Papers

This pictorial presents The Body in Algorithmic Colonialism, a screen-based interactive installation that critically examines Emotion-AI through artistic practice. The work uses facial tracking to simulate how commercial Emotion-AI systems translate facial movement into emotional categories and verdicts of compliance or non-compliance. Rather than treating explainability as technical transparency alone, the installation approaches XAI as an embodied encounter with misrecognition: viewers experience how the body is measured, classified, and returned to itself through algorithmic thresholds. Drawing on critical AI studies, affect theory, and practice-as-research, the pictorial positions artistic practice as a method for making affective classification visible and contestable. It documents the conceptual framework, technical process, and critical context of a prototype developed in the first year of a practice-based PhD. No audience responses, interviews, or participant data are analysed in this submission.

Learning world models across diverse domains calls for a unified control space, yet ac- tion interfaces vary across domains and action annotations are often scarce. Latent ac- tions learned directly from videos offer a promising way to establish such a space without explicit action supervision. However, learning shared latent actions across domains re- mains challenging, as variations in appearance, embodiment, and dynamics can entangle domain-specific visual cues with transferable transition structure. To this end, we intro- duce OneLA, a simple framework that learns continuous latent actions through a shared orthogonal dictionary and a common predictive objective, without task-specific designs or cross-domain alignment losses during latent action pretraining. These latent actions pro- vide a shared conditioning interface for world models, with lightweight adapters mapping native control commands into the same space. Conditioned on latent actions and the cur- rent visual state, the world model predicts context-dependent transitions for controllable rollout. OneLA improves action readout across domains and reduces action cycle-transfer rotation error by 24.8% over the strongest evaluated baseline. A joint-training comparison further shows a 20.5% improvement in Human Preference Score v3 (HPSv3) for real- scene generation. These results suggest a path toward general-purpose world models that learn from diverse videos and support flexible control through latent action space.

We revisit the effective radial Schrödinger equation for a charged spinless particle in collinear static electric and magnetic fields that was formulated in our earlier work within the tridiagonal representation approach (TRA). In that model the angular factor of the electric coupling is replaced by the orientation parameter $C=\cos\theta$, the linear Zeeman term is diagonal in the azimuthal quantum number, and the effective field-modified central potential takes the form $V(r)=\Vt(r)+qECr$. Here we show that the same equation can be solved exactly without the TRA, by the Nikiforov–Uvarov (NU) method, once the potential component $\Vt(r)$ is prescribed. Three examples are worked out in closed form: a Kratzer-type component, which yields a linear-plus-Coulomb-plus-inverse-square central potential; a pseudoharmonic-type component, which yields a linear-plus-harmonic-plus-inverse-square potential; and a Hulthén-type component with a repulsive core (Manning–Rosen form), which yields a screened confining potential with a finite number of $s$-wave bound states. For each case we obtain the energy spectrum, the normalized eigenfunctions in terms of Laguerre or Jacobi polynomials, the location of the potential minimum, and closed-form radial moments. The spectra are independent of the orientation $\theta$ and of the electric field, in agreement with the structure found by the TRA, while the field enters through the shape of $V(r)$, the radial moments, and the expectation value $\langle V\rangle_k$. All analytical results are confirmed by an independent finite-difference diagonalization.

Self-supervised joint embedding learning aligns representations of augmented views while regularizing their distribution to prevent collapse. Regularization aims to shape the overall representation distribution, but most joint embedding methods restrict its computation to the active minibatch. We propose out-of-batch regularization, which maintains a dataset-level state in CPU memory and uses globally constructed per-instance targets during minibatch training. Evaluated on natural images and biomedical datasets, these anchors approximate the full-dataset regularization gradient more faithfully than conventional minibatch estimation at small batch sizes. Applications to contrastive (SimCLR) and non-contrastive (SIGReg) approaches show consistent improvement over the corresponding in-batch objectives for image classification, minority-class performance under controlled imbalance, instance retrieval, and continued self-supervised pretraining, with the largest gains at small batch sizes. These results suggest that dataset-level regularization can be decoupled from minibatch optimization, making batch size an optimization choice that does not affect the geometry of the regularizer.

