Papers

Space-air-ground integrated networks (SAGINs) integrate heterogeneous platforms with different coverage, mobility, and payload constraints, which creates strong demands for flexible antenna architectures. Foldable antenna arrays (FAAs) provide a promising solution by reconfiguring array geometry, antenna positions, and orientations through controllable joints or hinges. However, existing studies on FAAs have mainly considered simple folding configurations and limited deployment benefits, while FAAs' architectures and broader potential for SAGIN remain largely unexplored. In this article, we present the fundamentals and architectures of FAA, including folding mechanisms, foldability levels, and functional capabilities. We then examine FAA applications for ground base stations, low-altitude platforms, high-altitude platforms, and satellites, together with their potential challenges and solutions. Finally, we highlight open issues and future research directions, and provide a case study to demonstrate the FAA-enabled communication performance improvement in SAGIN.

Cross-chain protocols enable applications to exchange messages across blockchains. Under point-to-point configurations, communication depends on a direct connection between the source and destination blockchains, limiting blockchain reachability and requiring additional configurations to connect more blockchains. To quantify this problem, this paper analyzes approximately 25 million mainnet cross-chain transaction events collected from six protocols (Axelar, CCIP, Hyperlane, LayerZero, Relay, and Wormhole) between January and October 2025. The resulting graph covers 286 active blockchains and 11,935 directly connected ordered blockchain pairs. These connections provide a direct reachability of 14.64%, while full direct connectivity would require 81,510 point-to-point configurations. We present XIR, a framework for interoperability across cross-chain protocols based on a verifiable intermediate representation. This representation binds an application message to an ordered record of authenticated cross-chain protocol deliveries, preserving message identity and verification history across protocol boundaries. XIR Gateways and XIR Adapters use this representation to compose existing connections into same-protocol and cross-protocol multi-hop paths. We implement an XIR prototype integrating Hyperlane and LayerZero and evaluate it in local and public-testnet environments. Theoretical analysis and evaluation show that, with correctly configured cross-chain protocol connections, XIR avoids 67,018 additional point-to-point configurations, equivalent to 84.88% of the total required by a point-to-point configuration baseline serving the same reachable pairs, and increases reachability from 14.64% to 96.86% of all ordered blockchain pairs.

Computing is evolving rapidly to cater to the increasing demand for sophisticated services, and Cloud computing lays a solid foundation for flexible on-demand provisioning. However, as the size of applications grows, the centralised client-server approach used by Cloud computing increasingly limits the applications' scalability. To achieve ultra-scalability, cloud/edge/fog computing converges into the compute continuum, completely decentralising the infrastructure to encompass universal, pervasive resources. The compute continuum makes devising applications benefitting from this complex environment a challenging research problem. We put the opportunities the compute continuum offers to the test through a real-world multi-view detection model (MvDet) implemented with the FastFL C/C++ high-performance edge inference framework. Computational performance is discussed considering many experimental scenarios, encompassing different edge computational capabilities and network bandwidths. We obtain up to 1.92x speedup in inference time over a centralised solution using the same devices.

Gromov–Wasserstein optimal transport (GW-OT) extends classical optimal transport by introducing structure-aware transport cost. This is particularly relevant for spatial transcriptomics, where dynamical reconstruction should preserve tissue structure in addition to matching expression patterns. While static formulations have been widely used for such structure-aware alignment, a general dynamic formulation for reconstructing continuous trajectories is still missing. We introduce Travelling Pair Dynamical Alignment and Trajectory Estimation (TP-DATE), a theoretical and computational framework to generalize GW-OT dynamically in a simulation-free manner. We formulate a broad class of static and dynamic Quadratic-form OT (QOT) through path actions and prove the static dynamic equivalence. We further develop travelling-pair flow matching, which allows interacting conditional paths and marginalizes their interactions into a single vector field. On synthetic and real spatial transcriptomics data, TP-DATE better preserves spatial structure and improves continuous 3D dynamics reconstruction.

