People increasingly use multiple Multimodal Large Language Models (MLLMs) concurrently, selecting each based on its perceived strengths. This cross-platform practice creates coordination challenges: adapting prompts to different interfaces, calibrating trust against inconsistent behaviors, and navigating separate conversation histories. Prior HCI research focused on single-agent interactions, leaving multi-MLLM orchestration underexplored. Through a diary study and semi-structured interviews (N=10), we examine how individuals organize work across competing AI systems. Our findings reveal that users construct primary and secondary hierarchies among models that shift over usage context. They also develop personalized switching patterns triggered by task aggregation to adjust effort and latency, and output credibility. These insights inform future tool design opportunities, supporting users to coordinate multi-MLLM workflows.

We investigate the Hamiltonian formulation of 1+1-dimensional staggered fermions and reconstruct the vector and axial charge operators, originally identified by Arkya Chatterjee et al., within the Wilson fermion formalism. These operators commute with the Hamiltonian and reduce, in the continuum limit, to the generators of the vector and axial $\mathrm{U}(1)$ symmetries. A notable feature of the axial charge operator is that it acts locally on operators and possesses quantized eigenvalues. Its eigenstates can therefore be interpreted as fermion states with well-defined integer chirality, analogous to those in the continuum theory. This structure enables the formulation of a gauge theory in which the axial $\mathrm{U}(1)_A$ symmetry is promoted to a gauge symmetry. We construct a Hamiltonian in terms of the eigenstates of the axial charge operator, thereby preserving exact axial symmetry on the lattice while recovering vector symmetry in the continuum limit. As applications, we study the implementation of the Symmetric Mass Generation (SMG) mechanism in the 3-4-5-0 models. Our framework admits symmetry-preserving interaction terms with quantized chiral charges, although further numerical investigation is required to confirm the realization of the SMG mechanism in interacting systems.

To achieve more controllable development of coherence in solids, we investigated the effect of a trigger laser tuned to the superradiance transition wavelength on periodic superradiance observed in an Er:YSO crystal. For period control, applying the trigger laser reduced both the superradiance period and its variance, demonstrating enhanced controllability of coherence development dynamics. As the trigger laser power increased, both the period and the number of emitted superradiance photons decreased while maintaining a proportional relationship. This behavior is explained by a reduced superradiance threshold under a constant excitation rate and is reproduced by numerical simulations based on the Maxwell-Bloch equations. For timing control, we found that superradiance could be triggered even when the excitation laser alone was insufficient. This enabled us to control the emission timing of superradiance using short trigger pulses and provided a device capable of generating superradiance at desired timing.

We consider a non-contextual inequality in the sequential measurement scenario and derive the optimal quantum violation of it without assuming the dimension of the system. Since the measurement is dichotomic and the dimension of the quantum system is arbitrary, we formulate the concept of degeneracy-breaking (DB) measurement depending on how many projectors are being used in the sequential measurement. We demonstrate that by increasing the number of projectors involved in the sequential measurement (thereby making the measurement more degeneracy breaking) the quantum violation of non-contextual inequality can be enhanced and can even reach up to its algebraic maximum. We demonstrate that the optimal quantum violations for different number of projectors serves as a quantum dimension witness.

A reconfigurable intelligent surface (RIS)-assisted non-orthogonal multiple access (NOMA) system is investigated, where the transmitter (Alice) is a dual functional radar communication (DFRC) base station (BS) that aims to sense the location of a potential warden (Willie), while simultaneously transmitting public and covert signals to the legitimate users, Carol and Bob, respectively. Both cases of known and unknown Willie locations are considered. For the known-location case, assuming perfect channel state information (CSI) at Willie, a covert rate maximization is formulated with the joint optimization of active and passive beamforming, which is solved using successive convex approximation (SCA), penalty method, and semidefinite relaxation (SDR). For the unknown-location case, we propose to estimate Willie's location via radar sensing and develop a sensing-based imperfect CSI model. In particular, the CSI error uncertainty is bounded by the sensing accuracy, which is characterized by the Cramer-Rao bound (CRB). Subsequently, a robust communication rate maximization problem is formulated under the constraints on quality-of-service (QoS) of Carol, sensing accuracy, and covertness level. The Schur complement and S-procedure are employed to handle the non-convex constraints. Numerical results compare the system performance under the two cases, and demonstrate the significant covert performance superiority of the sensing-based imperfect CSI model and NOMA over the general norm-bounded imperfect CSI model and the orthogonal multiple access scheme. Furthermore, the dual yet contradictory effects of sensing on covert communications are revealed. It is also found that Alice primarily utilizes Carol's signal for sensing, while allocating almost all of Bob's signal for communication.