World Action Models (WAMs) typically adapt video foundation models to robotic control through joint video and action prediction. However, task-level instructions may leave the next physical interaction underspecified, requiring semantic grounding and state-dependent motion selection. To this end, we introduce Atomic-WAM, an understand-then-execute framework that connects task reasoning to action generation through atomic-motion descriptions. Specifically, a vision-language reasoner describes the next local motion from the task instruction and visual context, guiding a WAM to predict future video and an action chunk. Furthermore, an automatic annotation pipeline curates temporally aligned, contact-enriched supervision from multi-view DROID demonstrations. To make atomic-motion predictions executable by the WAM, we post-train the reasoner with Execution-Grounded GRPO (EG-GRPO). With the WAM fixed, EG-GRPO adapts the reasoner through rewards for motion appropriateness and execution success. Finally, Atomic-WAM achieves 39.6% task success under default instructions and 31.6% under vague instructions in a real-to-sim evaluation on RoboLab. Project page: https://ssj9596.github.io/Atomic-WAM-project

Higher-order curvature corrections to general relativity arise naturally in many fundamental theories and provide phenomenologically interesting low-energy effective field theories. However, higher-derivative terms can introduce ghost modes, which call the validity of such models into question. Here, we thoroughly investigate the classical linear stability of the most general second-order scalar-tensor theory, namely the Horndeski theory of gravity, augmented by a Weyl-squared term in the action. We consider shift-symmetric de Sitter and power-law FLRW backgrounds and study the evolution of linear perturbations. Vector perturbations are free of gradient and tachyonic instabilities for a positive Weyl-induced mass squared. Tensor perturbations necessarily contain a ghost; their superhorizon branches are bounded, while a frozen-background calculation identifies a separate finite-momentum stability condition. The scalar sector instead contains a growing long-wavelength branch for all expanding backgrounds we consider, which is inherited by at least one of the gauge-invariant Bardeen potentials. We show that this growth enters the Bardeen potentials through an inverse-gradient factor $a^2/k^2$: the dynamical scalar fields remain bounded on the inflationary backgrounds, the direct contribution of the growing shear to the clock-matched tidal field cancels, and the electric part of the Weyl tensor decays. We discuss the implications of our results for perturbative control and for the phenomenological viability of theories including quadratic curvature corrections.

Banks rely on extensive internal policy frameworks to ensure that corporate lending decisions comply with regulatory requirements and internal risk standards. Reviewing loan applications against these policies is a largely manual and knowledge-intensive process, which can lead to overlooked policy violations or exceptions. This report presents a prototype of an LLM-based policy checker that automatically verifies loan applications against internal lending policies. The proposed approach follows a two-step architecture. First, compliance-relevant rules are automatically extracted from policy documents and converted into structured, machine-readable representations. Second, loan applications are evaluated against the extracted rules to identify policy exceptions, non-compliant findings, missing information, and cases requiring manual review. Evaluation results demonstrate that the rule extraction process is highly stable across multiple runs and consistently captures rules identified as critical by a domain expert. When evaluated on 11 real loan applications with known policy exceptions, the system successfully detected the majority of these exceptions and identified additional compliance issues that had previously been overlooked during manual reviews. While the approach produces false positives and cannot fully replace human judgement, the results indicate that LLM-based policy compliance verification can provide valuable support in the loan application verification process by highlighting potential compliance risks.

Fluorite-structured ferroelectrics, such as hafnium zirconium oxide ($Hf_{1-x}Zr_{x}O_{2}$), are among the most promising materials for developing the next generation of low-power nonvolatile memories. Emulating realistic operating scenarios for ferroelectric memories requires performing access rate-dependent endurance tests. The access rate-dependent reduction of the switched polarization cannot be described as a mere superposition of fatigue and imprint effects, which are observed in conventional endurance and retention tests, respectively. The interplay between these degradation mechanisms and its correlation with the delay time in between the programming pulses remains unclear. Combining electrical testing through voltage pulses with ab initio and drift-diffusion simulations provides a comprehensive perspective on the microscopic origin of the access rate-dependent reduction of the switched polarization of ferroelectric capacitors based on product-relevant 5 nm thick $Hf_{0.5}Zr_{0.5}O_{2}$ films. The role of oxygen vacancy defects accumulating at the electrode-ferroelectric interface with delay time in reducing the fraction of switchable domains is elucidated considering electron transfer processes which change the defect charge state. Building upon this theoretical understanding, a physics-based model to extrapolate the long-term access rate-dependent endurance performance is developed and experimentally verified for temperatures up to 135{\deg}C and delay times up to 2 ms.