A covariance matrix is often used as a compact surrogate for the statistical model of a measurement. We show that this replacement can change not only a fitted value or its uncertainty, but also the likelihood-ratio ordering of the experiment itself. In a tractable correlated-measurement model, the data separate into an informative component and an ancillary residual disagreement. Exact likelihood inference therefore does not use that disagreement, whereas data-dependent covariance matrices and Gaussian reconstructions after nonlinear transformations can reintroduce it into parameter inference. We quantify the resulting change by the probability mass of the symmetric difference of equally calibrated acceptance regions - the fraction of repeated experiments for which the confidence decision about a tested parameter value changes. In the small-uncertainty regime the ordering discrepancy is generically first order in the total relative uncertainty, whereas conventional and ancillary-conditioned coverage defects begin at second order. At 'coverage-blind' points the leading coverage difference vanishes while the ordering discrepancy remains nonzero. In a representative blind case with 10% total relative uncertainty, about 7.6% of experiments change their confidence decision despite exact calibration of both procedures. The same mechanism connects Peelle's Pertinent Puzzle to the classical literature on ancillarity and relevant subsets: a covariance approximation can manufacture inferential relevance for a goodness-of-fit statistic that is ancillary in the true model. We show directly how this changes the confidence interval reported for the same informative content, and extend the analysis beyond equal statistical uncertainties and beyond the two-measurement case.

The Alternating Direction Method of Multipliers (ADMM) is widely used for distributed optimization, but its synchronous implementation can suffer from efficiency loss in heterogeneous edge networks, where fast clients or groups need to wait for slower ones before global updates can be completed. Existing group-based ADMM methods reduce communication overhead through grouping, but their grouping rules usually focus on data similarity or network topology and do not explicitly account for computation heterogeneity. To address this issue, this paper proposes Parallel Group-Based ADMM (P-GADMM) for distributed optimization in heterogeneous edge networks. P-GADMM forms computation-aware edge groups according to client computational capabilities and local data sizes, which reduces training-speed variation within each group. It further combines edge-level aggregation with bounded asynchronous coordination at the cloud, allowing active groups to participate in global updates without waiting for slower groups while controlling stale group information through a delay threshold. For strongly convex group objectives, we establish convergence guarantees for an idealized form of P-GADMM under bounded group-level staleness, showing a time-averaged convergence behavior up to a staleness-induced asymptotic error neighborhood. Experiments show that P-GADMM reduces wall-clock training time compared with representative baselines while maintaining comparable final accuracy.

People increasingly turn to AI chatbots for news and explanations of world events. But do they receive the same political answers when they ask in different languages? Here we show that the language of a question can change how the same AI systems assess the war in Ukraine. We ask GPT, Claude and Gemini to evaluate twenty statements about the war in 112 languages, collecting 67,200 responses. The balance between Russia-leaning and Ukraine-leaning responses differs across languages. When we group responses by countries' official languages, they follow a pattern resembling worldwide political divisions: relatively more Russia-leaning answers correspond to more favourable public views of Russia, less support for Ukraine in United Nations votes, and less aid to Ukraine. The broad pattern recurs across all three models and remains when individual statement pairs are removed. Our findings suggest a possible route through which information warfare may shape the text used to train AI models, which may in turn spread geopolitical biases.

Reward-based reinforcement learning for language models, exemplified by Group Relative Policy Optimization (GRPO), collapses an entire stochastic trajectory into a single scalar reward. This is clean and scalable, but it explores and allocates reward inefficiently: a trajectory may contain many causal decisions, recovery attempts, and environment-randomness events, yet every token or action inherits one trajectory-level advantage. We study tree-based rollout construction as a compute-allocation problem for policy-gradient estimation. Our central claim is that branches should be placed not where the policy is merely uncertain, but where an additional branch most reduces uncertainty about the policy gradient per unit of compute. From a law-of-total-variance decomposition of the local policy-gradient random variable, we derive two allocation laws: new branches reduce decision uncertainty, while repeated suffix rollouts reduce continuation uncertainty. The resulting EPIG-Tree score allocates branches using the already computed rollouts. It estimates occupancy- and score-weighted value uncertainty, along with a suffix law $n_e \propto w_e \|\nabla_\theta \log \pi(a_e|h_e)\| \sigma_e / \sqrt{c_e}$. Empirically, EPIG reduces gradient MSE in cloned-state control, winning in all nine dense continuous-control environments of a 13-environment sweep and recovering the reference gradient direction near-perfectly, and it improves frozen-LLM gradient calibration relative to entropy branching. In online single-turn math, tree-local credit beats flat GRPO, while branch placement is secondary to token-level credit assignment. In online multi-turn Wordle, EPIG attains the highest final win rate (0.850), overtaking flat GRPO, which saturates early at 0.790, and entropy branching as training proceeds, confirming that the gradient-estimation advantage transfers to a stateful, large-action setting.

