We introduce a method based on semidefinite programming that produces rigorous two-sided bounds on ground state energy densities and correlation functions of translation-invariant classical spin models on infinite lattices. In this method, the challenge of non-convex optimization on an infinite lattice is replaced with a hierarchy of finite-size convex optimizations arising from positivity conditions that any probability distribution over spin configurations must satisfy. This adapts the Lasserre hierarchy in the theory of polynomial optimization to the context of frustrated magnetism, and we prove convergence of this hierarchy in the thermodynamic limit. Our method subsumes the Luttinger--Tisza method and further applies to non-quadratic Hamiltonians and non-Bravais lattices, thus addressing limitations of prior analytical methods. We apply the method to various two-dimensional frustrated spin models, where it brackets the energy densities and observables accurately across large parameter ranges with typical run times of seconds per parameter point.

Gravitational wave sources act as absolute distance indicators, making them powerful probes of the present-day expansion rate of the Universe, $H_0$. The cross-correlation method combines gravitational wave events with galaxy catalogues to constrain cosmological parameters through their shared large-scale structure. In this work, we investigate how key methodological choices -- including covariance treatment, bias parametrisation for galaxies and gravitational wave events, and distance and redshift binning width -- affect the inferred value of $H_0$. We also study catalogue incompleteness, showing that selection effects can be incorporated directly into the theoretical prediction, without the need to model the missing population explicitly, a key advantage over the standard galaxy catalogue approach. Our results indicate that, with appropriate modelling choices and a sufficiently large sample of precise gravitational wave events, the systematic biases considered here can be effectively mitigated, highlighting the potential of the cross-correlation method for future dark siren precision cosmology.

JWST has enabled the detection of galaxies in the earliest stages of cosmic history. We compare the ultraviolet luminosity functions (UVLFs) at redshifts $z=7-15$ predicted by the new cosmological hydrodynamics simulations, COLIBRE with observations, including those from JWST. The UV luminosities of COLIBRE galaxies are derived using the radiative transfer code SKIRT, which tracks stellar emission and its processing through the multi-phase interstellar medium and dust distribution predicted by COLIBRE. We find that although COLIBRE is consistent with the observed evolution of the stellar mass function up to $z=12$, its dust-attenuated UVLFs fall systematically below the observations at the bright end: at the number density of $10^{-6}\,\mathrm{Mpc^{-3}\,mag^{-1}}$, the brightest galaxies are underluminous by $\approx 1\,\rm mag$ at $z=7$, increasing to $\approx 2.5\,\rm mag$ at $z=15$. Accounting for observational uncertainties brings the COLIBRE UVLFs closer to the observational data, but does not fully resolve the discrepancy. Ignoring dust attenuation allows COLIBRE to produce sufficiently bright galaxies at $7\lesssim z \lesssim 12$, while at $z=15$, COLIBRE still underpredicts the luminosities of the brightest galaxies, indicating the need for additional physical mechanisms to boost the UV luminosities at the earliest cosmic epochs, such as a ''top-heavy'' stellar initial mass function. We fit the COLIBRE UVLFs with Schechter functions and calculate the evolution of the best-fit parameters. We find that the galaxy number density decreases, the characteristic luminosity becomes fainter and the faint-end slope becomes steeper towards higher redshifts. The UV luminosity density decreases by a factor of $\approx 300$ from $z = 7$ to $z = 15$.

