Papers

This study combines a stochastic renormalization group (RG) in space with a path-integral description of the time evolution, giving a spatial hierarchy of coarse-grained dynamics together with the distribution over spatiotemporal paths. The RG is defined as a diffusion process that applies scale-dependent Laplacian damping together with additive Gaussian noise. The time evolution at each spatial scale is formulated as an Onsager–Machlup action, with a drift (i.e., the predictor) and white noise whose amplitude is fixed by the RG. The predictor of these dynamics is optimized by minimizing the Kullback–Leibler divergence between the path distributions from the RG and from the path-integral description. The optimal predictor contains the score function that connects the spatial scales, so the predictor and the score are two aspects of the same multiscale path formulation. The formulation unifies simulation using the predictor, unconditional generation using the score, and super-resolution using both. The predictor has no closed form, so it is computed by a neural network in two realizations that differ only in how the score is computed. The first realization obtains the score by automatic differentiation of the path distribution, remaining faithful to that formulation. The second obtains the score as an additional network output trained by denoising, at a lower computational cost. These two realizations are validated through numerical experiments on two representative multiscale systems, the Kolmogorov flow and the two-timescale Lorenz-96 model.

We investigate derivation modules of hyperplane arrangements associated with the monomial groups $G(r,p,\ell)$. Motivated by Mühlherr's separator-based construction for graphic arrangements, we extend this approach by combining separators with integral expressions. For a subarrangement $\mathcal{A}_G$ defined by a graph $G$, we show that, when $p<r$, the derivation module $\mathcal{D}(\mathcal{A}_G)$ is generated by the standard derivations $\eta_0,\dots,\eta_{\kappa(G)}$ together with a finite set of separator-based derivations. This result generalizes the complete graph case and provides new evidence toward the broader applicability of separator methods in the study of hyperplane arrangements.

A growing brain tumor mechanically deforms the surrounding healthy brain through mass effect. This compression and displacement of tissue affects both patient prognosis and treatment planning, so biophysical models of tumor growth must account for this deformation. Established mass-effect models are partial differential equation (PDE) solvers that couple growth to elasticity and calibrate parameters per patient, at substantial implementation, runtime, and calibration cost. We ask how much of this deformation a fully differentiable energy-based model can recover, and introduce a potential-spring model: a regular grid over the brain is driven outward by a tumor-concentration potential and held by an elastic spring lattice with tissue-specific stiffness. It shares the tumor-gradient driving force of continuum elasticity but uses a single global parameter and no per-patient inversion, minimized end-to-end by gradient descent. We benchmark it against established solvers by applying each displacement field to a healthy atlas and measuring tissue overlap with the patient's own segmentation on 134 BraTS patients. Our model yields the largest overall local overlap improvement at up to 100$\times$ shorter computation time, establishing a lightweight, differentiable solver for tumor mass effect. Because our model obeys mechanical laws and reflects the tumor-induced tissue loading, it remains physically plausible, a prerequisite for clinical use.

In self-supervised learning, a third modality should improve, or at least preserve, performance. Across nine image-text-tabular datasets, we show that it instead harms performance: the trimodal model underperforms its own best bimodal subset in 55.6% of paired runs under VICReg. The same failure occurs in 51.1% of paired runs under SimSiam. We call this failure multimodal saturation. We propose that the failure lies in the alignment geometry. Riemannian VICReg (R-VICReg) generalizes classical VICReg: it aligns views by squared geodesic distance on learnable negative-curvature product factors and recovers VICReg exactly as curvature vanishes. Over the same 45 paired runs, R-VICReg raises the probability that the third modality helps from 44.4% to 64.4%, with gains concentrated where VICReg saturates.

Operator-learning benchmarks ship as full-precision arrays, and they have grown to terabyte scale. A curator who wants to distribute one has to decide how coarsely to store it. That decision is usually made by fixing a tolerance on the reconstruction error of the stored field. We show that this quantity is measured in the wrong place. It compares the stored field with the original, before any model is trained. What the curator is buying is the accuracy of a model trained on the compressed copy, and the two can disagree: two PDEBench families stored to identical field error differ threefold downstream. A solution operator smooths, so only part of the codec error ever reaches the model's output. That part can be measured. Push a compressed field through a model already trained at full precision, and compare its output against the output on the clean field. The measurement costs forward passes and no training on compressed data. It spans more than two orders of magnitude across six PDE families, and it orders downstream cost where field error does not, inverting 12 of 104 cross-dataset comparisons against field error's 36. Read as a budget, the same curve nominates a storage rate for each dataset. A back-check against trained models finds those rates wrong in both directions by up to a factor of two in stored bits.