In 2025, the affine dual curvature measures $\widetilde{C}_m^a(K,\cdot)$ of convex body $K$ were introduced by Cai, Leng, Wu, and Xi, and the even Minkowski problem for the affine dual curvature measures was solved. In this paper, the necessary and sufficient condition of affine dual Minkowski problem for general measures with $m>1$ are proposed.

Let $F$ be the unramified quadratic extension of $\mathbb{Q}_p$, and let $P_n(T)$ denote the polynomials arising in the $p$-adic Fourier theory of Schneider and Teitelbaum. We determine explicitly the norms of the functions $P_n(x\Omega)$ on the Banach spaces of locally $F$-analytic functions of fixed order, where $\Omega$ is a Lubin–Tate period. Our computation is based on a recent valuation formula of Ardakov and Berger for the values $P_n(\Omega)$. As an application, we obtain an explicit normalization of the basis constructed by Bannai and Kobayashi, whose elements all have norm one.

Automated video interview assessment integrates verbal content, acoustic delivery, and visual behavior, yet numerical predictions alone provide limited inspectable support. We present E-AVI, an evidence-grounded framework that extracts timestamped multimodal evidence and integrates dimension-conditioned evidence attention with source-level embeddings for scoring. A shared evidence pool further supports natural-language feedback and follow-up question answering. On RecruitView and a private hospitality dataset, E-AVI consistently outperforms fine-tuned multimodal baselines in rank correlation. Ablation, evidence-deletion, bootstrap, human-audit, and QA analyses characterize the predictive contribution, grounding, and practical utility of the evidence pathway. Together, these results demonstrate that our proposed E-AVI framework improves predictive performance while providing inspectable support for assessment, feedback, and interactive analysis.

Modern tabular datasets are becoming increasingly large, both in the number of samples and covariates, posing significant challenges for Bayesian variable selection due to the resulting computational burden. While there is extensive literature on scaling Bayesian inference to large numbers of observations or high-dimensional covariate spaces, comparatively little work addresses scalable Bayesian variable selection when both dimensions are large simultaneously in a practical setting. This paper presents a novel Bayesian variable selection framework for efficiently analysing data with a large number of both rows and columns. The proposed framework operates via a divide-and-conquer approach, splitting data into batches along both directions and analysing each batch independently in parallel, after which the results are combined together in a two-phase consensus procedure. Experiments demonstrate the computational gains while retaining strong variable selection performance, successfully identifying relevant signals in noisy, high-dimensional settings despite the loss of information induced by data splitting. Practical guidelines are also provided, including recommendations for tuning parameter choices and effective data partitioning strategies, followed by a real application to an H3N2 influenza dataset.

Adiabatic logic reuses the energy stored on load capacitances through quasi-reversible switching, enabling a lower minimum energy consumption than conventional static CMOS. Yet its practicality in FinFET technologies and at multi-GHz clock rates has yet to be investigated. This work provides a systematic evaluation of Positive Feedback Adiabatic Logic (PFAL) simulated in the TSMC 16nm FinFET process. A set of PFAL standard-cell gates were realised, along with two representative combinational circuits - a 2$\times$2 multiplier and a 4-bit comparator - and compared against static CMOS logic using the energy–delay product (EDP) and the energy advantage metric $\eta = E_{\mathrm{CMOS}} / E_{\mathrm{PFAL}}$. Transient simulations reveal three sources of non-adiabatic loss: two specific to the PMOS/NMOS latch, threshold-voltage-related loss and a previously unreported redundant charging of the output node and one related to the complexity of PFAL logic trees. The low-threshold Buffer/NOT cell achieves a minimum EDP of $1.23\times10^{-26}$J$\cdot$s at $V_{\mathrm{CLK}} = 0.6$V and $f_{\mathrm{CLK}} = 7.94$GHz, while PFAL preserves an energy benefit over static CMOS of up to roughly $5\times$ at reduced frequencies and elevated supply voltages. A parallel-coupled quadrature voltage-controlled oscillator is designed as a realistic four-phase power-clock generator. With this non-ideal supply, the Buffer/NOT energy stays within $2\%$ of the ideal sinusoidal case at $3$GHz. A loading study quantifies the phase shift and amplitude reduction induced by increasing fan-out. Overall, the results provide a design-oriented evaluation of PFAL in 16nm FinFET and a motivation to exploit adiabatic logic for future low-power system architectures.