A central question in galactic dynamics and galactic archeology is: how much do the orbits of stars redistribute (migrate) after birth? We use the FIRE-2 cosmological zoom-in simulations of 11 Milky Way-mass galaxies to quantify the change in the orbital specific angular momentum, j_phi, orbital radius, R_orbit, and azimuthal velocity, v_phi, of stars from birth to today. We examine the dependences on stellar age, present-day R_orbit, and birth R_orbit, characterizing both the median (net) change, Delta R_orbit, and its scatter, sigma(Delta R_orbit). We comprehensively compare five ways of measuring orbital radius; we find generally consistent trends, but only when measuring radius today and radial redistribution self-consistently. Stars selected by their birth R_orbit typically decreased in R_orbit, j_phi, and v_phi since birth. The trend for stars at a given R_orbit today depends on age: those younger than ~5 Gyr generally decreased in R_orbit, j_phi, and v_phi since birth, while those older generally increased in R_orbit, j_phi, and v_phi since birth. sigma(Delta R_orbit), a standard metric of radial redistribution, increases with stellar age only up to ~ 3 Gyr; it saturates at sigma(Delta R_orbit) ~2 kpc for older stars. This saturation contradicts a common expectation of a monotonic increase with age. Our results broadly agree with recent observational inferences of Delta R_orbit and sigma(Delta R_orbit) in the Milky Way. Across our FIRE-2 sample, the timing of disk formation does not correlate with sigma(Delta R_orbit), but it correlates with (net) Delta R_orbit.

Recent developments have shown that some semiclassical spacetimes cannot emerge from a traditional application of the rules of holography, prompting proposals for restoring their emergence with "observer rules". In this paper, we propose a general semiclassical diagnostic of such failures of emergence, and of the extent to which observer rules can fix them. Our diagnostic is the presence of certain "evanescent" quantum extremal surfaces, which are distinguished by an upper bound on their area rather than their generalized entropy. In particular, the generalized entropy of an evanescent QES may be large: even though its area term must be small, its bulk entanglement term is unconstrained. This feature is explained by an operational distinction between classical and quantum connectivity in semiclassical gravity, or equivalently between the two summands of the generalized entropy.

We present an HST imaging survey of a 1.1 sq. degree sky area toward the Milky Way Galactic Bulge. This field significantly overlaps with the upcoming Nancy Grace Roman Galactic Bulge Time Domain Survey (GBTDS). High angular resolution imaging of this area with HST before the start of the Roman Galactic Exoplanet Survey (RGES) will greatly strengthen Roman's ability to characterize detected exoplanet systems, as well as provide a rich and wide-field archive for use as a legacy dataset toward the Galactic Bulge for the broader community. We conduct coordinated-parallel imaging with both wide-field cameras on HST, Wide-field Camera 3 (WFC3) and Advanced Camera for Surveys (ACS), utilizing the F606W and F814W passbands. Approximately 70% of the survey was conducted during HST Cycle 32, with the remaining 30% conducted during Cycle 33. In this paper, the first in a series, we give a general overview of the program and the observing strategy, and present early results. This campaign secures HST's lasting impact on the high-precision study of stellar populations, dynamics, exoplanet systems, interstellar extinction, metallicities, cluster associations, and more toward the center of our Galaxy.

We show that Majorana neutrinos convert into antineutrinos in a background of ultralight vector dark matter coupled to lepton number, such as the gauge boson of $\text{U}(1)_{B-L}$ or $\text{U}(1)_{L_i - L_j}$ with $i, j = e , μ, τ$. This effect is suppressed by the small neutrino mass, but the enhancement by long astrophysical baselines can enable future searches for solar and supernova neutrinos to explore uncharted parameter space. For instance, for $\text{U}(1)_{B-L}$ dark matter, the observation of a supernova neutrino burst at DUNE, Hyper-Kamiokande, and JUNO could probe gauge couplings as small as $e^\prime \sim 10^{-32} - 10^{-25}$ for dark matter masses of $m_{A^\prime} \sim 10^{-22} \ \text{eV} - 10^{-14} \ \text{eV}$, beyond the capability of other future probes.

We propose a new class of infrared dualities relating three-dimensional $\mathcal{N}=2$ $USp(2N)$ Chern--Simons SQCD to planar Abelian quiver gauge theories. These dual descriptions are constructed via real mass deformations of established $\mathcal N=4$ mirror dualities between $\mathcal{N}=4$ $USp(2N)$ SQCD and unitary $D$-type quiver gauge theories. The resulting $\mathcal N=2$ dual pairs exhibit the characteristic exchange of topological and flavor symmetries. We provide nontrivial evidence for these dualities by matching $\mathbf{S}^3_b$ partition functions, superconformal indices, and gauge-invariant operator spectra. Furthermore, we systematically incorporate additional real mass deformations on both sides of the duality, allowing us to extend the construction to $\mathcal{N}=2$ symplectic SQCD with generic ranks, flavors, and Chern--Simons levels.

