GPU-initiated I/O has emerged as a key mechanism for achieving high-throughput storage access by leveraging massive GPU thread-level parallelism, while recent industry trends point toward SSDs optimized for ultra-high random-read IOPS. Together, these trends are enabling the emergence of IOPS-optimized, GPU-centric storage systems. Despite this momentum, no existing framework enables quantitative end-to-end evaluation of storage systems optimized for GPU-initiated I/O. While conventional SSD emulators provide a promising path toward end-to-end modeling in traditional storage systems, they face three key challenges in this GPU-centric setting: limited frontend scalability for ingesting massive request streams, high software overhead in emulating GPU-initiated I/O control and data paths, and excessive timing-model maintenance overhead at extremely high I/O request rates. We propose SwarmIO, an SSD emulator for massively parallel, GPU-centric storage. SwarmIO faithfully models IOPS-optimized SSDs at target performance levels of up to 40 MIOPS, achieving a 303.9x speedup over the state-of-the-art baseline SSD emulator under GPU-initiated I/O. We further demonstrate its utility through a vector search case study, showing that increasing SSD IOPS from 2.5 MIOPS to 40 MIOPS yields an average end-to-end speedup of up to 9.7x.

Computing-in-Memory (CiM) is a promising paradigm to address the memory bottleneck constraining traditional systems. Most power-efficient CiM variants can directly perform Boolean operations in non-volatile memory arrays. Higher microarchitectural activity due to CiM, however, can significantly increase power density (power per area) and result in thermal hotspots. In this paper, we provide a quantitative thermal characterization for CiM. We demonstrate that (i) the temperature remains mostly uniform due to lateral thermal conduction; (ii) the temperature increases linearly with the number of memory cells participating in computation; (iii) the temperature decreases linearly with the memory array size; (iv) the memory technology dictates the power density, hence the thermal characteristics.

The power sector is a major source of fossil fuel use and air pollutant emissions, making high-spatiotemporal-resolution emission accounting essential for effective mitigation policy and air quality management. Yet existing public inventories are often limited by low timeliness and coarse resolution. Here, we develop a global, plant-level, daily, multi-pollutant emission database for the power sector by integrating nearly 3 million hourly-to-daily near-real-time power generation records from 57 countries, representing about 81% of global fossil-fuel-based electricity generation, with fundamental information for more than 10,000 power plants worldwide, including location and installed capacity. The dataset substantially improves the timeliness and granularity of global power-sector emission estimates. From 2019 to 2025, emissions of most pollutants increased, with 2025 daily mean emissions reaching 0.274 kt/d for BC, 45.1 kt/d for CO, 0.418 kt/d for NH3, 52.2 kt/d for NOx, 3.01 kt/d for NMVOC, 0.418 kt/d for OC, 6.76 kt/d for PM10, 5.11 kt/d for PM2.5, and 78.5 kt/d for SO2. Compared with 2019, NMVOC showed the largest increase, whereas SO2 was the only pollutant to decline overall. Coal remained the dominant source of sulfur-, nitrogen-, and particulate-related emissions, while gas and biomass contributed more to carbonaceous species and reduced nitrogen. The dataset also captures pronounced seasonal, regional, and short-term variability. Against EDGAR for 2019-2022, our estimates agree well, with Pearson correlations of 0.92-0.99 and mean relative deviations of 8.8%-28.1%. This near-real-time, high-resolution dataset provides a strong foundation for air pollution control, carbon mitigation, emission monitoring, and satellite-based inversion.

Chiral liquid crystals and chiral magnets host a wide variety of topological solitons described by closely related continuum theories, namely the Frank-Oseen and Dzyaloshinskii models. Exploiting this correspondence, we develop a unified description of cholesteric fingers in confined liquid crystals and their magnetic counterparts. Within a continuum framework including bulk and surface anisotropies, we analyze the topology, structure, interactions, and collective states of the two main finger types, CF-1 and CF-2. We show that cholesteric fingers are composite chiral solitons built from merons. CF-2 corresponds to a bimeron with unit topological charge, while CF-1 is a topologically trivial composite of two merons with identical vorticities. From a homotopic viewpoint these textures correspond to skyrmions and droplets. Strong homeotropic anchoring induces confinement effects that reshape the meron structure and redistribute topological charge across the film thickness. Isolated fingers in the homogeneous state interact repulsively and behave as particle-like objects. Periodic phases emerge when the energy of an isolated finger becomes negative, leading to nucleation-type transitions with a diverging lattice period. Degenerate finger types allow mixed periodic sequences, analogous to stacking polytypes. In a conical background, interactions become attractive due to overlap of distortion regions. Film thickness controls stability and structure: at small thickness solitons collapse, while at large thickness bimerons exhibit bistability between surface-stabilized and bulk-like states.