Adiabatic logic can recover part of the energy stored on load capacitances through quasi-reversible switching, but its waveform-optimized operation in FinFET technology and at multi-GHz frequencies remains underexplored. This work investigates Positive Feedback Adiabatic Logic (PFAL) in the TSMC 16 nm FinFET process. A functionally complete PFAL gate library is designed, verified, and characterised using energy-delay product optimization over power-clock amplitude, frequency, and waveform shape. The power-clock sweep shows that the minimum-energy waveform approaches a triangular shape rather than a conventional trapezoid. The optimized single-gate PFAL cells achieve up to 3.83x lower energy than equivalent static CMOS gates, while sinusoidal excitation improves energy by up to 1.32x compared with trapezoidal excitation and extends the valid operating range. The library is then used to construct larger combinational circuits, including a 2:1 multiplexer, a 4-bit ripple carry adder, and a 4-bit Brent-Kung carry look-ahead adder. The carry look-ahead adder reaches a gain of up to 5.3x compared with the static CMOS energy estimate for the triangular power-clock.

The halo nucleon-core system is, by definition, a shallow state nucleus. The nucleon, in most cases a neutron, is loosely bound to the other nucleons forming the core. Accordingly, the system is located inside the universal window; the halo nucleon most likely resides far from the rest of the nucleons, showing a remarkable insensitivity to details of the interaction with the core. The system can be described using a simple nucleon-core interaction, and this description can be extended to the nucleon-nucleon-core system. Specifically, treating the neutron-core and the neutron-neutron-core systems with a Gaussian interaction as a reference, we show trajectories inside the universal window governed by interaction parameters determined from low-energy observables, such as the neutron-core binding energy, scattering length, and effective range. In this way, we can relate properties of different halo nuclei that might seem uncorrelated. In particular, we determine the three-body parameter, the binding momentum at the unitary limit, and show that the two-neutron halo nuclei emerge from that limit following a linear trajectory depending on the neutron-core range.

With the widespread use of online navigation and ride-hailing services, achieving optimal route planning for diverse user preferences has recently attracted increasing attention. Classic graph algorithms for pathfinding use heuristic cost functions to define edge weight, thus providing no optimality guarantee of route quality. Prior data-driven approaches equating ground truth of the optimal route with user trajectory, which is however moderately influenced by the navigation service, suffers from the feedback loop problem. To address these issues, we propose a deep architecture that is able to jointly optimize cost functions and route-ranking model towards any route preference. First, we run a multi-objective Dijkstra algorithm offline to collect the set of Pareto optimal routes, deeming it as the complete candidate set. Exploiting the property of such a set, we design a neural network structure that emulates shortest-path search and route ranking in an end-to-end differentiable manner. Second, we define route preference as a task of constrained optimization of route attributes, and propose a novel loss function that optimizes a single-objective variable, with other variables strictly under constraints. We conduct extensive experiments on real-world datasets. The results show that our architecture significantly outperforms state-of-the-art methods in route quality and customizability.