We present SIRENA, a Python and C++ implementation of the Laporta algorithm for the automatic reduction of multi-loop sum-integrals via integration-by-parts identities. The method builds on established techniques for zero-temperature Feynman integrals and extends them to finite-temperature quantum field theory by consistently accounting for the Matsubara sum structure. We validate the framework by reproducing several known results from the literature up to 3-loop order, and we further provide, for the first time, reductions for selected 3-loop fermionic sum-integrals. In addition to the package, we derive an analytic factorization formula for arbitrary 2-loop fermionic sum-integrals, extending on a previous result for the bosonic case.

We show that for both QED and perturbative quantum gravity, finite-time Faddeev-Kulish dressings can be fully constrained by symmetry, and that this gives the unique choice which reproduces the classical memory effect. For gravity, we show that using this dressing to construct finite-time Fock spaces, as well as a carefully defined finite-time memory detector allows us to recover both the first order gravitational memory, as well as higher order Christodoulou contributions from the gravitational field. We explain how these higher order perturbative corrections arise in inclusive in-in calculations.

Valleytronics and valley photonics exploit the valley degree of freedom to encode and manipulate information. Here we show that photonic valleys can be selectively addressed in quantum optics using a simple two-level emitter, provided it is coupled nonlocally to the field, thereby realizing a so-called giant atom. Specifically, we consider a qubit coupled at multiple points to an engineered honeycomb lattice of resonators with detuned sublattice frequencies. By tailoring the geometry of the coupling points, the giant atom can be made to emit selectively into a single valley. The emitted photons thereby acquire a well-defined valley character and inherit the associated Berry curvature. By placing the qubit near a domain wall between regions of opposite sublattice detuning, whose interface supports valley-polarized edge modes, emission becomes chiral along the domain wall. This provides a promising route toward implementation of single-photon disorder-robust chiral emission without breaking time-reversal symmetry of the electromagnetic medium in platforms such as circuit QED.

The competition between metal synthesis and feedback from massive stars establishes the mass-metallicity relation (MZR) at low-redshifts. Examining this relation at higher redshifts, particularly at the low-mass end $\lesssim10^{8}\,{\rm M_\odot}$, is essential for understanding chemical enrichment and stellar feedback. In this study, we utilize the deep ($\sim30\,$hrs) JWST/NIRSpec G395M GLIMPSE-D survey of the lensed field Abell S1063, to explore the low-mass end of the MZR at high redshift ($z\sim6-8$). We identify eight [OIII]$λ$4364 emitters, enabling the most reliable "direct" metallicity measurements in galaxies down to stellar masses of $\sim10^{6-8}\,{\rm M_\odot}$. By combining our sample and galaxies with [OIII]$λ$4364 detections from the literature, we calculate direct metallicities for 21 galaxies. We compare our direct metallicities to those derived from strong-line diagnostics, and find them to be consistent with previous calibrations. We fit the MZR at $10^{6.7-9}\,M_{\odot}$ with $\sim0.3-0.5$ dex lower metallicity than local galaxies at similar stellar mass. We find the slope to be $0.25\pm0.10$, comparable to the local MZR; and the MZR exhibits a scatter of $\sim0.2\,$dex, which is larger than the local MZR, The lower metallicities may reflect denser, more gas-rich early environments, with continuous inflow of metal-poor gas diluting the ISM metallicity. In addition, we show that in extremely high electron densities ($n_e \gtrsim 10^5\,{\rm cm^{-3}}$), metallicities can be significantly underestimated ($\sim0.5$ dex), if lower $n_e$ are assumed for galaxies with high $n_e$. In a nutshell, these observations provide the first glimpse of the low-mass MZR at $z\sim6-8$ using direct metallicity measurements. More deep spectroscopic observations in lensed fields will be critical to robustly characterize the MZR and chemical evolution in the early universe.