We investigate the low-energy phenomenology of the Type-I seesaw mechanism within a 3+3 framework containing three active and three sterile neutrinos. Using the exact seesaw relation as a bridge between the high-scale sterile-sector parameters and the standard oscillation observables, we perform a comprehensive Monte Carlo scan of the 21-dimensional sterile parameter space, retaining only those configurations consistent with current neutrino oscillation data within $3σ$. For the viable parameter points, we simulate the modified neutrino oscillation probabilities and event rates at the long-baseline experiments DUNE and NO$ν$A, and the medium-baseline reactor experiment JUNO, quantifying their sensitivity to sterile neutrino effects across the eV--GeV mass range. We find that eV-scale sterile neutrinos produce pronounced spectral distortions, while heavier states decouple progressively from oscillation experiments. In parallel, we confront the seesaw predictions with complementary probes: cosmological bounds on $\sum m_i$, the kinematic mass $m_β$ from beta decay, the effective Majorana mass $|m_{ββ}|$ from neutrinoless double beta decay ($0νββ$), and the charged-lepton-flavor-violating branching ratio $\text{BR}(μ\to eγ)$. The combination of all constraints significantly narrows the allowed parameter space: the predicted sum of neutrino masses clusters at $\sum m_i \sim 0.05$--$0.07$~eV, within reach of next-generation cosmological surveys, and eV-scale sterile neutrinos are found to be under significant tension from the current MEG bound on $μ\to eγ$.

Web-augmented large language models (LLMs) offer promising capabilities for automatic code generation. However, integrating live web search exposes models to unreliable or malicious content, leading to Search-Induced Issues (SII), a novel failure mode in which external pages mislead LLMs into producing incorrect code. This paper presents a comprehensive empirical study of the prevalence and impact of SII across three commercial search APIs and six advanced LLMs. Our analysis reveals that all evaluated web-augmented LLMs are vulnerable to SII, with root causes arising from either misaligned specifications or flawed code implementations in the searched Error-Inducing Pages (EIPs). To address this challenge, we propose Sherlock, an automated framework that enables LLM service providers to proactively safeguard web-augmented generation systems at scale. Sherlock operates as a continuous pipeline that first detects potential SII instances, then debugs them to identify the responsible EIPs and pinpoint their root causes, and finally repairs them by either annotating misaligned content or replacing erroneous code snippets with evaluated solutions from trusted sources. Experiments show that Sherlock identifies EIPs with an F1 score of up to 95% and repairs 71% to 100% of affected generations across the evaluated models, with modest computational overhead. Our findings and framework provide practical guidance for improving the reliability of web-augmented LLM-based code generation systems in real-world software engineering scenarios.

We develop a nonperturbative tensor-network framework for computing cosmological correlators in de Sitter space and use it to test the proposal that suitably defined in-in correlators can be obtained from an in-out formalism by gluing the expanding and contracting Poincaré patches. Focusing on interacting $1+1$-dimensional $ϕ^4$ theory, we formulate finite-time lattice observables using Matrix Product State (MPS) techniques and analyze the regulator subtleties associated with the singular behavior near the patching surface. Within this regulated framework, we find controlled nonperturbative evidence for the proposed relation between in-in and in-out correlators in several examples. We also find suggestive evidence that the perturbative obstructions present for sufficiently light fields can be softened nonperturbatively, albeit in a regime of substantially larger entanglement. A central outcome of our analysis is an entanglement-based picture of the computation: for in-in evolution the entanglement remains modest and can decrease toward late times, whereas in the patched in-out set-up it grows significantly after the gluing slice. Thus, although the in-out formalism is perturbatively economical, the in-in formulation is numerically more favorable. We briefly discuss how the same strategy extends to low-angular-momentum sectors in $3+1$ dimensions, and why regimes of rapid entanglement growth may eventually motivate quantum-computing implementations.

