This report describes the experimental strategy and technologies for XLZD, the next-generation xenon observatory sensitive to dark matter and neutrino physics. In the baseline design, the detector will have an active liquid xenon target of 60 tonnes, which could be increased to 80 tonnes if the market conditions for xenon are favorable. It is based on the mature liquid xenon time projection chamber technology used in current-generation experiments, LZ and XENONnT. The report discusses the baseline design and opportunities for further optimization of the individual detector components. The experiment envisaged here has the capability to explore parameter space for Weakly Interacting Massive Particle (WIMP) dark matter down to the neutrino fog, with a 3$\sigma$ evidence potential for WIMP-nucleon cross sections as low as $3\times10^{-49}\rm\,cm^2$ (at 40 GeV/c$^2$ WIMP mass). The observatory will also have leading sensitivity to a wide range of alternative dark matter models. It is projected to have a 3$\sigma$ observation potential of neutrinoless double beta decay of $^{136}$Xe at a half-life of up to $5.7\times 10^{27}$ years. Additionally, it is sensitive to astrophysical neutrinos from the sun and galactic supernovae.

Transformers exhibit in-context learning (ICL): the ability to use novel information presented in the context without additional weight updates. Recent work shows that ICL emerges when models are trained on a sufficiently diverse set of tasks and the transition from memorization to generalization is sharp with increasing task diversity. One interpretation is that a network's limited capacity to memorize favors generalization. Here, we examine the mechanistic underpinnings of this transition using a small transformer applied to a synthetic ICL task. Using theory and experiment, we show that the sub-circuits that memorize and generalize can be viewed as largely independent. The relative rates at which these sub-circuits learn explains the transition from memorization to generalization, rather than capacity constraints. We uncover a memorization scaling law, which determines the task diversity threshold at which the network generalizes. The theory quantitatively explains a variety of other ICL-related phenomena, including the long-tailed distribution of when ICL is acquired, the bimodal behavior of solutions close to the task diversity threshold, the influence of contextual and data distributional statistics on ICL, and the transient nature of ICL.

We generalize the Affleck-Kennedy-Lieb-Tasaki model to a spin-$S$ ferromagnetic model with exactly-written ground states, known as the partially-magnetized valence bond solid (VBS) states with magnetization $m=(S-1)/S$. We find that the VBS state and an antiferromagnetic ground state with magnetization $m=0$ are degenerate for $S=3/2$ and $S=2$ by using the Lanczos method and the density matrix renormalization group method (DMRG). However, increasing $S$, the magnetization of the ground states is uniquely determined as the fraction $m=(S-1)/S$. This is not just a ferromagnet, but a quantum ferromagnet due to quantum entanglement inherent in VBS states. In the low-energy excitation spectrum, we find the coexistence of the Haldane gap and Goldstone-like ferromagnetic magnon excitation. This ``magnetic chimera'' clearly appears under a finite magnetic field. Finally, we discuss an application to the measurement-based quantum computation and an extension of the Haldane's conjecture.

The metastable hypermassive neutron star produced in the coalescence of two neutron stars can copiously produce axions that radiatively decay into $\mathcal{O}(100)$~MeV photons. These photons can form a fireball with characteristic temperature smaller than $1\rm\, MeV$. By relying on X-ray observations of GW170817/GRB 170817A with CALET CGBM, Konus-Wind, and Insight-HXMT/HE, we present new bounds on the axion-photon coupling for axion masses in the range $1$-$400\,\rm MeV$. We exclude couplings down to $5\times 10^{-11}\,\rm GeV^{-1}$, complementing and surpassing existing constraints. Our approach can be extended to any feebly-interacting particle decaying into photons.

Deci-hertz Interferometer Gravitational Wave Observatory (DECIGO) is the future Japanese space mission with a frequency band of 0.1 Hz to 10 Hz. DECIGO aims at the detection of primordial gravitational waves, which could be produced during the inflationary period right after the birth of the universe. There are many other scientific objectives of DECIGO, including the direct measurement of the acceleration of the expansion of the universe, and reliable and accurate predictions of the timing and locations of neutron star/black hole binary coalescences. DECIGO consists of four clusters of observatories placed in the heliocentric orbit. Each cluster consists of three spacecraft, which form three Fabry-Perot Michelson interferometers with an arm length of 1,000 km. Three clusters of DECIGO will be placed far from each other, and the fourth cluster will be placed in the same position as one of the three clusters to obtain the correlation signals for the detection of the primordial gravitational waves. We plan to launch B-DECIGO, which is a scientific pathfinder of DECIGO, before DECIGO in the 2030s to demonstrate the technologies required for DECIGO, as well as to obtain fruitful scientific results to further expand the multi-messenger astronomy.