Generalized complex structures give rise to generalized Dolbeault, Bott–Chern and Aeppli cohomologies. We study these cohomologies for invariant generalized complex structures on almost abelian Lie algebras $\mathfrak g_A=\mathbb R e_0\ltimes_A\mathfrak h$, where $\mathfrak h$ is an abelian ideal of codimension one, and on almost abelian solvmanifolds $\Gamma\backslash G_A$. We first show that the invariant generalized Dolbeault cohomology can be identified with the Chevalley–Eilenberg cohomology of the complex Lie algebra $\mathcal L$, the $+{\mathbf i}$-eigenspace of the structure, with coefficients in a character determined by the canonical line. We prove that $\mathcal L$ is itself an almost abelian Lie algebra; thus the cohomology reduces to the kernel and cokernel of an explicit derivation. Assuming that $A$ is diagonalizable over $\mathbb R$, we obtain a combinatorial formula for the cohomology dimensions and compare with compact quotients. Other comparison results are discussed, noting that the cohomology of invariant forms always injects into that of $\Gamma\backslash G_A$.

We demonstrate that a single repulsive impurity enhances directional confinement of a high-energy wavepacket in a finite two-dimensional lattice of quantum wells. At energies far above the interwell barriers, the impurity suppresses transverse dynamical tunneling while preserving propagation along a principal lattice axis, even as the corresponding classical regular region shrinks. Confinement is strongest at impurity amplitudes comparable to the barrier height and weakens for stronger perturbations. Spectral analysis links this nonmonotonic pinning to the wavepacket's overlap with eigenstates concentrated along the launch direction, consistent with impurity-induced variational scarring. These results show that a single local defect can control quantum propagation over distances far exceeding its own spatial extent.

We investigate directional quantum steering harvesting by two uniformly accelerated non-identical Unruh-DeWitt detectors coupled to a massless scalar field in the Minkowski vacuum. We show that the Unruh effect does not simply degrade harvested steering, but can reshape its directionality in a strongly parameter-dependent manner. At low detector energy scales, acceleration predominantly suppresses steering, whereas at higher energy scales, moderate acceleration can enhance steering and even activate an initially absent steering direction. Most strikingly, within a finite acceleration window, the Unruh effect enhances steering in one direction while simultaneously suppressing it in the opposite direction, revealing an intrinsically directional influence of relativistic acceleration. Detector non-identicality provides an additional control mechanism: a moderate energy-gap mismatch can enhance or activate one steering direction while suppressing the reverse one, whereas sufficiently large energy-gap mismatch suppresses the harvested steering. Moreover, the maximal steering asymmetry approximately marks the transition from two-way to one-way steering. Our results establish relativistic acceleration and detector non-identicality as means of controlling the strength and directionality of vacuum-harvested quantum correlations.

Robotic vacuum cleaners must extend their functionality within tight limits on size, cost, and durability. Deployable linkages with few actuators meet these limits, but their backdrivability makes it difficult to hold a load-bearing posture without continuous motor torque. We introduce a two-stage deployable linkage that is self-locking, in the sense used for toggle and singularity-based linkages, through its own geometry: at full deployment an internal three-link subchain approaches a collinear, near-singular configuration where the input–output transmission ratio approaches zero, so that plate loads produce a small residual back-driving torque, absorbed by drive-train friction and a lightweight magnetic detent. A single actuator performs deployment and locking, a passive chassis-mounted trigger releases the lock for retraction, and a second actuator adjusts only the deployment height. We integrate the mechanism into a commercial robotic vacuum cleaner and validate it through a 100-cycle repeatability test, a quasi-static load test in which the locked plate held 50\,N with the motor de-energized, and 140 pushing-type tidying trials on seven household objects against a commercial arm-equipped vacuum robot under a documented protocol. We derive the analytical holding condition and discuss its limits as an intermediate design point between passive compliance and articulated manipulation.

A graph $H$ is called common if the number of monochromatic copies of $H$ in a $2$-edge-coloring of a large complete graph is asymptotically minimized by the random coloring with an equal proportion of each color. The house graph and $K_4$ with one edge subdivided once are known to be common by the flag algebra method. Grzesik, Lee, Lidický, and Volec asked for “human-friendly” proofs of the commonality of these two graphs. In this short note, we give flag-algebra-free proofs that both graphs are common.