High-dimensional data in modern applications, such as COVID-19 mortality, often span multiple domains. Leveraging auxiliary information from source domains to improve performance in a target domain motivates the use of transfer learning. However, a practical issue that has been overlooked is data contamination, which induces heterogeneity and can significantly degrade transfer learning performance. To address this challenge, we propose a novel approach that tackles transfer learning under data contamination within a structured regression setting. By employing the robust L2E criterion, we develop the TransL2E method that accounts for contamination in both target and source data while effectively transferring relevant information. Beyond robust estimation, TransL2E introduces a data-driven bi-level source detection mechanism, operating at both individual and cohort levels, which possesses multiple advantages over existing source detection approaches. Comprehensive simulation studies and a real data application demonstrate the superior performance of TransL2E in both robust estimation and structure recovery in the presence of data limitation and contamination.

As a key capability of 6G, sensing-communication (S&C) coexistence over distributed infrastructure is expected to support next-generation ultra-reliable and low-latency communication (xURLLC) applications, which demand both robust connectivity and real-time environmental awareness. This paper investigates network-wide information freshness in large-scale cell-free massive multiple-input multiple-output (CF-mMIMO) with S&C coexistence. A challenge arises from the spatial partitioning of access points (APs) into S&C roles: allocating more APs to sensing improves update generation, whereas allocating more APs to communication enhances reliable short-packet delivery. To address this, we develop a unified analytical framework by combining stochastic geometry and stochastic network calculus (SNC) to characterize the peak age of information (PAoI) violation probability (PAVP). Specifically, we derive the moment generating functions (MGFs) of sensory packet inter-arrival and service times, accounting for the joint stochastic spatial distribution of APs and users, imperfect channel state information (CSI), and finite blocklength coding (FBC). This facilitates the derivation of a tractable upper bound on the PAVP, which is minimized to determine the optimal AP partitioning. The derived bound accurately captures the performance trend and yields a minimizing partition factor that closely matches simulations. Therefore, the framework provides an efficient and low-complexity tool for network-wide PAoI guarantee and coexistence-oriented design in CF-mMIMO networks toward xURLLC.

Oxide semiconductors have emerged as common channel materials in transistors and hold promise for next-generation electronics, yet achieving high mobility typically requires costly vacuum-based techniques. Here, ultrathin (5-nm) indium native oxide (InOx) prepared by ambient-air liquid-metal printing (LMP) at low temperature (250 °C), is applied as semiconducting channel in field-effect transistor (FET). The resulting InOx is found to be polycrystalline with large lateral grains that extend vertically throughout the film thickness. InOx FETs in a transfer length method (TLM) configuration demonstrate a high conductivity mobility (uCON) of 125 cm2 V-1 s-1, with systematic analysis of contact resistance confirming potential for channel length scaling. Integration with atomic-layer-deposited (ALD) gate dielectrics further reveals excellent compatibility, for instance, InOx FET integrated with HfO2 exhibits a high field-effect mobility (uFE) of 107 cm2 V-1 s-1, an on/off current ratio (ION/IOFF) of >107, a subthreshold swing (SS) of 204 mV dec-1, a gate leakage of <10-6 A cm-2, while maintaining stable performance over 104 endurance cycles without degradation. Post-fabrication oxygen-plasma treatment is applied to achieve enhancement-mode operation and a depletion-load inverter is demonstrated, exhibiting a voltage gain of 69.8 V/V. These results demonstrate the great potential of LMP InOx as semiconducting channel in high-performance and power-efficient transistors for next-generation oxide electronics.

The extraordinary one-dimensional properties of carbon nanotubes have captivated scientists and engineers since their discovery in the early 1990s. In particular, semiconducting single-wall carbon nanotubes (SWCNTs) are highly promising for optoelectronic applications because of their diameter-dependent direct band gaps and strong, tunable light-matter interactions. However, the prevalence of structural disorder, misalignment, and chirality heterogeneity in macroscopic assemblies has hindered their practical applications. Recently, advanced assembly methods, combined with post-growth chirality separation techniques, have enabled the fabrication of wafer-scale, nearly crystalline films of highly aligned and densely packed SWCNTs with tailored properties. In this Opinion, we discuss how these films provide a transformative platform for engineering "Single$^3$" heterostructures-assemblies that are simultaneously single-crystal, single-chirality, and single-wall. Stacking these layers with nanometer-scale precision and tunable thicknesses allows for the realization of artificial bilayer junctions, quantum wells, and superlattices. We posit that these architectures will enable a new generation of high-performance devices, including lasers, photodiodes, solar cells, and single-photon emitters.

