also predicts the time evolution and temperature of the emission, and possibly the spectrum, by analogy with the closely related phenomenon of the dynamic Casimir effect.

We propose a unified model of dark energy and inflation through the Markov-Mukhanov modification of the Einstein-Hilbert action, where the matter sector is coupled to gravity via a scalar coupling function depending only on the energy density of the matter content. We assume that the coupling function encodes the UV corrections to the standard model of cosmology and we determine the form of the coupling that allows for the dark energy component to be dynamical and act as the inflaton field in the early universe. Interestingly we show that our model, in order to account for inflation, prefers a dark energy equation of state with $w$ close but not equal to $-1$ in agreement with the latest DESI data.

We report on the observation of thermal photons from an accelerated electron via examination of radiative beta decay of free neutrons measured by the RDK II collaboration. The emitted photon spectrum is shown to corroborate a thermal distribution consistent with the dynamical Casimir effect. Supported by a robust chi-squared statistic, we find the photons reside in a one-dimensional Planck spectrum with a temperature predicted by the moving mirror model.

Thermal radiation from a moving point charge is found. The calculation is entirely from a classical point of view, but is shown to have an immediate connection to quantum field theory.

Erasing a black hole leaves spacetime flat, so light passing through the region before any star forms and after black hole's evaporation shows no time delay, just like a flying mirror that returns to its initial starting point. Quantum radiation from a round-trip flying mirror has not been solved despite the model's mathematical simplicity and physical clarity. Here, we solve the particle creation from worldlines that asymptotically start and stop at the same spot, resulting in interesting spectra and symmetries, including the time dependence of thermal radiance associated with Bose-Einstein and Fermi-Dirac Bogolubov coefficients. Fourier analysis, intrinsically linked to the Bogolubov mechanism, shows that a thermal Bogolubov distribution does not describe the spin statistics of the quantum field.

We investigate how entanglement entropy can drive particle creation, deriving explicit relations between entropy and the radiated particle spectrum, the total number of particles, and the total energy. Particle production is computed for scenarios that include accelerated motion, black hole evaporation, and beta decay, validating against known results while also extending them. We focus primarily on the low-entropy limit (analogous to non-relativistic motion), but also examine cases of significant particle production arising from harmonic cycles. The results establish an explicit operational link between information flow and matter creation, providing a concrete demonstration of 'it from bit'.

As the use of Large Language Models (LLMs) becomes more widespread, understanding their self-evaluation of confidence in generated responses becomes increasingly important as it is integral to the reliability of the output of these models. We introduce the concept of Confidence-Probability Alignment, that connects an LLM's internal confidence, quantified by token probabilities, to the confidence conveyed in the model's response when explicitly asked about its certainty. Using various datasets and prompting techniques that encourage model introspection, we probe the alignment between models' internal and expressed confidence. These techniques encompass using structured evaluation scales to rate confidence, including answer options when prompting, and eliciting the model's confidence level for outputs it does not recognize as its own. Notably, among the models analyzed, OpenAI's GPT-4 showed the strongest confidence-probability alignment, with an average Spearman's $\hat{\rho}$ of 0.42, across a wide range of tasks. Our work contributes to the ongoing efforts to facilitate risk assessment in the application of LLMs and to further our understanding of model trustworthiness.

We demonstrate that if the universe started as a vacuum fluctuation rather than from a singular Big Bang state, the universe must have a late-time cosmic acceleration. This is required by a ``cosmological sum rule'' derived using the Schwarzian form of the Friedmann equations. We discuss possible connections to conformal and M\"obius transformations, and also compute that the best fit present cosmic data is consistent with the necessary crossing of the Schwarzian through zero having occurred (while it would not yet have happened in a $\Lambda$CDM cosmology).

We examine the extreme situation of radiation from an electron that is asymptotically accelerated to the speed of light, resulting in finite emission energy. The analytic solution explicitly demonstrates the difference between radiation power loss and kinetic power loss (null).