We develop an encounter-based formulation of restricted diffusion in a bounded domain with multiple targets, resolving separately the boundary local time accumulated on each target. Unlike the single-target problem, the target projection operators do not generally commute with the governing Dirichlet-to-Neumann (DtN) operator, so that its eigenbasis does not diagonalize the local-time dependence. We overcome this difficulty by decomposing the DtN operator into target-restricted blocks. Multi-dimensional Laplace inversion then yields a convergent switching expansion, whose successive terms describe alternating diffusive transfers between the targets. We also derive an equivalent operator-valued renewal equation. For two targets, diagonalizing the target-restricted blocks provides an explicit spectral representation in terms of two DtN block spectra and an inter-target coupling matrix. When the coupling preserves spectral modes, the switching series can be resummed exactly in terms of modified Bessel functions. In general, as the off-diagonal DtN block is smoothing for separated targets, one can resort to low-rank finite-dimensional approximations. We examine an effective two-mode reduction for small, well-separated targets and identify a regime in which repeated inter-target transfers are progressively suppressed. Probabilistic interpretations and implications for diffusion-controlled reactions are discussed.

Graphical multiple testing procedures are a flexible and transparent way to control the family-wise error rate when a confirmatory trial pursues several label claims, but they leave open the question of which graph to use. In practice sponsors often fall back on fixed-sequence or Holm procedures that may poorly reflect what the trial is actually trying to achieve. We propose to instead choose the graph that maximises an explicit gain function, which states what each possible set of rejected hypotheses is worth to the sponsor: typically nothing until a regulatory hurdle is cleared, then an incremental amount for each further claim. Conventional power criteria are recovered as special cases, and because the search is confined to graphical procedures, family-wise error rate control holds whichever graph is selected. We set out a structured procedure for eliciting the gain function from the clinical, commercial, and regulatory members of a trial team. The proposed framework also handles uncertainty in the treatment effects and correlations assumed at the design stage, by averaging performance over their plausible values rather than fixing a single assumption. For group sequential designs, we describe how it can further reward early claims. Five examples, drawn from real and hypothetical pharmaceutical trials, show that the best graph depends on how trial success is defined, how uncertain the design assumptions are, and, in a group sequential setting, when claims can be established.

Active galactic nuclei (AGN) feedback can alter the abundance and spatial distribution of galaxies within large-scale structure. We apply topological data analysis (TDA) to Simba hydrodynamical simulations with varying AGN feedback at $z=0$ to determine which galaxy populations, environments, and spatial scales are most affected. The galaxy two-point correlation function is most sensitive to feedback at $r<1 h^{-1} Mpc$. We then use Betti curves, persistence diagrams, and the 2-Wasserstein distance to quantify feedback-dependent changes in topology. Persistence diagrams show clearer differences than Betti curves, particularly for satellite galaxies. For satellites, loop-like features show significant differences at $\alpha=0.1$–$0.4 h^{-1} Mpc$ for all feedback channels, together with a weaker cumulative difference at $\alpha=1$–$3 h^{-1} Mpc$. The differences are strongest for satellites in quenched halos with $M_h\gtrsim10^{12} M_\odot$. Optimal matching shows that they are driven mainly by shifts in the birth and death scales of loop-like features rather than by their appearance or disappearance. These results indicate that persistent homology can provide complementary constraints on AGN feedback and environmental quenching in galaxy surveys.

This white paper presents the current status of the ecosystem of statistical tools used in High Energy Physics (HEP), and attempts to summarize the views on the future R&D directions. These views have been collected during the latest IRIS-HEP statistical ecosystem blueprint, which took place in February 2026.

Omni models can describe video content, but can they locate events in time, preserve event order, and judge audio-visual synchronization? We introduce AVTrace (Audio-Visual Temporal Reasoning Assessment and Capability Evaluation), a silver-standard diagnostic suite spanning onset and span grounding, synchronization, next-step prediction, cross-modal localization, chain parsing, and event-conditioned comprehension. It contains 34,114 training examples and category-balanced development and test splits of 3,500 and 7,000 examples. We evaluate five open omni models under their respective input configurations using reference-blind response normalization followed by deterministic scoring. All five off-the-shelf systems score below the test split's majority-label baseline of 0.556 on synchronization verification, and obtain low scores on chain parsing and event-conditioned grounding and comprehension. Development-set perturbations reveal task-dependent sensitivity in Qwen3-Omni-30B to modality removal and changes in visual input processing, without isolating their underlying causes. Parameter-efficient temporal post-training improves Gemma4-E4B-it on several benchmark metrics. On three external image benchmarks, task metrics change modestly, including some degradations, while teacher-forcing perplexity decreases. Together, these findings show that semantic reference-text overlap should not be treated as a proxy for temporal localization, and that AVTrace can identify task-specific weaknesses while providing a testbed for temporal post-training.