Local projections (LPs) are widely used for impulse response analysis, but Bayesian methods face challenges due to the absence of a likelihood function. Existing approaches rely on pseudo-likelihoods, which often result in poorly calibrated posteriors. We propose a quasi-Bayesian method based on the Laplace-type estimator, where a quasi-likelihood is constructed using a generalized method of moments criterion. This approach avoids strict distributional assumptions, ensures well-calibrated inferences, and supports simultaneous credible bands. Additionally, it can be naturally extended to the instrumental variable method. We validate our approach through Monte Carlo simulations.

Extreme mass ratio inspirals (EMRIs) are among the most interesting gravitational wave (GW) sources for space-borne GW detectors. However, successful GW data analysis remains challenging due to many issues, ranging from the difficulty of modeling accurate waveforms, to the impractically large template bank required by the traditional matched filtering search method. In this work, we introduce a proof-of-principle approach for EMRI detection based on convolutional neural networks (CNNs). We demonstrate the performance with simulated EMRI signals buried in Gaussian noise. We show that over a wide range of physical parameters, the network is effective for EMRI systems with a signal-to-noise ratio larger than 50, and the performance is most strongly related to the signal-to-noise ratio. The method also shows good generalization ability towards different waveform models. Our study reveals the potential applicability of machine learning technology like CNNs towards more realistic EMRI data analysis.

Formula-driven supervised learning (FDSL) is a pre-training method that relies on synthetic images generated from mathematical formulae such as fractals. Prior work on FDSL has shown that pre-training vision transformers on such synthetic datasets can yield competitive accuracy on a wide range of downstream tasks. These synthetic images are categorized according to the parameters in the mathematical formula that generate them. In the present work, we hypothesize that the process for generating different instances for the same category in FDSL, can be viewed as a form of data augmentation. We validate this hypothesis by replacing the instances with data augmentation, which means we only need a single image per category. Our experiments shows that this one-instance fractal database (OFDB) performs better than the original dataset where instances were explicitly generated. We further scale up OFDB to 21,000 categories and show that it matches, or even surpasses, the model pre-trained on ImageNet-21k in ImageNet-1k fine-tuning. The number of images in OFDB is 21k, whereas ImageNet-21k has 14M. This opens new possibilities for pre-training vision transformers with much smaller datasets.

A large dataset of annotated traffic accidents is necessary to improve the accuracy of traffic accident recognition using deep learning models. Conventional traffic accident datasets provide annotations on traffic accidents and other teacher labels, improving traffic accident recognition performance. However, the labels annotated in conventional datasets need to be more comprehensive to describe traffic accidents in detail. Therefore, we propose V-TIDB, a large-scale traffic accident recognition dataset annotated with various environmental information as multi-labels. Our proposed dataset aims to improve the performance of traffic accident recognition by annotating ten types of environmental information as teacher labels in addition to the presence or absence of traffic accidents. V-TIDB is constructed by collecting many videos from the Internet and annotating them with appropriate environmental information. In our experiments, we compare the performance of traffic accident recognition when only labels related to the presence or absence of traffic accidents are trained and when environmental information is added as a multi-label. In the second experiment, we compare the performance of the training with only contact level, which represents the severity of the traffic accident, and the performance with environmental information added as a multi-label. The results showed that 6 out of 10 environmental information labels improved the performance of recognizing the presence or absence of traffic accidents. In the experiment on the degree of recognition of traffic accidents, the performance of recognition of car wrecks and contacts was improved for all environmental information. These experiments show that V-TIDB can be used to learn traffic accident recognition models that take environmental information into account in detail and can be used for appropriate traffic accident analysis.

We demonstrate a high-efficiency, free space optical delay line utilizing a nested multipass cell architecture. This design supports extended optical paths with low loss, aided by custom broadband dielectric coating that provides high reflectivity across a wide spectral bandwidth. The cell is characterized using polarization-entangled photon pairs, with signal photons routed through the delay line and idler photons used as timing reference. Quantum state tomography performed on the entangled pair reveals entanglement preservation with a fidelity of $99.6(9)\%$ following a single-transit delay of up to $687$~ns, accompanied by a photon retrieval efficiency of $95.390(5)$. The delay is controllable and can be set between $1.8$~ns to $687$~ns in $\sim12.6$~ns increments. The longest delay and wide spectral bandwidth result in a time-bandwidth product of $3.87\times 10^7$. These results position this delay line as a strong candidate for all-optical quantum memories and synchronization modules for scalable quantum networks.