Massive multiple-input multiple-output (MIMO) has enabled substantial spatial multiplexing and array gains in real-world systems, while distributed MIMO (D-MIMO) improves macro-diversity over wide areas at the cost of deployment complexity. Repeater-assisted massive MIMO (RA-MIMO) is a lower-cost alternative that can recover key distributed-MIMO advantages. This paper asks whether repeater assistance can also enhance frequency diversity. We study an uncoded discrete Fourier transform-spread orthogonal frequency-division multiplexing (DFT-s-OFDM) uplink with one-tap single-carrier frequency-domain equalization (SC-FDE) based on minimum mean-square error (MMSE) and derive a receiver-matched semi-analytic bit-error rate (BER) expression by averaging over channel and interference realizations, without Gaussian approximation of residual despreading interference. The analysis clarifies how repeater delay reshapes frequency correlation, and waveform simulations confirm tight agreement with the derived expression together with improved high-signal-to-noise ratio (SNR) BER decay, highlighting delay as a practical tuning knob.

In this work, we develop a rigorous theoretical framework for the evolution of linear viscoelastic properties across the sol-gel transition. More specifically, we derive general admissible expressions for the relaxation modulus and dynamic moduli as the critical gel state is approached from the pre-gel or the post-gel side. These expressions possess a generalized multi-mode series representation and recover the critical gel power law spectrum in the limit of vanishing distance from the gel point. We validate these expressions against the experimental data for various polymeric and colloidal systems. A central finding of the present work is the requirement of continuity of the dynamic moduli and their derivatives at the critical gel point, which imposes a profound physical constraint, necessitating the relaxation dynamics on both sides of the transition to be symmetric. This, in turn, leads to the hyper-scaling relation, which is a theoretical requirement rather than an empirical proposal. We further show that the critical relaxation exponent (n) always remains above the relaxation scaling exponent (κ), establishing a previously unrecognized lower bound on n. We also analytically estimate, for the first time, the parameter C that characterizes the relative evolution of the storage modulus with respect to the loss modulus as the critical state is approached. These results reveal that the symmetry, scaling, and hyperscaling properties of the sol-gel transition are all consequences of a single unifying physical requirement originating from the continuity of the linear viscoelastic properties at the critical gel point.

A gradient-based method is proposed for solving the linear quadratic regulator (LQR) problem for linear systems with nonlinear dependence on time-invariant probabilistic parametric uncertainties. The approach explicitly accounts for model uncertainty and ensures robust performance. By leveraging polynomial chaos theory (PCT) in conjunction with policy optimization techniques, the original stochastic system is lifted into a high-dimensional linear time-invariant (LTI) system with structured state-feedback control. A first-order gradient descent algorithm is then developed to directly optimize the structured feedback gain and iteratively minimize the LQR cost. We rigorously establish linear convergence of the gradient descent algorithm and show that the PCT-based approximation error decays algebraically at a rate $O(N^{-p})$ for any positive integer $p$, where $N$ denotes the order of the polynomials. Numerical examples demonstrate that the proposed method achieves significantly higher computational efficiency than conventional bilinear matrix inequality (BMI)-based approaches.

The orbital dynamics of magnons have recently drawn interest due to their potential roles in thermal and orbital transport phenomena in magnetic insulators. In this study, we investigate the orbital magnetic moment (OMM) and orbital angular momentum (OAM) of magnons in a Kagome antiferromagnet with negative vector chirality, focusing on the distinction between thermodynamic and wave-packet-based definitions. We compute the Berry curvature, the OMM, and the OAM in momentum space under an external magnetic field. Our results reveal a quantitative difference between the OMM and OAM, yet their associated Nernst coefficients exhibit similar temperature and field dependence in transport. Our results provide a quantitative comparison between the thermodynamic and wave-packet formulations of magnon orbital dynamics.