The high visual-token load in multimodal large language models (MLLMs) motivates training-free pruning to reduce later-layer computation, but under a fixed budget, pruning must preserve query-relevant evidence while avoiding redundancy. Existing methods rank tokens, diversify selected subsets, or optimize coverage without using a shared per-visual query utility to weight both visual targets and candidate representatives. We introduce QCPruner, which makes both roles query-conditioned through bilateral utility weighting. Using keyword-matched query anchors, QCPruner fuses two cross-modal cues into utility and applies it to both visual targets and candidate representatives within visual-affinity-based coverage. The resulting nonnegative facility-location objective is monotone and submodular, retains the standard (1-1/e) greedy guarantee, and requires no model training or parameter updates. Across LLaVA-1.5, LLaVA-NeXT, LLaVA-Video, and Qwen2.5-VL, QCPruner achieves the highest average relative performance among evaluated complete-system pruning methods at every reported token budget. At 32 of 576 tokens on LLaVA-1.5-7B, it retains 96.1% of unpruned performance, versus 93.9% for the strongest evaluated baseline. At 256 of 1296 tokens on Qwen2.5-VL-7B, the corresponding values are 96.7% and 92.5%.

The widespread adoption of LLMs has led to escalating content compliance risks. Prior works have contributed to addressing these risks in the English context, downplaying the complexity of Chinese language content. This paper follows China's current AI-Generated content compliance requirements and provides evaluation results on 20 notable LLMs, offering insight into China's regulatory landscape. We design a novel framework to assess the compliance and refusal rates with 2303 questions spanning six distinct dimensions, including 203 self-constructed constitutional questions. The framework employs several judges to generate verdicts independently based on their hierarchical alignment memory. Our findings show that international models also exhibit high levels of compliance despite the use of standard Chinese questions, and the main differences may stem from dimensions closely related to ideological alignment. We establish a regulatory benchmark that enables the global AI community to evaluate both Chinese and non-Chinese LLMs under a unified set of legally grounded compliance requirements.

In this paper, three operations on Gibbs probability measures are studied. The first operation, often referred to as renormalization, takes one Gibbs probability measure and generates a new Gibbs measure by normalizing a power of its density. This normalization has a twofold effect: it changes the regularization factor and concentrates the support within a subset of the original support. Interestingly, these effects can be independently controlled by different parameters. The second operation consists of a normalized log-linear combination of the densities of Gibbs probability measures. The third operation takes two Gibbs probability measures and changes the reference measure of the latter with the former. Hence, the former is said to be "nested" within the latter, yielding a new Gibbs probability measure. The resulting measures from the second and third operations are also Gibbs probability measures and are shown, respectively, to solve optimization problems involving the expectations of linear combinations of the objective functions of the given measures, subject to a relative entropy regularization. These optimization problems differ exclusively in the coefficients of the linear combinations. This leads to the conclusion that there exists a set of parameters for which nesting one Gibbs probability measure into another has the same effect as log-linearly combining them. These operations have relevant applications in statistical learning. As an example, a one-shot federated learning system in which clients send their locally trained Gibbs algorithms to the server for log-linear combination is shown to achieve the same performance as a Gibbs algorithm trained upon the aggregation of all local training datasets.

Cardiac T1 and T2 quantitative magnetic resonance imaging (MRI) is a technique that provides characterization of a multitude of myocardial pathologies. However, each parameter requires its own specialized acquisition sequence, which may be time-consuming. In this work, we demonstrate that MOLLI, an acquisition sequence commonly used to perform T1 mapping, is also sensitive to T2, and can also be used for myocardial T2 mapping. Although MOLLI is not designed for T2 mapping, its signal evolution inherently depends on both T1 and T2 relaxation. Here, we demonstrate that this dependence is sufficiently strong to allow T2 estimation directly from MOLLI T1-weighted images, using appropriate signal modelling. This finding opens the possibility of obtaining simultaneously T1 and T2 maps from a single MOLLI acquisition, reducing total scan time and potentially improving patient comfort and clinical throughput.