Modeling the collapse of an extreme Reissner-Nordstr\"om (ERN) black hole by solving the corresponding moving mirror model for the trajectory that asymptotically approaches uniform acceleration, we obtain the non-zero beta coefficients for all times. Finite energy is emitted, the radiation spectra is non-thermal (non-steady / not Planck), soft particles characterize the evaporation, and particle production at ultra-late times is damped. Entanglement entropy diverges with no Page curve turn-over, demonstrating non-thermal information loss. The radiation obeys time-reversal symmetry.

A confined, slow-moving, accelerating electron is shown to emit thermal radiation. Since laboratories face spatial constraints when dealing with rectilinear motion, focusing on a finite total travel distance combines the benefits of simple theoretical analysis with prospects for table-top experimentation. We demonstrate an accelerated moving charge along an asymptotically static worldline with fixed transit distance and slow maximum speed, emitting self-consistent analytic power, spectra, and energy. The classical radiation is Planck distributed with an associated acceleration temperature. This is the first fully parametrized, spectrum-solved, finite-distance worldline.

In this article, we derive and discuss the properties of the symplectic group Sp(2), which arises in Hamiltonian dynamics and ray optics. We show that a symplectic matrix can be written as the product of a symmetric dilation matrix and a rotation matrix, in either order. A symplectic matrix can be written as the exponential of a generating matrix, and there is a one-to-one relation between the coefficients of the symplectic and generating matrices. We also discuss the adjoint and Schmidt decompositions of a symplectic matrix, and the product of two symplectic matrices. The results of this article have applications in many subfields of physics.

A black mirror is an accelerated boundary that produces particles in an exact correspondence to an evaporating black hole. We investigate the spectral dynamics of the particle creation during the formation process.

This work aims to build a multilingual text-to-speech (TTS) synthesis system for ten lower-resourced Turkic languages: Azerbaijani, Bashkir, Kazakh, Kyrgyz, Sakha, Tatar, Turkish, Turkmen, Uyghur, and Uzbek. We specifically target the zero-shot learning scenario, where a TTS model trained using the data of one language is applied to synthesise speech for other, unseen languages. An end-to-end TTS system based on the Tacotron 2 architecture was trained using only the available data of the Kazakh language. To generate speech for the other Turkic languages, we first mapped the letters of the Turkic alphabets onto the symbols of the International Phonetic Alphabet (IPA), which were then converted to the Kazakh alphabet letters. To demonstrate the feasibility of the proposed approach, we evaluated the multilingual Turkic TTS model subjectively and obtained promising results. To enable replication of the experiments, we make our code and dataset publicly available in our GitHub repository.

The Schwarzschild-de Sitter (SdS) metric is the simplest spacetime solution in general relativity with both a black hole event horizon and a cosmological event horizon. Since the Schwarzschild metric is the most simple solution of Einstein's equations with spherical symmetry and the de Sitter metric is the most simple solution of Einstein's equations with a positive cosmological constant, the combination in the SdS metric defines an appropriate background geometry for semi-classical investigation of Hawking radiation with respect to past and future horizons. Generally, the black hole temperature is larger than that of the cosmological horizon, so there is heat flow from the smaller black hole horizon to the larger cosmological horizon, despite questions concerning the definition of the relative temperature of the black hole without a measurement by an observer sitting in an asymptotically flat spacetime. Here we investigate the accelerating boundary correspondence (ABC) of the radiation in SdS spacetime without such a problem. We have solved for the boundary dynamics, energy flux and asymptotic particle spectrum. The distribution of particles is globally non-thermal while asymptotically the radiation reaches equilibrium.

This study evaluates the trade-offs between convolutional and transformer-based architectures on both medical and general-purpose image classification benchmarks. We use ResNet-18 as our baseline and introduce a fine-tuning strategy applied to four Vision Transformer variants (Tiny, Small, Base, Large) on DermatologyMNIST and TinyImageNet. Our goal is to reduce inference latency and model complexity with acceptable accuracy degradation. Through systematic hyperparameter variations, we demonstrate that appropriately fine-tuned Vision Transformers can match or exceed the baseline's performance, achieve faster inference, and operate with fewer parameters, highlighting their viability for deployment in resource-constrained environments.