Although stellar bars are prevalent in local galaxies, unbarred galaxies constitute a significant fraction, particularly at high redshifts. While some galaxies are unbarred by nature due to stability against the bar instability, several mechanisms capable of transforming barred galaxies into unbarred systems have also been proposed, such as central mass concentration, specific dark matter halo morphologies and tidal interactions. Regarding galactic interactions, mergers can undoubtedly disrupt bars while potentially destroying the entire disk. However, the effects of pure collisions (non-merging scenarios) on bars remain unclear, with limited existing studies yielding contradictory conclusions. Here we aim to systematically investigate the disruptive effects of collisions on bars hosted by Milky Way-like galaxies using N-body/SPH simulations. We model collisions between the barred galaxy and a spherical intruder, conducting multiple simulations by varying interaction parameters, with mass ratios set at 1:3, 1:5, and 1:15. We find that bars are remarkably robust, with most interactions failing to significantly reduce their strength or pattern speed. Only off-center high-inclination retrograde collisions can effectively destroy bars, while central high-inclination collisions can substantially decrease the pattern speed. Such destruction and deceleration primarily arise from gravitational forces rather than gas-related processes. Notably, compared to collisions occurring along the bar's major axis, those along the minor axis cause greater weakening but can slow the bar's natural deceleration. Furthermore, changes in mass resolution do not significantly affect the results when the resolution is better than ~10^5 Solar mass.

We present a generalized Ernst-type framework for stationary, axisymmetric spacetimes in which a scalar field is coupled to the electrodynamic field, with a particular focus on the ModMax theory. Our approach relies on the Weyl stationary-axisymmetric ansatz and explicitly allows for a nonzero rotational metric function, $ω\neq 0$. The resulting setup is broad enough to encompass wide classes of scalar couplings, including dilatonic and phantom-like sectors, and can be tailored to specific models such as Einstein-ModMax, Kaluza-Klein theories, low-energy string-inspired scenarios, entanglement relativity and related generalizations. Within this scheme, we derive two novel families of exact rotating solutions in the sector where the electromagnetic invariants obey $\mathcal F/\mathcal G=\mathrm{constant}$. This regime is particularly significant for ModMax, as it preserves genuinely nonlinear features while still admitting an analytically manageable description.

Non-terrestrial networks (NTNs) are considered a promising technology for seamless, universal communication in the 6G era. However, signals from NTN elements to ground users are often blocked by high-rise buildings in dense urban environments. To quantify this blocking effect, in this paper, we propose a novel analytical framework by modeling the location of buildings as a 3D skyline process based on stochastic geometry and we derive closed-form expressions for the distribution of the blockage elevation angles. Furthermore, we extend our analysis to include spatially correlated blockage effects and analyze the spectral properties of the Skyline process using power spectral density (PSD) and autocorrelation function (ACF). Based on these theoretical findings, we present numerical results that provide insights into the design of LEO satellite networks by computing the mean number of visible satellites and the outage probability. We provide a decorrelation angle that serves as a useful threshold for obtaining the satellite diversity gain. These findings provide first analytical steps toward designing and user connection strategies to NTNs in urban environments.

We analytically sum the leading bubble diagrams that contribute to the elastic scattering amplitude of a kink and a meson in the $ϕ^4$ double-well model. We find a single peak, corresponding to the unstable kink state in which the shape mode is excited twice. The peak has the usual Breit-Wigner form, and its imaginary part agrees with the shape mode decay rate found by Manton and Merabet.