We propose a loss-tolerant measurement-device-independent quantum key distribution (MDI-QKD) protocol using time-bin-multiplexed single photons entangled with local qubits stored in nitrogen-vacancy (NV) centers. In each round, Alice and Bob each send a single multiplexed photon to an untrusted intermediary, Charlie. Each photon first becomes entangled with Charlie's NV-center qubits; Charlie then performs entanglement swapping between his NV qubits to establish EPR pairs between Alice's and Bob's local qubits. With $d$ time bins per photon, $\log_2 d + 1$ EPR pairs can be generated per transmitted photon, substantially increasing the probability of EPR-pair creation under limited channel bandwidth. To further mitigate the effects of prolonged memory storage, the protocol employs a memory-insensitive (MI) architecture, where multiple banks of qubits and multiplexed photon routing reduce the effective memory lifetime requirements. This approach enhances loss tolerance, enabling higher secret key rates and secure key distribution over longer distances compared to conventional single-photon protocols. Using the Lo-Chau EPR-QKD framework and Shor-Preskill reduction, we prove that the protocol is secure without trusting intermediate nodes, making it suitable for quantum networks with untrusted repeaters.

Echo State Networks (ESNs) are fundamentally important from the conceptual perspective of AI, because they provide simple and tractable frameworks for understanding temporal information processing, memory, and learning in dynamical systems. They are composed by an interconnected reservoir whose connectivity matrix remains constant over time and only the output weights are updated during the training period. In particular, ESNs employing online training allow the incremental adaptation of the output weights as data is received, making them particularly suitable for applications in which new input data becomes available over time (non-stationary environments). A key objective in the study of such learning structures is to determine which statistical properties of the reservoir enhance computational efficiency and maximize prediction performance. To this end, we investigate the effects of reservoir topology on learning performance, considering weight matrices generated from Erdos-Renyi random networks and Watts & Strogatz small-world networks. As test data we use the Lorenz and Rossler models in the chaotic regime and the FitzHugh-Nagumo model in the limit cycle regime. In the case of random Erdos-Renyi reservoirs, our simulations show that best performances are obtained for sparse reservoirs and especially when the link-density is near and immediately above the critical point of mean-field bond percolation. Watts & Strogatz reservoirs are shown to demonstrate best performance also for small values of link-density, compatible with the mean-field bond percolation when taking into account the specific nonlocal ring structure of the Watts & Strogatz reservoirs. These studies on the influence of the reservoir architecture and hyperparameters on enhancing the accuracy of ESN predictions provide an efficient framework for real-time forecasting in dynamically evolving systems.

Let $W$ be a complex reflection group and $X$ the De Concini-Procesi wonderful model of its reflection arrangement. Inside Etingof's sheaf of Cherednik algebras on $X$ we define a subsheaf generated by the Dunkl operators that extend the standard ones on the hyperplane complement. We conjecture that at $t=0$ the image of the logarithmic vector fields in a natural module over this subsheaf is the sheaf of Gaudin algebras on $X$ defined by Aguirre, extending a construction of Bonnafé and Rouquier from the hyperplane complement to $X$. We prove the conjecture for $W=S_{3}$.

We propose and analyze a projection-based model reduction method for the rapid and reliable solution of time-harmonic acoustic wave propagation problems over a frequency band, with particular emphasis on seismic applications. The high-fidelity model is based on a hybridizable discontinuous Galerkin (HDG) discretization, formulated as a generalized saddle-point system coupling the volume and trace unknowns. The basis construction combines residual-based greedy sampling of the frequency parameter, enrichment with frequency derivatives of the HDG solution, and proper orthogonal decomposition (POD) compression. We establish the well posedness of both the high-fidelity and reduced problems and characterize their stability with inf-sup constants. We investigate the interaction between static condensation, inherent to the HDG discretization, and projection onto the reduced spaces, characterizing its implications for the reduced formulation and approximation errors. Since these operations do not commute in general, we derive a priori error estimates for both the trace and volume variables to quantify the gap between the two operations. We further identify a lifting condition on the volume reduced space that eliminates these gaps and yields a trace error estimate independent of the volume approximation error. Numerical experiments on the two-dimensional Marmousi model and a three-dimensional heterogeneous medium, with more than $10^{7}$ unknowns, illustrate the theoretical estimates and show that the reduced model achieves speedups of several orders of magnitude compared to the high-fidelity solver.