Recent work has established that the trajectory of the Nesterov ODE, a the continuous-time model of Nesterov's accelerated gradient method, exhibits point convergence towards a minimizer of a convex potential. A natural next question is whether this point convergence can be upgraded to rectifiability, namely whether the convergent orbit has finite path length. This work provides the answer in the negative by constructing a differentiable convex potential in $\mathbb{R}^2$ for which the flow converges to its minimizer but still accumulates infinite path length. All proofs of this work are due entirely to an internal model at OpenAI.

Clinical decisions for unruptured intracranial aneurysms depend on detecting growth on follow-up magnetic resonance angiography (MRA). Growth is typically judged from manual 2D diameters on few slices, which vary across clinicians and frequently miss subtle 3D change. Even with 3D segmentations, apparent differences can reflect resolution, segmentation, surface processing, or registration mismatch rather than true growth; most criteria remain heuristic and binary. We show that a Bayesian displacement-based model using the surrounding vessel as an internal reference achieves strong discrimination of aneurysm growth (AUC 0.86-0.87) and improves agreement with expert labels (Cohen's kappa up to 0.66 vs. 0.35 for volumetric criteria), while providing calibrated posterior probabilities with uncertainty bounds. The method registers baseline and follow-up surfaces, computes normal-directed displacements, and summarizes change as the difference between mean aneurysm displacement and mean displacement on the surrounding non-aneurysmal vessel segment. The vessel segment serves as an internal control for imaging and processing variability, assuming negligible structural change over the surveillance interval. We evaluate two cohorts spanning time-of-flight and contrast-enhanced longitudinal MRA studies: a public dataset labeled from neuroradiologist-provided measurements and an institutional dataset labeled by senior and junior raters. Performance is preserved when training on lower-expertise labels, indicating robustness to label variability. Calibrated probabilities may aid clinical decision-making in borderline cases, where high uncertainty can motivate repeat imaging. This framework provides interpretable probabilistic growth assessment from longitudinal MRA, reduces dependence on clinician expertise, and supports cross-center surveillance across scanners and angiography sequences.

Accurate localization in non-line-of-sight (NLoS) environments remains challenging even with both angle-of-arrival (AoA) and time-of-arrival (ToA) measurements. In complex urban scenarios, the absence of line-of-sight (LoS) paths and the lack of environment prior knowledge make geometric based localization methods inapplicable, while prior-based approach such as fingerprinting is sensitive to environmental perturbations. This paper proposes a novel environment-aware localization framework enabled by the emerging concept called channel knowledge map (CKM). In the offline stage, AoA-ToA path signatures are learned by the CKM, with each path mapped to one candidate scatterer, thereby forming geometric priors within the environment. In the online stage, observed paths are matched to the CKM to extract high-confidence scatterers. Nonlinear least squares (NLS) method is then applied to jointly estimate the user and dominant scatterer locations. Even with imperfect CSI matching, geometric feasibility consistent with CKM scatterer priors provides corrective information and suppresses ambiguity. Simulations demonstrate that the proposed scheme outperforms fingerprinting and offers a robust and scalable solution to address the challenging NLoS localization for integrated sensing and communication (ISAC) systems.

This paper develops a unified framework for testing monotonicity of Bayesian Nash equilibrium strategies in unobserved types in games of incomplete information. We show that, under symmetric independent private types, monotonicity of differentiable equilibrium strategies is equivalent to monotonicity of a quasi-inverse strategy identified from observed actions. This allows the problem to be reformulated as testing a countable set of moment inequalities involving unconditional expectations. We propose a Cramer-von Mises-type statistic with bootstrap critical values. The method accommodates covariates and game heterogeneity. Monte Carlo simulations demonstrate finite-sample performance, and an application to procurement auctions illustrates cartel detection.