Cone-beam computed tomography (CBCT) is widely used in image-guided radiotherapy, but scatter, beam hardening, noise, truncation, and other artifacts limit image quality and CT number accuracy. Paired CBCT and CT data are difficult to obtain clinically because of motion, anatomical changes, and acquisition mismatch. We present RefineCBCT, an unpaired CBCT refinement framework that uses pseudo-label guidance and short-step diffusion to reduce artifacts while preserving patient-specific anatomy. RefineCBCT was trained and evaluated on unpaired CBCT and planning CT data from public LUNG TCIA and PELVIC TCIA datasets and compared with representative GAN and diffusion based methods. On LUNG TCIA, it achieved the best results across all metrics, with MAE 19.411, RMSE 62.758, PSNR 30.845 dB, and SSIM 0.931. On PELVIC TCIA, it achieved the best MAE, PSNR, and SSIM, with values of 14.905, 36.671 dB, and 0.876. The refined images showed fewer streaking and shading artifacts, clearer anatomical boundaries, and improved soft tissue uniformity, with line profile and ROI analyses showing closer agreement with planning CT. These results suggest that RefineCBCT provides efficient and effective CBCT refinement under clinically realistic unpaired training conditions and may support more reliable CBCT use in image-guided radiotherapy workflows. Code is publicly available on GitHub, and the evaluated datasets are available from The Cancer Imaging Archive.

The robustness of Personally Identifiable Information (PII) protection in Large Language Models (LLMs) is a critical concern, yet the risks associated with cross-lingual data extraction remain under-explored. This study evaluates the vulnerability of English-centric and multilingual models to Training Data Extraction (TDE) attacks when prompted in non-English languages. We construct a multi-domain PII dataset comprising social media handles, email addresses, and phone numbers and translate the attack contexts into Italian, Spanish, French, and German. Our results show that TDE attacks against both English-centric and multilingual models transfer to different languages: the attacks are successful on translated prompts, even though only the original English prompt might have been included in the pre-training data. A web-presence check on a sample of the translations confirms that they are not available online. The share of English leaks recovered in other languages grows with the multilingual capability of the model, and it drops sharply when the original wording is lost, even without a change of language. This suggests that native multilingual pre-training facilitates the emergence of latent cross-linguistic bridges that simplify the retrieval of personally identifiable information (PII). We analyze the activations of multilingual large language models (LLMs) and find that different translations of the same prompt are bridged in similar representations, with the strongest alignment in the middle layers. Our results highlight a fundamental security gap in modern LLMs, necessitating more robust, language-agnostic sanitization strategies for future model alignment.

We study the verification problem for deep narrow ReLU neural networks: given a network of bounded width computing a piecewise-affine map on [0,1], does some input satisfy a prescribed output constraint? Classical NP-hardness proofs for ReLU verification use one neuron per Boolean variable and say nothing about networks of small constant width, while width-1 networks are easy to verify. We show that verification of ReLU networks is NP-complete at width 4 for arbitrary inputs in [0,1]. When inputs are restricted to a natural discrete encoding set, NP-completeness already holds at width 3. Together with polynomial-time decidability at width 1, this leaves open only width 2 on the encoding set, and widths 2 and 3 on [0,1]. The technical core is a fractal preprocessing gadget: a width-2 ReLU subnetwork whose iterate vanishes precisely near a finite Cantor-like subset of [0,1] with 2^n points. It reduces verification of a continuous function on [0,1] to verification on 2^n discrete points without increasing the width, and is the missing ingredient for width-bounded hardness reductions. The same construction yields further results at width 3 on the encoding set: the universal problem is coNP-complete, counting zeros is #P-complete, a majority variant is PP-complete, and approximating the minimum output within a constant gap inherited from Max-3Sat is NP-hard. The NP, coNP and inapproximability results lift to all of [0,1] at width 4; lifting counting and majority, and lifting at width 3, remain open.