Optical emission-line ratios in star-forming galaxies at $z \sim 3$-8, such as [OIII]/H$β$ and [OIII]/[OII], are strongly offset from those at $z \sim 0$-2, pointing to more extreme ionization and ISM conditions in the early Universe. To constrain the physical origin of these offsets, we developed Lumen, a framework for modelling nebular emission from spatially distributed HII regions in cosmological simulations. We apply Lumen to IllustrisTNG50, validate its predictions at low redshift, and test a suite of proposed mechanisms for producing extreme line ratios at $z = 3$-8. We focus on the [NII]/H$α$ versus [OIII]/H$β$ (N2-BPT) diagram, the [SII]/H$α$ versus [OIII]/H$β$ (S2-VO87) diagram, and the [OIII]/[OII] versus ([OII]+[OIII])/H$β$ (O32-R23) diagram. We find that $α$-enhancement alone cannot explain the bulk of observations. Moderate offsets emerge from the combined effects of $α$-enhancement, a higher IMF upper-mass cutoff, and AGN contributions. The most extreme [OIII]/H$β$ and [OIII]/[OII] values require high ionization parameters powered by massive star clusters of $\gtrsim 10^5$-$10^6\,\mathrm{M}_\odot$, consistent with recent JWST observations. Reproducing the highest [NII]/H$α$ ratios additionally requires enhanced nitrogen abundances. Although gas densities of $n \sim 10^4\,\mathrm{cm}^{-3}$ can boost several diagnostic ratios, they suppress [SII]/H$α$ and are therefore in tension with current observations. Overall, models combining harder ionizing spectra, elevated ionization parameters from massive star clusters, and enhanced nitrogen abundances reproduce the observed high-$z$ galaxy population across the N2-BPT, S2-VO87, and O32-R23 diagrams. This successful model also motivates new demarcation lines for star-forming galaxies in the N2-BPT and S2-VO87 diagrams.

We introduce an unbinned method for extracting the CKM angle $γ$ from the decay chain $B^\pm \to (D \to K_S π^+ π^-) K^\pm$ using normalizing flows (NFs). The NFs, trained on $D$ decay data, learn a faithful continuous representation of the amplitude and strong phase variation over the $D\to K_Sπ^+π^-$ Dalitz plot whose fidelity improves with increased data sample sizes. With this input, the $B$ decay data can be used to extract the parameters $r_B$, $δ_B$, and $γ$. We test the method on Monte Carlo generated data, where it successfully recovers the injected value of $γ$ within uncertainties. The present implementation propagates statistical uncertainties from finite training data via an ensemble of independently trained flows, and does not attempt to capture the effects of systematic experimental errors. We explore two versions of the method that differ in how the trigonometric constraint on phase variation is encoded, and comment on the possible extension to Bayesian NFs, which would provide direct uncertainty estimates on the learned densities without requiring ensemble training.