Cost-effective, low-complexity and spectrally efficient interconnection can offer fundamental guiding law for future datacenter. In this work, we demonstrate a cost-efficient SSBI-free direct detection for datacenter interconnection, leveraging the phase diversity of residual optical carrier caused by finite-extinction ratio (ER) IQ modulators, combining the device cost-effective IQ modulator with finite-ER and efficient SSBI-free phase-diverse direct detection receiver. Specifically, the proposed solution transforms the inherent limitation of finite-ER of cost-effective IQ modulator into the residual optical carrier advantage of SSBI-free direct detection systems, eliminating SSBI without additional hardware and control complexity. A digital pre-distortion and offset correction algorithms, and a PD-thermal-noise constrained SSBI-free direct detection and signal recovery algorithms are derived and implemented. Comprehensive simulations are conducted. A Global-SNR gain of 1.78 dB and 400 Gb/s data rate are achieved in 100-km SSMF transmission when (ER_i, ER_o)= (7 dB, 25 dB) of IQ modulator. The proposed solution enables low-complexity, cost-effective, and spectrally-efficient interconnects for next-generation datacenters.

Tag-based physical layer authentication (PLA) has garnered significant attention due to its low complexity and enhanced security. However, existing PLA schemes encounter two challenges. First, unintended user interference, which overlaps with the authentication signal, corrupts the tag and degrades authentication performance. Second, the vulnerability introduced by direct embedding of the raw tag exposes the tag to the adversary and degrades the security. To address these challenges, this paper proposes a novel frozen-tag-based PLA framework. Different from typical schemes that directly embed the uncoded tag into the signal, a well-designed frozen tag is inserted for authentication, where the frozen tag is generated based on the concept of polar codes with the anchor information as information bits and raw tags as frozen bits. Accordingly, the proposed PLA framework offers two principal advantages. First, the authentication performance is improved since the legitimate receiver can decode the frozen tag and mitigate unintended user interference. Second, the authentication process becomes indecipherable to the illegitimate receiver due to the concealment of the raw tags. Furthermore, we conduct a comprehensive analysis of the proposed framework in terms of robustness, security, and compatibility. Theoretical analysis and simulation demonstrate that the proposed frozen-tag-based PLA framework not only enhances the detection performance but also significantly degrades Eve's capability to estimate the raw tags.

In this work we give a complete description of the collection of curves of tangencies induced by germs of foliation pairs -- non dicritical and dicritical -- given by analytic differential equations with degenerated non dicritical and dicritical singularities, satisfying some genericity assumptions. To this purpose we use local models and analytic normalizing transformations. Moreover, for each natural number $k$ we obtain $k$-normal forms for the normalizing transformations. These normal forms are used to give parametrizations, up to a finite jet, of the branches of the curves of tangencies. We also prove that under natural genericity assumptions any germ of analytic curve having pairwise transversal smooth branches is realized as curve of tangencies of a -- non dicritical and dicritical -- foliation pair.

Shor's algorithm outperforms its classical counterpart in efficient prime factorization. We explore the coherence and entanglement dynamics of the evolved states within Shor's algorithm, showing that the coherence in each step relies on the dimension of register or the order, and discuss the relations between geometric coherence and geometric entanglement. We investigate how unitary operators induce variations in coherence and entanglement, and analyze the variations of coherence and entanglement within the entire algorithm, demonstrating that the overall effect of Shor's algorithm tends to deplete coherence and produce entanglement. Our research not only deepens the understanding of this algorithm but also provides methodological references for studying resource dynamics in other quantum algorithms.

Sparse matrix-dense matrix multiplication (SpMM) is a critical kernel in scientific computing, graph analytics, and machine learning, whose performance is often constrained by memory bandwidth. In this work, we investigate the applicability and limitations of roofline modeling for SpMM by explicitly accounting for the impact of matrix sparsity structure on arithmetic intensity and attainable performance. We evaluate three SpMM implementations: Compressed Sparse Row (CSR), Compressed Sparse Blocks (CSB), and Intel's Math Kernel Library (MKL). Each implementation was tested using large-scale matrices from the SuiteSparse collection and grouped by sparsity pattern, including block-structured, banded (diagonal), scale-free, and uniformly random matrices. We derive sparsity-aware roofline models that incorporate memory traffic, cache locality, and blocking behavior, and demonstrate that a single model is insufficient to accurately predict performance across diverse structures. Experiments were conducted on an AMD-based Perlmutter compute node with a varying number of columns in the dense matrix. In particular, blocking and structured sparsity significantly alter effective arithmetic intensity. The results show that accurate roofline-based performance analysis of SpMM requires sparsity-aware modeling, and that data layout and blocking strategies must be evaluated in the context of matrix structure rather than through a single unified model.