We present here the FetalSenseM V2 (FSM V2), a wireless and wearable, multi-wavelength, NIRS device, based on LED illumination at six NIR wavelengths (782, 800, 818, 830, 848 and 888 nm). It utilises two sources located on the flexible wings of the sensor body, and three photodiodes placed symmetrically in its centre at 1 cm from each other. The configuration provides two sets of three source-to-detector distances (SDS) at 3, 4 and 5 cm. FSM V2 allows estimation of dynamic, relative changes in placental concentrations of HbO2, HHb and oxCCO. At the same time, the dual set of SDS also enables the use of multi-distance algorithms, such as spatially-resolved spectroscopy (SRS) and Dual Slope (DS), for estimating placental oxygen saturation (PlO2). FSM V2 is the first example of a new generation of wearable NIRS devices, paving the way for smart and advanced monitoring of placental health at bedside. The FSM V2 was successfully validated on a bench-testing, metrological platform based on optical phantoms, as well as against a gold-standard, time-domain (TD) NIRS system, before moving to preliminary examples of clinical application and usability on pregnancy monitoring trials at University College Lon-don Hospital (UCLH), conducted as part of an in vivo proof-of-concept study.

Continuous asynchronous replanning is essential for real-time generative robot policies, but independent stochastic initialization can cause mode switching and inconsistent continuation across action chunks. We propose Execution-Aligned Progressive Noise (EAPN), which introduces structured stochasticity at both inter-chunk and intra-chunk levels. Across replanning steps, EAPN propagates a shared noise trajectory and aligns it with the actual execution displacement, establishing execution-aligned inter-chunk correlation. Within each action chunk, it models temporal correlation along action time. The aligned stochastic history is further combined with committed action context to condition subsequent generation, allowing new chunks to continue from execution-consistent generative states rather than restart from independent noise. We evaluate EAPN on D3IL, Kinetix, LIBERO, and real-world manipulation tasks. EAPN improves multimodal behavior consistency on D3IL and achieves an average success rate of 88.59% on Kinetix. On LIBERO, it remains robust and maintains strong task performance even under long inference delays. Real-robot experiments further achieve 90.0% success on Object Storage and 96.7% on bimanual Cloth Folding, demonstrating reliable continuous execution under asynchronous replanning.

Diffusion- and flow-based robot policies have recently become widespread in robotic Imitation Learning (IL) due to their high performance and ability to model continuous and multimodal distributions. However, the iterative denoising procedure used by these models introduces significant prediction latency, hindering high-frequency closed-loop robot control and leading to jittery, unstable motion when frequent updates to the robot's action predictions are used. Therefore, it is common practice to train models to predict chunks of actions that can be executed sequentially without feedback, even when this reduces responsiveness and may mean the most recent state information is not used. In this article, we present Adaptive Mean Flow (AMF), a flow-based IL method that enables smooth and responsive, fully closed-loop robot control. AMF uses Mean Flow, which is an accelerated form of Flow Matching (FM), to minimize prediction latency. To ensure smoothness and consistency across predictions, AMF uses a corrupted version of the trajectory from the previous step when predicting new robot actions, with the signal-to-noise ratio increasing over the time parameter of the trajectory. This discourages large changes in the prediction from one step to the next, while allowing freedom to adapt the predictions for future steps. We evaluate AMF across a wide range of simulated and real robot tasks and demonstrate significantly improved performance compared with baselines. Code: https://github.com/akselva/Adaptive-mean-flow-RoboticIL.

Personality is associated with economic outcomes, but whether these relationships are causal remains unclear. We exploit within-family variation in genetic propensity toward the Big Five generated by random genetic transmission from parents to offspring. Combining polygenic indexes from the Lifelines cohort study with 19 years of administrative tax records in the Netherlands, we find that greater genetic propensity toward emotional stability increases earnings and household wealth, while genetic propensity toward conscientiousness increases household wealth. The other three traits show no consistent pattern across outcomes and specifications. The effect of genetic propensity toward emotional stability on earnings remains positive as we progressively account for genetic propensity toward educational attainment, cognitive performance, and depression. These findings strengthen the evidence that personality causally affects earnings and wealth.