Cosmic rays (CRs) are important drivers for molecular chemistry in star-forming regions, and laboratory experiments have shown that CRs can stimulate the release of complex organic molecules (COMs) such as methanol. Observationally, this has primarily been tested in cold, low-mass cores, so studying how CRs affect COM formation in a high-mass star-forming environment is of great interest. We performed a high-sensitivity wide-band spectral line survey with the Onsala 20 m telescope towards the high-mass protostar Cepheus A HW2, which is known to host an ionised jet. Consistent with previous studies, two primary velocity components ($-11$ km s$^{-1}$ and $-5$ km s$^{-1}$) were identified. Column densities and relative abundances of the detected ions and COMs were estimated from rotational diagrams, single transitions and RADEX grid searches (CH$_3$OH: $1.6\times10^{-9}$, CH$_3$CN: $5.9\times10^{-11}$, t-HCOOH: $7.9\times10^{-11}$, H$_2$CCO: $1.7\times10^{-11}$, CH$_3$CHO: $1.9\times10^{-11}$, CH$_3$OCHO: $7.6\times10^{-10}$ at $-11$ km s$^{-1}$). Deuterium fractions were also estimated (in range $0.002-0.3$ at $-11$ km s$^{-1}$), and the volume density of molecular hydrogen ($2.6\times10^5$ cm$^{-3}$ at $-11$ km s$^{-1}$) was constrained from the RADEX grid searches. Electron fractions and CR ionisation rates (CRIR, $6.8\times10^{-17}$ s$^{-1}$ at $-11$ km s$^{-1}$, $\leq9.2\times10^{-19}$ s$^{-1}$ at $-5$ km s$^{-1}$) were estimated through analytic chemistry using different ions as probes. The gas-grain chemical code Nautilus reproduced the observed abundances of CH$_3$OH, CH$_3$CN, HCO$^+$, N$_2$H$^+$ at the observed density, temperature and CRIR within the uncertainty of the model. The results indicate that the CR ionisation rate of the kinematic component associated with most of the COMs' emission in the region is locally enhanced.

Fast radio bursts (FRBs) which are well localised ($<$1") to their host galaxy are tools for studying cosmology and the intergalactic medium. Furthermore, high-time resolution datasets of their polarisation properties can enable testing of the numerous models on their potential progenitors. To that end, the CELEBI (CRAFT Effortless Localisation and Enhanced Burst Inspection) pipeline was conceived to enable data reduction from raw antenna voltages to detect fast radio transient events, localise them to sub-arcsecond precision, and produce polarimetric data at time resolutions as fine as 3 ns. Here we present a slew of updates to the CELEBI pipeline. Improvements to the astrometry correction for FRB localisations has aided our ability to determine what part of a galaxy more nearby FRBs have occurred in, which can have its own implication on the progenitor. We also have implemented time and frequency gating on detected fast transients to enable a boost to signal-to-noise, particularly useful for high dispersion measure or faint fast radio transients. We give examples of our improvements to the localisation, including for the currently 'hostless' FRB 20251019A. The polarisation calibration process has been overhauled, resulting in much more accurate measurements of derived polarisation fraction and rotation measures. Furthermore, we now have incorporated tools for structure-maximisation of the dispersion measure of fast radio transients, a software container which enables the installation of CELEBI on other machines, and improved the pipeline efficiency. Together these updates (named 'Polarisation and astrometry Improvements for New Knowledge', or PINK) greatly improve our ability to keep up with the expected detection rate from the CRAFT COherent (CRACO) upgrade to the real-time fast transient detection system of the Australian SKA Pathfinder.

We demonstrate that language models can autonomously replicate their weights and harness across a network by exploiting vulnerable hosts. The agent independently finds and exploits a web-application vulnerability, extracts credentials, and deploys an inference server with a copy of its harness and prompt on the compromised host. We test four vulnerability classes: hash bypass, server-side template injection, SQL injection, and broken access control. Qwen3.5-122B-A10B succeeds in 6-19% of attempts, and the smaller Qwen3.6-27B reaches 33% on a single A100. This already matches the current-generation GPT-5.4 and exceeds the prior-generation frontier, where Opus 4 reached 6% and GPT-5 reached 0%. Replicating Qwen weights, frontier models reach 81% (Opus 4.6) and 33% (GPT-5.4). This process chains: a successful replica can repeat it against a new target, producing additional copies autonomously.

System dynamics is a methodology that is widely used in many academic fields. It explains the behavior of social and economic systems with models that capture complex causality and feedback effects. This 'practice paper' discusses the opportunities and barriers for introducing feedback thinking and system dynamics models in the economics curriculum. We start by providing a pricing feedback model that illustrates some of the benefits that system dynamics can provide in enhancing economics education. Then we summarize the experiences of each of the authors in teaching system dynamics on economics educational programs. This includes different approaches to teaching economics with system dynamics that depend on the learning objectives, the preparation of students, and the background of the instructor. We also develop a four-level course hierarchy for using system dynamics in economics teaching. We then point out the tradeoffs that instructors must consider as they introduce new pedagogies for delivering economics material. Finally, we provide some concluding comments with some suggestions for future work. The expected audiences for this paper are instructors as well as graduate students who are considering academia as a profession.