We develop a preference elicitation method for a Von Neumann-Morgenstern (VNM)-type decision-maker from pairwise comparison data in the presence of response errors. We apply the maximum likelihood estimation (MLE) method to jointly elicit the non-parametric systematic VNM utility function and the scale parameter of the response error, assuming a Gumbel distribution. We incorporate structural preference information known in advance about the decision-maker's risk attitude through linear constraints on the utility function, including monotonicity, concavity, and Lipschitz continuity. Under discretely distributed lotteries, the resulting MLE problem can be reformulated as a convex program. We derive finite-sample error bounds between the MLE and the true parameters, and establish quantitative convergence of the MLE-based VNM utility function to the true utility function in the sense of the Kolmogorov distance under some conditions on the lotteries. These conditions may have potential applications in the design of efficient lotteries for preference elicitation. We further show that the optimization problem maximizing the expected MLE-based VNM utility is robust against the response error and estimation error in a probabilistic sense. Numerical experiments in a portfolio optimization application illustrate and support the theoretical results.

Quantum error mitigation is essential for extracting trustworthy results from noisy intermediate-scale quantum (NISQ) processors. Yet, current approaches face a core scalability bottleneck: unbiased methods such as probabilistic error cancellation (PEC) incur exponential sampling overhead, while approximate techniques like zero-noise extrapolation trade accuracy for efficiency. We introduce and experimentally demonstrate MoSAIC (Modular Spatio-temporal Aggregation for Inverted Channels), a scalable quantum error mitigation framework that preserves the unbiasedness of PEC while dramatically reducing sampling costs. MoSAIC partitions a circuit into noise-aligned blocks, learns an effective block noise model using classical variational optimization, and applies quasi-probabilistic inversion once per block instead of after every layer. This blockwise aggregation reduces both sampling overhead and circuit-depth overhead, enabling mitigation far beyond the operating regime of standard PEC. We also experimentally validate MoSAIC on IBM's 156-qubit Heron processors, performing the largest PEC-based mitigation demonstration on hardware to date. As a physically meaningful benchmark, we prepare the critical one-dimensional transverse-field Ising (TFIM) ground state for system sizes up to 50 qubits. We show that MoSAIC can achieve at least 1 to 2 orders of magnitude better accuracy than standard PEC under identical sampling budgets. This enables MoSAIC to recover accurate observables for larger system sizes, even when standard PEC fails due to its prohibitive sampling overhead. We also present CUDA-Q accelerated simulations to validate performance trends under a range of different noise models. These results demonstrate that MoSAIC is not only theoretically scalable but also practically deployable for high-accuracy, large-scale quantum experiments on today's quantum hardware.

We introduce an iterative scheme for discrete convex minimization problems of $p$-Laplace type such as variational graph $p$-Laplace problems and $\ell^p$ regression. In each iteration, the scheme solves only a weighted least-squares problem. We verify linear convergence for suitably regularized problems and derive convergence to any prescribed tolerance.

Large language models (LLMs) exhibit substantial potential across diverse scientific disciplines, including materials science. A property prediction framework, ZEBRA-Prop (Zero-Shot Embedding-Based Rapid and Accessible Regression Model for Materials Properties), is presented here as an extension of LLM-Prop. In contrast to LLM-Prop, which requires task-specific fine-tuning of the LLM, ZEBRA-Prop eliminates fine-tuning, thereby reducing computational cost and enabling rapid model training. The framework employs MatTPUSciBERT, an LLM specialized for materials science, to enhance predictive capability. Multiple textual embeddings are incorporated through a learnable weighting mechanism, which alleviates the context-length constraints inherent in LLM-Prop and facilitates effective integration of diverse textual representations. Evaluation is conducted using two datasets: the TextEdge dataset (approximately 140,000 entries) and an in-house dataset (approximately 2,000 entries) derived from the Materials Project database, with physical properties obtained from first-principles calculations. The predictive performance of ZEBRA-Prop is close to that of LLM-Prop, while the training time is reduced by approximately 95%. The performance improvements are attributable to three principal factors: domain-specific LLM utilization, diversified textual descriptions, and systematic text preprocessing. ZEBRA-Prop constitutes a scalable and computationally efficient framework for materials property prediction and supports accelerated materials discovery, particularly under limited computational resources.