Game-theoretic attribution explains a model by assigning credit to its features or training examples. The least core has attracted interest as an alternative to Shapley-style averaging because it can expose players that cause substantial harm in rare, high-value contexts. However, least-core allocations are generally nonunique, and the choice of allocation can affect the resulting explanation. In this study, we investigate how payoff selection and coalition sampling affect least-core attribution. Our experiments show that selector choice matters for distinguishing useful and harmful contributions, and that sampling can degrade harmful-player identification across the tested selectors even when useful players remain well identified. These observations motivate efficient computation with all coalition constraints and a well-defined selector. We introduce entropic least core (ELC), a smooth approximation whose unique minimizer follows a continuous path along the temperature to the nucleolus, a classical refinement of the least core. Our experiments show that ELC approximates the nucleolus faster than an LP-based nucleolus solver while retaining small payoff errors, with further GPU acceleration at larger problem sizes. In the tested full-coalition contextual-distractor games, ELC matches the minimum-norm selector in identification accuracy and more accurately ranks distractors by harm.

Learning directed acyclic graphs (DAGs) to uncover causal mechanisms has attracted substantial attention in machine learning. While most existing methods focus exclusively on observed variables, many variables of substantive interest are latent constructs defined by statistical measurement models. Such constructs are particularly common in the social and behavioral sciences. In this paper, we propose a general statistical framework for DAG learning when some or all nodes of the graph are latent variables, accommodating non-Gaussian manifest variables such as binary and categorical data in the measurement model. To overcome the computational burden of high-dimensional integration in standard marginal likelihoods, we develop DAG learning via Composite-Likelihood-based screening and Iterative Pruning (DAG-CLIP), a two-step composite-likelihood-based learning algorithm. This algorithm first solves a smooth acyclicity-constrained optimization problem to screen out misspecified DAG structures, and subsequently performs BIC-guided composite-likelihood backward deletion within the Markov equivalence class (MEC) space to identify a sparse DAG. We establish the statistical consistency of the algorithm in recovering the MEC of the true DAG, and demonstrate its effectiveness through extensive simulation studies and a real-world application to large-scale educational survey data.

We present a computational design study of a stellarator-mirror hybrid concept, Stellarators Linking Axisymmetric Mirrors (SLAMs), in which an optimized poloidally omnigenous (OP or QI) stellarator serves as the rotational-transform-generating element that links long, axisymmetric (circular cross-section) magnetic mirrors. Starting from an OP precursor configuration obtained from the DESC omnigenity database, we construct hybrid coil sets for field period nFP = 2 SLAMs by separating the OP modular coils along the device midplane and inserting axisymmetric planar mirror coils into the resulting gap. Field-line tracing shows that nested vacuum flux surfaces survive this extrusion across a range of mirror-section lengths, with the cross-section qualitatively preserved at the point where the mirror is inserted. We use the field-line tracing data to map the shape of a surface and use the GVEC fixed-boundary solver to calculate an ideal magnetohydrodynamic (MHD) equilibrium and evaluate physics metrics such as the neoclassical transport proxy ($\epsilon_{\rm eff}^{3/2}$), the vacuum magnetic well (W), and the gyrokinetic heat fluxes (Q), providing a first-principles analysis of such concepts using modern stellarator analysis tools.

More people across the UK are today exposed to more AI-generated and AI-altered disinformation and misinformation than ever before - exposing individuals and society to a growing range of harmful consequences and potential consequences. We analyse evidence from a dataset of 112 pieces of AI-generated or AI-altered disinformation or misinformation seen tens of millions of times across the UK between 1 January 2025 and 31 March 2026. The dataset of examples studied does not, of course, provide an exhaustive sample of all AI-generated false information in circulation in the period, but rather a snapshot of examples showing some, but not all, of the potential effects. Our analysis of this sample found a substantive risk of causing or contributing to harm to individuals and society in eight distinct fields, from contributing to incidents of serious social unrest and vigilante violence to causing direct harms to health and causing the sort of serious financial loss that can be caused by online scams and fraud. Other fields of risk included: abuse serious enough to affect individuals' health and behaviour; public engagement with the police and justice systems; susceptibility to false conspiracy theories with the potential to cause direct harms; and broader changes to social and political attitudes with potential to affect political, social events over the longer term. More than four in five pieces of content we assessed added to reasons for the public to distrust information as not merely inaccurate but substantively false or misleading: a broad disinformation effect with potential for significant effects for society.