Based on (2712.4+-14.3)*10^{6} psi(3686) events collected with the BESIII detector, the decays Xi(1530)^{-} to Xi^{0} pi^{-} and Xi(1530)^{-} to Xi^{-} pi^{0} are investigated jointly via the process psi(3686) to anti-Xi^{+} Xi(1530)^{-} + c.c. Under the assumption of isospin symmetry, the two decay modes are treated as fully correlated, and we report the first measurement of their absolute branching fractions. The results are B(Xi(1530)^{-} to Xi^{0} pi^{-})=(61.4+-4.5+-4.6)% and B(Xi(1530)^{-} to Xi^{-} pi^{0}) =(29.7+-2.2+-2.2)%. The combined branching fraction of the two decays is B(Xi(1530)^{-} to (Xi pi)^{-})=(91.1+-6.7+-6.8)%, with uncertainties accounting for the correlations between the two modes. Here, the first uncertainties are statistical, while the second are systematic. Additionally, we update the branching fraction of the decay psi(3686) to anti-Xi^{+} Xi(1530)^{-} + c.c. The updated measurement is B(psi(3686) to anti-Xi^{+} Xi(1530)^{-} + c.c.)=(8.67+-0.52+-0.58+-0.57)*10^{-6}, where the first uncertainty is statistical, the second is systematic related to event selection and the fit model, and the third is associated with the interference effect.

Quantum spin liquids (QSLs) represent highly entangled states of matter in which frustration-induced quantum fluctuations suppress any symmetry-breaking phase transition down to absolute zero, giving rise to fractionalized excitations and emergent gauge fields. Theoretically, bond anisotropy can stabilize QSLs even on bipartite lattices, as exemplified by the Kitaev honeycomb model; however, no material has so far been established to realize such a state as its true ground state. Here we identify the three-dimensional spin-trimer magnet KBa$_3$Ca$_4$Cu$_3$V$_7$O$_{28}$ as a promising candidate for a bipartite quantum spin liquid persisting to the lowest temperatures. Strongly coupled Cu$^{2+}$ trimers form effective pseudospin-1/2 degrees of freedom upon cooling, which in turn constitute a three-dimensional bipartite network. Bulk thermodynamic measurements, neutron scattering, $μ$SR, and NMR detect no spin freezing or symmetry-breaking phase transition down to 20 mK, but instead reveal a gapless dynamical ground state with algebraic spin autocorrelations. Complementary Monte Carlo and exact-diagonalization calculations show that this state is stabilized by a strong enhancement of effective anisotropy: a weak microscopic Cu-Cu exchange anisotropy of approximately 15 percent is generically amplified at the trimer level, producing effective pseudospin-pseudospin interaction anisotropies of 60 to 100 percent. Our results establish trimer-based networks as a promising platform for realizing anisotropy-stabilized quantum entangled states, even in three-dimensional bipartite systems with only weak microscopic anisotropy.

The security of wireless challenge-response Physical Layer Authentication (PLA) based on Orthogonal Frequency Division Multiplexing (OFDM) relies on a sufficiently random fading channel condition, which is commonly assumed in existing studies. However, in practical scenarios, such a condition is not always guaranteed and the responses of OFDM subchannels may exhibit correlation.} Consequently, ensuring the security of such PLA systems remains an unsolved problem. In this paper, we propose a novel adversary model, called Maximum Differential Likelihood Generator (MDLG), which exploits the weak correlation property in practical wireless channel to launch effective attacks against PLA. Based on this model, we create a measurable guideline using randomness testing to decide when we can in fact use PLA in a practical wireless channel condition. Extensive real-world experiments validate the effectiveness of the MDLG attack and demonstrate how the proposed guideline can help protect the security of PLA.