In the present paper, we introduce and analyze elephant random walks (ERWs) on bipartite periodic lattices arising as coverings of dipole graphs. We focus on lattices whose admissible step directions in the two parts of the bipartition are negatives of each other and disjoint. On such graphs, we define an ERW in which each step is chosen by referring to the entire history of the walk. The ERW on the hexagonal lattice is a prototypical example of our model. The definition and asymptotic analysis of such ERWs are not straightforward because both depend strongly on the underlying geometric structure. Our analysis is based on a combination of the Pólya-type urn techniques and the martingale approach, two standard methods for analyzing ERWs. We find that the counting process of the ERW forms a Pólya-type urn with two-periodic generating matrices. By analyzing for such urn models, we show the strong law of large numbers for the counting process. Combining the result for the counting process with the martingale approach, we derive non-standard strong laws of large numbers and central limit theorems for the position process of the ERW in the diffusive and critical regimes, as well as almost sure and $L^2$ scaling limits in the superdiffusive regime.

The Weil representation is a particularly significant linear representation of the metaplectic group, used in the study of theta correspondence. In this paper, I introduce a derived category version of the Weil representation in the local field case. For the dual pair $ (\mathrm{GL}_n,\mathrm{GL}_m) $, I give a coherent description of this category, in the philosophy of relative Langlands duality.

We investigate the relationship between two interpretations of equivariant Riemann-Roch defects of complex spaces with conic singularities; as (i) equivariant $η_{T}$ and $ξ_{T}$ invariants, and as (ii) supertraces over local cohomology groups. This leads to a novel threefold partitioning of the $L^{2}$-spinor space on the link and a corresponding splitting of $ξ_{T}$. Two partitions correspond to cohomological contributions coming from $\bar\partial$-Neumann and $\bar\partial$-Dirichlet operators on the cone, while the third partition makes no contribution to the equivariant index defect, which we show is due to supersymmetric cancellations on the cone that we call lifted supersymmetry. We use this to define complex equivariant $ξ_T$ and $η_T$ invariants, which are equivalent to the usual invariants but are easier to compute. We highlight connections to related algebraic and analytic descriptions of Riemann-Roch defects in the literature, both at the level of numbers and their categorifications, and explore connections to existing notions of supersymmetric cancellations in physics and mathematics.

We study logit-based multi-purchase choice models and develop an exact solution methodology for the resulting assortment optimization problems, which we show are NP-hard to approximate. We introduce a hypergraph representation that captures general bundle-based choice structures and subsumes several models in the literature, including the BundleMVL-K and multivariate MNL models (Tulabandhula et al. 2023, Jasin et al. 2024). Leveraging this representation, we derive mixed-integer programming (MIP) formulations by integrating polyhedral relaxations from multilinear optimization with a perspective reformulation of the logit choice model. Our approach preserves the strength of the underlying polyhedral relaxations, yielding formulations with provably tighter linear programming (LP) bounds than the prevalent Big-M approach. We further characterize structural conditions on the hypergraph under which the formulations are locally sharp, thereby generalizing existing LP characterizations for path-based models. The framework extends naturally to heterogeneous and robust settings. Computational experiments demonstrate that the proposed formulations significantly improve both solution quality and scalability.

Modern multicore system-on-chips (SoCs) share off-chip DRAM across cores, where bank-level interference can significantly degrade performance and threaten real-time guarantees. While prior work has focused on per-core bandwidth regulation, these approaches treat main memory as a monolithic resource and overlook DRAM's inherent bank-level parallelism. We show that DRAM interference is fundamentally a bank-level phenomenon. We characterize the guaranteed bandwidth of modern DRAM, demonstrate that it remains effectively constant across generations, and show how this limitation can be exploited by single-bank attacks. These results highlight the need for bank-aware memory management for predictable and efficient real-time systems. We design and implement a novel per-bank memory bandwidth regulator in an open-source RISC-V SoC and evaluate it using FireSim with both synthetic and real-world workloads. Our evaluation demonstrates that per-bank regulation effectively mitigates adversarial bank contention and achieves a 5.74x average throughput improvement for best-effort workloads over traditional bank-oblivious approaches while providing the same-level of performance isolation guarantees for real-time workloads.