Most adversarial attacks on deep reinforcement learning (DRL) assume white-box access to the victim policy, which rarely holds in practice. This paper studies transfer-based black-box attacks on DRL: the attacker crafts observation perturbations on a white-box surrogate agent and feeds them to an unknown victim. We formulate the attack as return minimization under a per-step perturbation budget. We first show that transplanting transferable image-classification attacks (FGSM, MI-FGSM, and NI-FGSM) with a per-step objective yields perturbations that transfer but are no stronger than random noise of the same budget. We then propose a trajectory-level attack that optimizes a sequence of perturbations over a receding horizon through a differentiable model of the environment and a temperature-smoothed surrogate policy, with the same optimizers. On CartPole-v1 with ten DQN and DDQN agents and 100 surrogate–victim pairs, the trajectory-level attack outperforms per-step attacks and random noise in the white-box, cross-model, and cross-algorithm settings.

Modern asteroid surveys routinely acquire sparse photometric measurements for hundreds of thousands of asteroids. These data can be exploited to reconstruct asteroid spin states and shapes, providing statistical constraints on the rotational evolution of different dynamical populations.Our goal is to derive spin-state and shape models from publicly available survey photometry and use the enlarged sample to investigate the distributions of asteroid physical properties across the Solar System. We combined sparse photometry from Catalina, Mt. Lemmon, Pan-STARRS, ZTF, USNO, ATLAS, ASAS-SN, TESS, and Gaia and applied convex light-curve inversion to determine sidereal rotation periods, spin-axis orientations, convex shapes, phase-function parameters, and colors. The extensive parameter-space search was performed using the distributed-computing project Asteroids@home.We derived approximately 30000 asteroid models, including about 22000 unique spin-state solutions and almost 10000 partial pole solutions, tripling the number of available asteroid models. The enlarged sample confirms the dependence of the rotation-period distribution on taxonomic class, reveals a pronounced lack of elongated asteroids near the spin barrier, and shows that slow rotators exhibit systematically larger photometric residuals, supporting the interpretation that many are in non-principal-axis rotation. We derived spin-state distributions for 47 near-Earth asteroids, 247 Jupiter Trojans, and 157 Hildas. The expanded Trojan sample no longer shows a statistically significant excess of prograde rotators. The combination of modern survey photometry and large-scale light-curve inversion enables population studies based on tens of thousands of asteroid models. The resulting database provides improved constraints on the rotational evolution of asteroids

We construct explicit off-diagonal Ramsey graphs from polynomial families of incident point–hyperplane flags. Universal interpolation translates the exact dimension of ordered clique configurations into a bound on their coefficient–label incidence. At the critical dimension, a polynomial separates the forbidden-pair image from its diagonal; deleting the corresponding edges preserves a triangular rank certificate. Restriction of scalars realizes the rational endpoint without rounding loss. The resulting fixed-$s$ exponent has leading scale $s/(2\log_2s)$, and short Frobenius relations give the explicit example $R(16,t)\ge\Omega(t^{2.0539221767\ldots})$. A finite-fiber construction gives the range $s^2(\log s)^2=o(\log t)$. For fixed parameters, algebraic preprocessing terminates and field initialization, vertex decoding and adjacency take deterministic time polynomial in the extension degree. We also retain an elimination-free variant and formulate the filtering argument for general forbidden configurations with a distinguished edge.

Gaussian-process limits are powerful in describing overparameterized machine learning models, yet their validity in structured tensor-network architectures remains unclear. Here we analytically present a moment-based approach that identifies precise conditions for the emergence and breakdown of Gaussian universality in tensor-network learning models, with a focus on matrix product states. We prove that in the large bond dimension limit, the learning models with both local and global observables converge to Gaussian processes, with explicit finite-size bounds on higher-order moment deviations. Whereas in the large physical dimension limit, the Gaussian universality no longer persists: while the models with local observables retain Gaussian-process behavior, those global cases exhibit persistent non-Gaussian corrections. Our results reveal that Gaussian-process behavior in tensor-network learning is controlled not only by parameter number, but also by architectural scaling, observable locality, and the spectral properties.