We have discovered a memory effect in the temperature sensitivity of a cryogenic sapphire microwave resonator, at the heart of the ultra-stable Cryogenic Sapphire Oscillators (CSOs). Such effect is due to the relaxaxtion time of Cr3+ impurities, and results in hysteresis in the frequency vs temperature behavior, These paramagnetic impurities, always present in synthetic sapphire, produce a temperature turning point which is necessary to achieve ultimate frequency stability. The practical implication on the CSO is that the sapphire resonators's frequency depends on the rate of temperature change. This dynamical thermal sensitivity results in a wide bump in the Allan deviation at 10 s integration time, where the frequency stability is degraded. The actual degradation depends on the specie and on the amount of the dominant paramagnetic impurity.

A hallmark of life on Earth is the ability of agents to exert causal power and be drivers of subsequent events. This is key to cognition at all scales. Causal emergence, measuring the degree to which an agent exerts unique predictive power on its future, is one consequence of causal power. Indeed, recent discoveries have shown that biological agents, even minimal ones, increase their causal emergence after learning new memories. However, there is a major knowledge gap regarding how causally emergent artificial agents are. We focused on Reinforcement Learning (RL) of neural-network agents across an array of environmental conditions, encompassing different algorithms, agent architectures, and six environments arranged on a complexity spectrum. For consistency, we computed the causal emergence of their latent-space representations over their lifetimes. We used the recently proposed ΦID to estimate causal emergence and tested how it related to learning performance. Our results suggested a Causally Emergent Alignment Hypothesis: successful agents exhibited causal emergence that was consistently predictive of final reward early in training and whose representational dynamics aligned with reward improvement in most tasks. This idea suggests that causal emergence may be a previously undisclosed axis of reorganization of neural representations in RL agents, with the potential to establish causal relationships and interventions that will lead to better RL agents. Our work also highlights the alignment between causal emergence and learning as another way biological and artificial creatures compare.

Computational complexity and storage requirements are crucial factors influencing the performance and efficiency of convolutional neural networks (CNNs) in resource-constrained environments. This paper presents a high-performance embedded target detection system based on FPGA and YOLOv3-Tiny, specifically designed for embedded artificial intelligence applications. By integrating lightweight CNN optimization techniques with hardware accelerator design, significant improvements are made in both computational efficiency and resource utilization. Key optimizations, including low-bit quantization, batch normalization fusion, and table lookup mapping, reduce model parameters and computational complexity. Additionally, an FPGA hardware accelerator with a pipelined architecture is developed to enhance the efficiency of convolution operations while minimizing off-chip data transmission through modular design and on-chip cache optimization. On the ZYNQ-XC7Z035 platform, the system achieves an inference latency of 0.211 seconds, outperforming comparable designs by 75.58% in speed. The system achieves an power efficiency of 10.11 GOPS/W, surpassing comparable designs by at least 29.45%. Furthermore, hardware resource utilization is reduced by up to 51.94% compared to similar systems. This study offers innovative design methodologies and practical application examples for the efficient deployment of deep learning models on embedded platforms.

Attackers willing to compromise computing systems can use malicious peripherals as an attack vector, threatening users that cannot verify the hardware's authenticity. To address this problem, our work uses the Security Protocol and Data Model to propose a UEFI system capable of authenticating PCIe and USB devices trying to connect with it. We also develop an open source proof-of-concept using emulation to evaluate and illustrate our proposal, which is capable of restricting the devices' connections to only those allowed, thus protecting the system against malicious peripherals. Then, using kernel virtualization features to evaluate the emulation, we collect the number of instructions and CPU cycles during boot. Our experiments reveal that, during firmware execution, the number of instructions and the number of CPU cycles increased respectively 13% and 8% on average. This processing overhead is acceptable in view of enhanced security. Institutions requiring high security levels can leverage our proof-of-concept to tailor their own system based on their own requirements.