We report results from a series of three-dimensional (3D) rotational core-collapse simulations for $11.2$ and 27 $M_{/odot}$ stars employing neutrino transport scheme by the isotropic diffusion source approximation. By changing the initial strength of rotation systematically, we find a rotation-assisted explosion for the 27$M_{/odot}$ progenitor, which fails in the absence of rotation. The unique feature was not captured in previous two-dimensional (2D) self-consistent rotating models because the growing non-axisymmetric instabilities play a key role. In the rapidly rotating case, strong spiral flows generated by the so-called low $T/|W|$ instability enhance the energy transport from the proto-neutron star (PNS) to the gain region, which makes the shock expansion more energetic. The explosion occurs more strongly in the direction perpendicular to the rotational axis, which is different from previous 2D predictions.

The multi-messenger observation of compact binary coalescence promises great scientific treasure. However, a synthetic observation from both gravitational wave and electromagnetic channels remains challenging. Relying on the day-to-week long macronova emission, GW170817 remains the only event with successful electromagnetic followup. In this manuscript, we explore the possibility of using the early stage X-ray afterglow to search for the electromagnetic counterpart of gravitational wave events. Two algorithms, the sequential observation and the local optimization are considered and applied to three simulated events. We consider the proposed Einstein probe as a candidate X-ray telescope. Benefiting from the large field of view and high sensitivity, we find that the sequential observation algorithm not only is easy to implement, but also promises a good chance of actual detection.

This work characterises the sky localization and early warning performance of networks of third generation gravitational wave detectors, consisting of different combinations of detectors with either the Einstein Telescope or Cosmic Explorer configuration in sites in North America, Europe and Australia. Using a Fisher matrix method which includes the effect of earth rotation, we estimate the sky localization uncertainty for $1.4\text{M}\odot$-$1.4\text{M}\odot$ binary neutron star mergers at distances $40\text{Mpc}$, $200\text{Mpc}$, $400\text{Mpc}$, $800\text{Mpc}$, $1600\text{Mpc}$, and an assumed astrophysical population up to redshift of 2 to characterize its performance for binary neutron star observations. We find that, for binary neutron star mergers at $200\text{Mpc}$ and a network consisting of the Einstein Telescope, Cosmic Explorer and an extra Einstein Telescope-like detector in Australia(2ET1CE), the upper limit of the size of the 90% credible region for the best localized 90% signals is $0.25\text{deg}^2$. For the simulated astrophysical distribution, this upper limit is $91.79\text{deg}^2$. If the Einstein Telescope-like detector in Australia is replaced with a Cosmic Explorer-like detector(1ET2CE), for $200\text{Mpc}$ case, the upper limit is $0.18\text{deg}^2$, while for astrophysical distribution, it is $56.77\text{deg}^2$. We note that the 1ET2CE network can detect 7.2% more of the simulated astrophysical population than the 2ET1CE network. In terms of early warning performance, we find that a network of 2ET1CE and 1ET2CE networks can both provide early warnings of the order of 1 hour prior to merger with sky localization uncertainties of 30 square degrees or less. Our study concludes that the 1ET2CE network is a good compromise between binary neutron stars detection rate, sky localization and early warning capabilities.

Understanding stars and their evolution is a key goal of astronomical research and has long been a focus of human interest. In recent years, theorists have paid much attention to the final interior processes within massive stars, as they can be essential for revealing neutrino-driven supernova mechanisms and other potential transients of massive star collapse. However, it is challenging to observe directly the last hours of a massive star before explosion, since it is the supernova event that triggers the start of intense observational study. Here we report evidence for a final phase of stellar activity known as a ``shell merger'', an intense shell burning in which the O-burning shell swallows its outer C-/Ne-burning shell, deep within the progenitor's interior moments before the supernova explosion. In the violent convective layer created by the shell merger, Ne, which is abundant in the stellar O-rich layer, is burned as it is pulled inward, and Si, which is synthesized inside, is transported outward. The remnant still preserves some traces of such Ne-rich downflows and Si-rich upflows in the O-rich layer, suggesting that inhomogeneous shell-merger mixing began just hours ($\lesssim 10^4$ s) before its gravitational collapse. Our results provide the first observational evidence that the final stellar burning process rapidly alters the internal structure, leaving a pre-supernova asymmetry. This breaking of spherical symmetry facilitates the explosion of massive stars and influences various supernova and remnant characteristics, including explosion asymmetries and the neutron star's kick and spin.