To account for the soft transverse dynamics overlooked by the $k_T$ factorization formalism of exclusive QCD processes, we propose introducing the intrinsic transversal momentum distribution functions (iTMDs), in conjunction with the light-cone distribution amplitudes (LCDAs), to define the probability amplitude for finding a hadron state where partons move fast along the longitudinal axis while undergoing gentle oscillations in the transverse plane. We implement this framework in the calculation of the electromagnetic form factors of the pion and the proton. Our iTMDs-improved perturbative QCD predictions show good agreement with available experimental data and lattice QCD results. More importantly, we find that higher-twist, especially subleading-twist contributions, are indispensable for explaining the scaling behavior of both pion and proton form factors at intermediate momentum transfers. Furthermore, the inclusion of iTMD effects significantly improves the perturbative QCD predictions for the pion form factor in the few GeV$^2$.

This paper asks whether VLA generalization is grounded in a global understanding of objects' physical properties that enables policies to adapt their motion to unseen setups, or if policies simply replay the motions they've learnt that happen to succeed in new setups. The former reflects genuine generalization; the latter reflects incidental robustness. We first examine awareness of physical properties in seven VLAs by applying linear probing and representational similarity analysis (RSA) to their activations. We find that physical properties, including mass, fragility, deformability, friction and size are less decodable than non-physical properties such as semantic category, material, sound and price in nearly every modality stream. Compared with their base VLMs, robot pre-training weakens the linear encoding of physical properties in the language stream. Neither pre-training nor downstream fine-tuning strengthens the alignment between physical-property differences and activation distances. We then ask whether the weak physical information present in these activations shapes the actions a VLA generates. In a controlled LIBERO case study, we increase the mass of an in-domain object and signal the change through language or vision. Most VLAs use similar lifting behaviour for the heavier and original-mass objects, leading to task success declines. The few exceptions change their behaviour in response to lexical or visual cues rather than to mass itself. These results suggest that VLAs encode physical properties weakly and do not reliably use them to adapt their motion.

Impact-driven atmospheric loss plays a key role in shaping the diversity of planetary atmospheres. However, existing prescriptions for calculating loss from a given impact neglect the critical role of atmospheric mass on loss, and thus often severely miscalculate the extent of loss. Using 3D smoothed particle hydrodynamics simulations, we quantify atmospheric loss during giant impacts onto planets with 1-20% mass fraction atmospheres and show that less massive (lower-pressure) atmospheres are significantly easier to remove. We derive a new scaling law for atmospheric loss that includes the effects of atmospheric mass and is applicable to collisions between volatile-poor planetary embryos through to super-Earths and sub-Neptunes. Applying our scaling law to $N$-body simulations of solar system formation, we show that including the coupling between atmospheric mass and loss efficiency results in significantly more atmosphere removal. Substantial atmospheric loss events are common for smaller planets ($\lesssim$0.6 M$_{\oplus}$), whilst even roughly Earth mass bodies lose 30-100% of their atmospheres through multiple impacts. We also show that atmospheric loss during the Moon-forming impact strongly depends on both the pre-impact atmospheric mass and impact style. For a representative $\sim$100 bar pre-impact atmosphere, Earth could have lost between 20-80% of its atmosphere. In any system that undergoes a phase of giant impacts, the atmospheres of its host planets will be unavoidably shaped by the stochasticity of impacts. Our new scaling law can be readily incorporated into models of planet formation and system-wide evolution, offering new insights into the origins of planetary diversity.

Preparing quantum states from classical data can cost more than the computation they serve; most loaders tailor a circuit to each input. Here we show that the signals of a real dataset share structure that can be learned once and reused. Our quantum-native loader learns a low-dimensional description of a dataset and prepares every signal with one fixed circuit set by a few numbers. Across seven views of five public datasets it meets the targets of the strongest structured loader at equal gate cost with several times fewer numbers per signal. These numbers can be inferred from a random subset: in a preregistered blind replication the subset needed to come within ten per cent of full-signal accuracy stayed constant within a prespecified margin as signals grew sixteenfold, whereas the structured loader needed ever more. It declines what it cannot represent, covering fewer cases than that baseline and no electrocardiogram.

Hadron resonances are intrinsically dynamical states whose observable properties are governed by the interplay between quark-model configurations and hadronic channels. Hamiltonian effective field theory provides a practical framework for connecting hadron dynamics with experimental observables and lattice QCD results. In this review, we outline the infinite- and finite-volume formulations of Hamiltonian effective field theory and survey its applications to light-hadron spectroscopy and resonance structure. By constraining the same Hamiltonian framework with experimental scattering data and lattice QCD spectra, Hamiltonian effective field theory can help distinguish among scenarios that describe the available scattering data comparably well but imply different roles for bare-state and multihadron configurations. We also review recent extensions of the Hamiltonian framework developed to address more complicated dynamical settings.