Social contact matrices are essential tools in infectious disease epidemiology as they quantify close-range human contact patterns which directly drive the transmission of airborne infectious diseases. In this work we propose a Bayesian modeling framework for inferring generalized contact matrices which stratify contact matrices beyond contemporary age dimensions. The model is designed to satisfy fundamental structural assumptions of contacts while leveraging tensor structures and smoothing constraints to make high-dimensional matrix estimation computationally feasible and statistically stable. We discover a link between multi-dimensional matrix stratification subject to structural constraints with the theory of contingency tables. This enables us to approach a challenging missing-data problem commonly encountered in real-world analysis where feature information on the contacts is unobserved. We benchmark the framework against existing methods through simulation studies and illustrate the framework's practical utility through two real-world datasets: BICS (United States) and COVIMOD (Germany). Our models are implemented in an open-source Python package to facilitate adoption in the wider scientific community.

We extend the concept of two-way forest diagrams, introduced by Belk and Brown in 2003, to represent elements of $F(n)$ as a pair of infinite, bounded $n$-ary forests together with an order-preserving bijection of the leaves. This representation allows us to develop an alternative way to compute the length of an element of $F(n)$, distinct from the formula established by Fordham and Cleary in 2009. As an application of our length formula, we re-prove the existence of dead end elements in $F(n)$ and show that their depth is always two, first proved by Wladis in 2009.

Integration of Ordinary Differential Equations (ODEs) using Backward Difference formula (BDF) methods with p backward steps achieves order p accuracy if specific conditions are met. This work extends the composition technique with complex coefficients to the implicit BDF schemes, increasing the approximation order by one without additional backward points. The imaginary part of the composed flow provides an error estimate of order p + 1. Linear stability analysis reveals that the composed schemes break the Dahlquist barrier, achieving stability up to order eight. The computational performance of the composed flow outperforms BDF schemes when using the same number of backward points, allowing for higher accuracy with lower CPU time. For non-uniform meshes, the ratio of consecutive time steps, which influences stability, appears as a parameter in the roots of algebraic equations relative to the composed flow. Having a complex root with a real positive part implies a lower bound to this ratio depending on the order. For example, the bound is 0.4506 for order three and 0.6806 for order four. Numerical tests demonstrate the effectiveness of this technique in improving the accuracy and stability compared to BDF methods.

Blow-up in graph theory is a procedure in which each vertex is replaced by copies of itself, and two copies are adjacent if and only if the original vertices are adjacent. In this paper, we extend the concept of graph blow-up to a more general combinatorial context and discuss its potential applications.

We prove the necessity part of the higher-order Szegő theorem on the unit circle for the single-critical-point weights $H_m(e^{iθ})=(1-\cosθ)^m$, $m\ge1$. If $\{α_n\}_{n\ge0}$ are the Verblunsky coefficients of a nontrivial probability measure $dμ=w(θ)dθ/(2π)+dμ_{\mathrm s}$, then the weighted Szegő condition $\int_0^{2π} (1-\cosθ)^m\log w(θ)\frac{dθ}{2π}>-\infty$ implies $Δ^mα\in\ell^2, \,\, α\in\ell^{2m+2}.$ The proof uses a finite-volume version of Yan's higher-order sum rule. The quadratic part yields the $m$-th difference energy, and the logarithmic tail yields the $\ell^{2m+2}$-control. The non-sign-definite critical terms are treated in two steps. First, the quartic principal critical block is isolated using the Yan quotient-algebra normal representative and shown to have a positive semidefinite Gram representation. Second, the remaining non-principal critical terms are controlled by the diagonal-vanishing property $\mathcal Y_{k,\mathrm{crit}}^{(m)} \in \mathfrak I_k^{\,m+1-k}, \,\, 2\le k\le m,$ together with the Breuer--Simon--Zeitouni normal form, discrete interpolation, and Young's inequality. These estimates yield a uniform finite-volume coercive bound, from which the necessity theorem follows for all $m\ge1$.