We consider a class of holographic tensor networks that are efficiently contractible variational ansatze, manifestly (approximate) quantum error correction codes, and can support power-law correlation functions. In the case when the network consists of a single type of tensor that also acts as an erasure correction code, we show that it cannot be both locally contractible and sustain power-law correlation functions. Motivated by this no-go theorem, and the desirability of local contractibility for an efficient variational ansatz, we provide guidelines for constructing networks consisting of multiple types of tensors that can support power-law correlation. We also provide an explicit construction of one such network, which approximates the holographic HaPPY pentagon code in the limit where variational parameters are taken to be small.

We present formalism necessary to determine weak-scale matching coefficients in the computation of scattering cross sections for putative dark matter candidates interacting with the Standard Model. Particular attention is paid to the heavy-particle limit. A consistent renormalization scheme in the presence of nontrivial residual masses is implemented. Two-loop diagrams appearing in the matching to gluon operators are evaluated. Details are given for the computation of matching coefficients in the universal limit of WIMP-nucleon scattering for pure states of arbitrary quantum numbers, and for singlet-doublet and doublet-triplet mixed states.

We describe a search for gravitational waves from compact binaries with at least one component with mass 0.2 $M_\odot$ -- $1.0 M_\odot$ and mass ratio $q \geq 0.1$ in Advanced LIGO and Advanced Virgo data collected between 1 November 2019, 15:00 UTC and 27 March 2020, 17:00 UTC. No signals were detected. The most significant candidate has a false alarm rate of 0.2 $\mathrm{yr}^{-1}$. We estimate the sensitivity of our search over the entirety of Advanced LIGO's and Advanced Virgo's third observing run, and present the most stringent limits to date on the merger rate of binary black holes with at least one subsolar-mass component. We use the upper limits to constrain two fiducial scenarios that could produce subsolar-mass black holes: primordial black holes (PBH) and a model of dissipative dark matter. The PBH model uses recent prescriptions for the merger rate of PBH binaries that include a rate suppression factor to effectively account for PBH early binary disruptions. If the PBHs are monochromatically distributed, we can exclude a dark matter fraction in PBHs $f_\mathrm{PBH} \gtrsim 0.6$ (at 90% confidence) in the probed subsolar-mass range. However, if we allow for broad PBH mass distributions we are unable to rule out $f_\mathrm{PBH} = 1$. For the dissipative model, where the dark matter has chemistry that allows a small fraction to cool and collapse into black holes, we find an upper bound f_{\mathrm{DBH}} < 10^{-5} on the fraction of atomic dark matter collapsed into black holes.

To study properties of magneto-hydrodynamic (MHD) convection and resultant dynamo activities in proto-neutron stars (PNSs), we construct a "PNS in a box" simulation model with solving compressible MHD equation coupled with a nuclear equation of state (EOS) and a simplified leptonic transport. As a demonstration, we apply it to two types of PNS models with different internal structures: fully-convective model and spherical-shell convection model. By varying the spin rate of models, the rotational dependence of convection and dynamo that operate inside the PNS is investigated. We find that, as a consequence of turbulent transport by rotating stratified convection, large-scale structures of flow and thermodynamic fields are developed in all models. Depending on the spin rate and the convection zone depth, various profiles of the large-scale structures are obtained, which can be physically understood as steady-state solutions to the "mean-field" equation of motion. Additionally to those hydrodynamic structures, the large-scale magnetic component with $\mathcal{O}(10^{15})$ G is also spontaneously organized in disordered tangled magnetic fields in all models. The higher the spin rate, the stronger the large-scale magnetic component is built up. Intriguingly, as an overall trend, the fully-convective models have a stronger large-scale magnetic component than that in the spherical-shell convection models. The deeper the convection zone extends, the larger the size of the convection eddies becomes. As a result, the rotationally-constrained convection seems to be more easily achieved in the fully-convective model, resulting in the higher efficiency of the large-scale dynamo there. To gain a better understanding of the origin of the diversity of NS's magnetic field, we need to study the PNS dynamo in a wider parameter range.

We argue that the neutrino halo, a population of neutrinos that have undergone direction-changing scattering in the stellar envelope of a core-collapse supernova (CCSNe), is sensitive to neutrino emission history through time of flight. We show that the constant time approximation, commonly used in calculating the neutrino halo, does not capture the spatiotemporal evolution of the halo neutrino population and that correcting for time of flight can produce conditions which may trigger fast neutrino flavor conversion. We also find that there exists a window of time early in all CCSNe where the neutrino halo population is sufficiently small that it may be negligible. This suggests that collective neutrino oscillation calculations which neglect the Halo may be well founded at sufficiently early times.