We study massive modes on a magnetized blow-up manifold of $T^2/\mathbb{Z}_N$. The blow-up manifold can be constructed by appropriately replacing orbifold singular points with a part of $S^2$. To ensure a smooth connection between the massive modes on magnetized $T^2/\mathbb{Z}_N$ orbifold and those on magnetized $S^2$, it is required that not only the total magnetic flux as well as the total curvature but also the effective magnetic flux on the connected line remain invariant under the blow-up procedure. Furthermore, we find that the number of the localized modes at each orbifold singular point increases by one for each unit increment of the mass level.

Deep-UV ~ 250 nm (4.96 eV) tilted in-situ co-illumination of the sample under imaging by scanning electron microscope (SEM) is developed at a robust and practical instrument level. Precise mechanical control of the lateral position and tilt angle (within 6.5$^\circ$ from a 42$^\circ$ baseline) of the UV-C LED source is achieved using mechanisms external to the vacuum chamber. The incorporated linear polariser (for s-pol. mode illumination) with external polarisation plane adjustment allows for modulation and tuning of tangential $E^{(t)}$ and normal $E^{(n)}$ electric field components and their enhancement for controlled directional electron emission from the surface of the sample. Numerical modelling of E-field enhancement corroborates the expected enhancement in the production of secondary electrons. This modality of SEM imaging does not require metal coatings, preserving sample integrity for subsequent analysis. The feasibility of having linearly polarised incident UV-C light with azimuthal orientation control in $(s,p)$-plane is modeled and discussed.

Let $(\mathrm{X},σ)$ be a holomorphic symplectic manifold. We study the differential graded category of canonical Lagrangian $\mathrm{D}$-branes $\mathcal{D}_\mathrm{Lag}(\mathrm{X},σ)$ along with its deformation quantisation, spanned by quantised orientations, $\mathcal{DQ}(\mathrm{X},σ)$, and the virtual de Rham category $\mathcal{DR}^{\mathrm{vir}}(\mathrm{X},σ)$. We prove the formality of these dg categories when localised at a countable collection of orientable compact Kähler Lagrangian submanifolds with pairwise clean intersections. Along the way, we define Kaledin classes of minimal $\mathrm{A}_\infty$-categories and show that they are the obstructions to formality. In addition, we obtain a formality criterion for flat weakly proper Calabi-Yau dg categories.

In this paper, we investigate the blow-up in finite time and the corresponding lifespan estimates for a weakly coupled system of wave equations on a compact Lie group. In particular, we show how the Cauchy data and the presence of lower order terms affect the lifespan of local in-time solutions.

We investigate the Bishop-Phelps-Bollobás property for the numerical radius (BPBp-nu) through a Zizler-type perspective on the classical Bishop-Phelps-Bollobás property (BPBp). This approach allows us to establish two new results: the real Banach space $\ell_\infty$ satisfies the BPBp-nu, while the complex space $\ell_1 \oplus_\infty c_0$ does not. Note that the latter provides the first natural example (constructed without renorming techniques) of a Banach space where the numerical radius attaining operators are dense but the BPBp-nu fails. Along the way, we strengthen the main results of the paper [Kim et al, On the Bishop-Phelps-Bollobás theorem for operators and numerical radius, Studia Math., 2016] concerning the interplay between the BPBp for the pair $(X,Y)$ and the BPBp-nu for a direct sum $X\oplus Y$ of Banach spaces. We further explore the validity of the Zizler-type BPBp across different pairs of Banach spaces, and how this property relates to the classical BPBp and the BPBp-nu. Finally, we specialize our analysis to the framework of compact operators.

The rapid proliferation of data centers is reshaping modern power system dynamics. Unlike legacy industrial loads, data centers have power-electronic interfaces whose multi-timescale dynamics can interact strongly with the grid, inducing oscillatory behavior. However, analytical models that are grid-integratable for revealing the underlying resonance mechanisms remain largely unexplored. To fill this research gap, this paper derives an explicit, component-informed dynamic model of data-center power-delivery chains, which preserves component-level fidelity and captures inter-stage control interactions. This model is formulated as a time-invariant representation in the positive-sequence domain, enabling seamless integration with the phasor (or RMS) domain power-system dynamic models. The analytical derivation reveals how realistic server-load fluctuations at specific frequencies can excite coupled control modes, thereby inducing oscillation amplification and propagation in power grids with heterogeneous dynamic resources, including synchronous machines and grid-forming/following inverters. Case studies on test systems with some realistic data center data demonstrate the effectiveness of the proposed solutions.