Voice anonymization (VA) is used to conceal the voice identities in speech data. Performance is usually estimated using automatic speaker verification (ASV) as a proxy to judge the capability of an attacker to re-identify speakers after anonymization. The defender anonymizes test utterances; the attacker compares them to equally anonymized reference utterances to infer voice identity, with degraded ASV performance indicating successful anonymization. Thus, the defender's protection is one-sided and only applied to test utterances via VA. Though unexplored so far, there is an opportunity to enhance anonymization by protecting reference utterances too. We present a new, preemptive VA paradigm: reference utterances are protected using adversarial noise to degrade their potential to infer voice identity. Preemptive protection does not degrade the perceived quality of reference utterances and is independent of the specific ASV system used for identity inference. By combining preemptive protection and regular test-side VA, ASV equal error rates can be increased from 14% to 45% (near-perfect privacy).

Investigating the statistics of the sum of random variables (RVs) is fundamental in wireless communication systems, as it can be used to accurately estimate their performance. In this paper, exact closed-form expressions are derived for the probability density function (PDF) and cumulative distribution function (CDF) of the sum of inverse-gamma (IG) RVs. The IG distribution is especially suitable for performance analysis purposes because it provides a tractable yet accurate model of shadowing/local mean power variations. Unlike prior studies on sums of shadowing RVs, the resulting formulas are simple to evaluate and numerically stable. Their utility is demonstrated in two practical settings: unmanned aerial vehicle (UAV)-assisted networks and high-speed railway communications. Across all considered scenarios, the analytical results are in excellent agreement with the Monte Carlo simulation results, while the asymptotic approximations closely track the exact results

We classify, up to graded isomorphism, graded bimodules over split group algebras of finite abelian subgroups of an arbitrary group. The classification is governed by double cosets, stabilizer characters, and coefficient modules. Over a field, we determine the essential image of the induction functor on Yetter-Drinfel'd modules, giving a negative answer to a question of Dos Santos and Yasumura in the nonabelian case. We also compute the Peirce components and characterize multiplicity-free bimodules through extensions of stabilizer characters.

Magneto-mechanical resonators (MMRs) have emerged as promising passive sensors for wireless localization and remote inductive sensing in industrial and medical environments. However, fundamental limits on achievable localization accuracy using the spatial sensitivity profiles of receive coils remain a key challenge. In this work, we demonstrate that an MMR - comprising two permanent magnets in an anti-parallel arrangement - exhibits a measurable shift in its mechanical resonance frequency in response to an externally applied magnetic field, a property we exploit for spatial encoding. We derive and experimentally validate a method that maps frequency shifts of 1D positions within a known magnetic gradient field, demonstrating accurate localization along a linear trajectory. This gradient-field frequency encoding approach constitutes a conceptually distinct alternative to conventional sensitivity encoding with receiver coil arrays, with the potential to simplify hardware requirements while improving localization accuracy.

Normalized gradient descent is a widely studied adaptive optimization method. Most existing analyses focus on the best iterate or a weighted average of the iterates, whereas practical implementations typically return the last iterate. In this paper, we study the last-iterate convergence of normalized gradient descent for convex, $(\nu,M_\nu)$-Hölder-smooth objectives. For a constant stepsize, we establish an upper bound of $\mathcal{O}\bigl((\log^2(T)/T)^{(1+\nu)/2}\bigr)$, which contains a logarithmic overhead relative to the known $\mathcal{O}\bigl(T^{-(1+\nu)/2}\bigr)$ guarantees for the best and weighted-average iterates. For $\nu = 0$, this overhead is known to be unavoidable. We complement this analysis with numerical results based on the performance estimation problem (PEP), investigating the finite-horizon worst-case behavior in the smooth setting and whether the logarithmic overhead reflects an intrinsic limitation of constant-step normalized gradient descent. We then show that a linearly decreasing stepsize yields a last-iterate guarantee of $\mathcal{O}\bigl(T^{-(1+\nu)/2}\bigr)$, matching the order of the best-iterate/weighted-average guarantees without requiring knowledge of $\nu$ and $M_\nu$.