The Brunner-Munzel (BM) test is a nonparametric test for two independent samples that evaluates whether observations from one group tend to be greater than observations from another group, or vice versa. The BM test has a broader scope of application than the Mann-Whitney $U$ test because it does not assume equal variances between the two groups. However, the meaning of the BM test statistic is difficult to understand intuitively, which may be one of the factors hindering the widespread use of the BM test. To alleviate this problem, in this paper, I introduce an alternative interpretation of the BM test statistic from the viewpoint of local linear approximation. It is shown that the variance estimator for the sample stochastic superiority used in the BM test can be derived using local linear approximation, in which the influence of each observation on the sample stochastic superiority is assumed to be additive. This simple interpretation will help practitioners decide to use the BM test without hesitation.

Fine-tuning foundation models on new tasks inevitably suffer from catastrophic forgetting. While existing works attempt to mitigate this on the basis of parameter-efficient fine-tuning methods, they adopted an overly restrictive Subspace Orthogonality condition. In this paper, we introduce a purely post-hoc and tuning-agnostic weight rectification framework that achieves Parameter Space Orthogonality, which is the necessary and sufficient condition for preserving historical performance to the first order. By projecting parameter updates into the JAcobian NUll Space (JANUS), our method significantly recovers compromised historical knowledge without interfering with the underlying fine-tuning process. To overcome the local validity of the Jacobian approximation, we further propose a Multi-step Adaptive Rectification mechanism that utilizes the JANUS shift to dynamically verify the valid trust region and adjust step sizes. Coupled with our proposed ghost projection, ghost orientation comparison, and sequence-level singular value decomposition compression techniques, JANUS also achieves great temporal and spatial efficiency. Experiments demonstrate that JANUS seamlessly integrates with various fine-tuning methods, significantly mitigating the stability-plasticity dilemma by recovering historical knowledge while preserving downstream task adaptation.

Recurrent structure in supernova (SN) light curves may reveal physical clocks associated with central-engine dynamics, binary orbital motion, or interaction with structured circumstellar material (CSM). We present a multiband photometric reanalysis of the nearby hydrogen-poor superluminous SN (SLSN-I) 2018bsz (z=0.0267), using Swift/UVOT and GROND observations spanning approximately 110 rest-frame days after maximum light. After independently modelling and subtracting the smooth decline in each band, Generalized Lomb-Scargle periodograms reveal recurrent residual variations at broadly consistent phases across ten bands from u to Ks. The strongest individual-band detections, in the g, r, i, z, and J bands, yield periods of 31.5-31.8 days, while the joint multiband analysis gives a rest-frame period of P=31.61 (+/-0.03) days. The modulation is traced over approximately three cycles and is therefore described as quasi-periodic. The signal is robust to the adopted detrending procedure, while comparison-star and lunar-cycle tests disfavour an observational systematic. The modulation is broadly phase-coherent from the optical to the near-infrared, but its amplitude increases towards shorter wavelengths; lower-significance bands also show possible wavelength-dependent shifts in the best-fitting period. The present observations do not uniquely determine the physical origin. Lense-Thirring precession, interaction with CSM structured before the explosion, and post-SN interaction with a surviving companion remain possible. SN 2018bsz provides one of the clearest examples of coherent, month-scale photometric modulation in an SLSN-I.

Graph Neural Networks (GNNs) achieve state-of-the-art performance on graph-structured data, but training and inference on large graphs are often bottlenecked by memory constraints and sparse linear-algebra workloads. Quantum computing offers an alternative set of primitives that may improve scalability for graph learning. Building on the quantum graph neural network (QGNN) framework of Liao et al., this work implements two representative architectures — the Simplified Graph Convolution (SGC) and Linear Graph Convolution (LGC) models — and evaluates them on open benchmark graph datasets and semi-supervised learning tasks using quantum simulation. We compare predictive performance and optimization behavior against classical baselines, showing that the quantum models achieve competitive performance with fewer parameters. Finally, we present a cost gradient analysis that identifies the tasks for which the models showcased are trainable. This is followed by a classical simulability study to find regimes in which the proposed circuits remain robust during training.

We prove a prescribed-index-law comparison for canonical processes with independent symmetric coordinates having log-concave tails, with no doubling assumption