Let $F_0$ be the function field of a curve over a $p$-adic field $K,$ and let $F$ be a quadratic extension over $F_0$. Let $A$ be a central simple algebra over $F$ of period $2,$ and let $τ$ be a $F/F_0$-involution on $A$. We show the triviality of the reduced unitary Whitehead group $SK_1U( A, τ)$ if $p\neq 2$.

We give a quintet of proofs resulting from questions posed by Erdős. These questions concern ordinary lines in planar point sets, sequences with uniformly small exponential sums, $K_4$-free $4$-critical graphs with few chords in any cycle, a counterexample to a "fewnomial" version of the Erdős--Turán discrepancy bound, and a finiteness theorem for integers $n$ such that $n-a k^2$ is prime for all $k\leq \sqrt{n/a}$ coprime to $n$ (for fixed $a\in\mathbb Z_+$). Each proof is due to an internal model at OpenAI.

We present an enhanced entangled quantum clock protocol that incorporates a quantum phase estimation algorithm to directly estimate proper-time differences as an unknown phase. By employing highly entangled multi-clock states, the achievable uncertainty scales inversely with the total number of quantum clocks, surpassing the standard projection-noise limit. This approach extends the original EQC framework and provides a systematic method for high-precision relativistic time comparison.

In this paper, we consider the weighted $p$-Laplacian equation $$ Δ_{p,f}u+au^σ\ln u=0$$ defined on a complete smooth metric measure space under the conditon that the $m$-Bakry-Émery Ricci curvature has a lower bound, where $a$, $σ$ are two nonzero real constants. By applying the Nash-Moser iteration, we obtain sharp gradient estimates and thereby establish Liouville theorems for the above equation.

Magic states play an important role in fault-tolerant quantum computation, and so the quantification of magic for quantum states is of great significance. In this work, we propose two new magic quantifiers by introducing two versions of quantum $(α,β)$ Jensen-Shannon divergence based on the quantum $(α,β)$ entropy and the quantum $(α,β)$-relative entropy, respectively. We derive many desirable properties for our magic quantifiers, and find that they are efficiently computable in low-dimensional Hilbert spaces. We also show that the initial nonstabilizerness in the input state can boost the magic generating power for our magic quantifiers with appropriate parameter ranges for a certain class of quantum gates. Our magic quantifiers may provide new tools for addressing some specific problems in magic resource theory.

We show that the space of sections of any line bundle on the augmented wonderful variety of a hyperplane arrangement has the structure of a coalgebra. These coalgebras correspond to the hyperplane arrangement power ideals of Ardila and Postnikov, which include zonotopal algebras as a special case. By proving cohomology vanishing results on augmented wonderful varieties, we recover many results about zonotopal algebras. We also interpret the "superspace" zonotopal algebras of Rhoades, Tewari, and Wilson in terms of the sections of vector bundles on the augmented wonderful variety, and we use this interpretation to prove a formula that they conjectured for the Hilbert series of the superspace version of the central zonotopal algebra.

Concept drift and adversarial evasion are two major challenges for deploying machine learning-based malware detectors. While both have been studied separately, their combination, the adversarial robustness of drift-adaptive detectors, remains unexplored. We address this problem with AdvDA, a recent malware detector that uses adversarial domain adaptation to align a labeled source domain with a target domain with limited labels. The distribution shift between domains poses a unique challenge: robustness learned on the source may not transfer to the target, and existing defenses assume a fixed distribution. To address this, we propose a universal robustification framework that fine-tunes a pretrained AdvDA model on adversarially transformed inputs, agnostic to the attack type and choice of transformations. We instantiate it with five defense variants spanning two threat models: white-box PGD attacks in the feature space and black-box MalGuise attacks that modify malware binaries via functionality-preserving control-flow mutations. Across nine defense configurations, five monthly adaptation windows on Windows malware, and three false-positive-rate operating points, we find the undefended AdvDA completely vulnerable to PGD (100% attack success) and moderately to MalGuise (13%). Our framework reduces these rates to as low as 3.2% and 5.1%, respectively, but the optimal strategy differs: source adversarial training is essential for PGD defenses yet counterproductive for MalGuise defenses, where target-only training suffices. Furthermore, robustness does not transfer across these two threat models. We provide deployment recommendations that balance robustness, detection accuracy, and computational cost.