Unlike the von Neumann architecture, which separates computation from memory, the brain tightly integrates them, an organization that large language models increasingly resemble. The crucial difference lies in the ratio of energy spent on computation versus data access: in the brain, most energy fuels compute, while in von Neumann architectures, data movement dominates. To capture this imbalance, we introduce the \emph{operation-operand disjunction constant} $G_d$, a dimensionless measure of the energy required for data transport relative to computation. As part of this framework, we propose the metaphor of \emph{data gravity}: just as mass exerts gravitational pull, large and frequently accessed data sets attract computation. We develop expressions for optimal computation placement and show that bringing the computation closer to the data can reduce energy consumption by a factor of $G_d^{(β- 1)/2}$, where $β\in (1, 3)$ captures the empirically observed distance-dependent energy scaling. We demonstrate that these findings are consistent with measurements across processors from 45\,nm to 7\,nm, as well as with results from processing-in-memory (PIM) architectures. High $G_d$ values are limiting; as $G_d$ increases, the energy required for data movement threatens to stall progress, slowing the scaling of large language models and pushing modern computing toward a plateau. Unless computation is realigned with data gravity, the growth of AI may be capped not by algorithms but by physics.

We formulate the Gao-Jafferis-Wall traversable wormhole protocol as a quantum channel and compute its quantum channel capacity. We show that this capacity is governed by the time derivative of an out-of-time-ordered correlator, hence by operator size growth in the holographic dual, and that its growth is bounded above by the Einstein gravity limit. The channel capacity therefore provides a natural benchmark for quantum simulations of traversable wormholes.

Heating, ventilation, and air conditioning (HVAC) systems account for a substantial share of building energy consumption. Environmental uncertainty and dynamic occupancy behavior bring challenges in decarbonized HVAC control. Reinforcement learning (RL) can optimize long-horizon comfort-energy trade-offs but suffers from exponential action-space growth and inefficient exploration in multi-zone buildings. Large language models (LLMs) can encode semantic context and operational knowledge, yet when used alone they lack reliable closed-loop numerical optimization and may result in less reliable comfort-energy trade-offs. To address these limitations, we propose a hierarchical control framework in which a fine-tuned LLM, trained on historical building operation data, generates state-dependent feasible action masks that prune the combinatorial joint action space into operationally plausible subsets. A masked value-based RL agent then performs constrained optimization within this reduced space, improving exploration efficiency and training stability. Evaluated in a high-fidelity simulator calibrated with real-world sensor and occupancy data from a 7-zone office building, the proposed method achieves a mean PPD of 7.30%, corresponding to reductions of 39.1% relative to DQN, the best vanilla RL baseline in comfort, and 53.1% relative to the best vanilla LLM baseline, while reducing daily HVAC energy use to 140.90~kWh, lower than all vanilla RL baselines. The results suggest that LLM-guided action masking is a promising pathway toward efficient multi-zone HVAC control.

Given a pair $(V,C)$ that admits a stable minimal model with fixed dimension, fixed coefficient set, and bounded relative volume, we study when the linear system $|m(K_V+C)|$ induces an Iitaka fibration, assuming the Iitaka volume of $K_V+C$ is sufficiently large.

Observation of charged lepton flavor violation would constitute unambiguous evidence for physics beyond the Standard Model (SM). We identify a previously unexplored same-sign dimuon signature in electron-photon collisions, $γe^- \to e^+μ^-μ^-$, mediated by an axionlike particle (ALP) with flavor-violating $e$-$μ$ couplings. The absence of irreducible SM backgrounds and the on-shell production of the ALP render this channel intrinsically clean and highly sensitive, with only small residual backgrounds arising from detector effects. Such collisions can be realized via laser Compton backscattering at $e^+e^-$ colliders including BEPC-II with the BESIII detector, STCF, and ILC. We find that STCF and ILC can probe couplings one to two orders of magnitude below existing bounds. This combination of resonant production, vanishing irreducible background, and same-sign topology would be difficult to achieve in conventional $e^+e^-$ or hadron-collider environments, establishing electron-photon collisions as a uniquely powerful probe of charged lepton flavor violation.