News outlets in poorer countries cover the world much as the leading outlets of rich countries do. Two explanations compete. One says they depend on Western news agencies for lack of resources. The other says they choose to copy prestigious outlets or whoever moves first. Observation cannot tell these apart. This study measured how closely 18 national outlets in Bangladesh, Kenya, South Africa, India, Malaysia and Nigeria followed four core outlets in how much attention they gave, which angle they took and when they published, on the wars in Ukraine and Gaza and two UN climate conferences in 2022-2024. It then ran a simulation in which language models acted as the outlets' editors and dependence, imitation and herding were switched on and off. Of nine tests, one held: outlets that relied more on wire copy published sooner after the core did. Tier and local stake showed no effect distinguishable from none. The herding test was supported and the imitation test was not. The simulation did not decide between the mechanisms, because the dependence rule, as fitted, reduced to publishing wire copy. Wire copy sets the clock of the periphery's foreign news. Whether it sets the agenda stays open.

Measurements of optical turbulence are essential in both ground-based astronomy and free-space optical communication systems contexts for multiple reasons including supporting and optimizing Adaptive Optics. Similarly, forecasting these parameters is highly valuable for planning observational campaigns; however, it requires at least two years of continuous data to generate a reliable training dataset. The most commonly used instruments to monitor continuously the optical turbulence include DIMM, MASS, SLODAR, RINGSS sensors. The Arcetri team is leading two operational forecast systems for two different Observatories: ALTA above LBT, and FATE above VLT. Both systems are in continuous evolution and obviously forecast performances depend also on available measurements. Our long-term goal is to develop our own instruments to obtain continuous measurements directly from local observations. In this work, we present an automated system designed for continuous seeing measurements using multiple sensors. This portable station integrates a Cyclope, a RINGSS unit, a scintillometer which operate together to provide comprehensive night-time atmospheric information. We will compare the performance of low-cost sensors, such as the Cyclope and the scintillometer, with that of the RINGSS turbulence profiler, which has been fully installed and programmed by our team in collaboration of LBT team. A dedicated insturment for night and day-time measurements is planned for near future. This system incorporates a fully automatic on-site system that monitors operational status, delivers real-time information on sensor conditions and measurements, and manages synchronization with the data storage unit. The system's real-time acquisition capability enables continuous turbulence analysis, supports R&D activities, facilitates local data generation, and allows on-site testing of forecasting methodologies.

Movable antenna (MA) has emerged as a promising technology for enhancing wireless communication performance by flexibly reconfiguring wireless channels via antenna movement. In this paper, we propose an MA relay-enhanced multiuser communication system that adopts a half-duplex decode-and-forward (DF) relay scheme. The relay node is equipped with an MA array, and a buffer is employed to enhance the end-to-end transmission performance. Specifically, we formulate an optimization problem to maximize the minimum average achievable rate among multiple users over a given transmission period by dynamically optimizing the positions of MAs, transmit power, and beamforming matrices, subject to the constraints on antenna movement, inter-antenna spacing, and total energy budget for antenna movement and transmission. To address this non-convex problem with highly coupled variables, we propose a reinforcement learning (RL)-based algorithm, where a twin delayed deep deterministic policy gradient (TD3) algorithm is utilized to optimize the antenna positions and power allocation, while the transmit and receive beamforming solutions are obtained through closed-form expressions in each time slot. Numerical results validate the efficacy of the proposed RL-based algorithm and demonstrate the superiority of the proposed MA relay-enhanced multiuser communication system over the conventional fixed-position antenna (FPA)-based relay system and other MA-based benchmarks.

In various industries, smoke and fire are two of the most important threats in the workplace. One of the common methods for detecting smoke and fire is the use of infrared thermal and smoke sensors, which cannot be used in outdoor applications. Therefore, the use of vision-based methods seems necessary. The problem of smoke and fire detection is spatiotemporal and requires spatiotemporal solutions. This paper presents a method that uses spatial features along with temporal-based features to detect smoke and fire in the scene. It consists of three main parts; the task of each part is to reduce the error of the previous part so that the final model has a robust performance. This method also uses transformer modules to increase the accuracy of the model. The results of our model show the proper performance of the proposed approach in solving the problem of smoke and fire detection and can be used to increase workplace safety.

We give a randomized polynomial-time $0.48$-approximation for maximizing a nonnegative non-monotone $k$-submodular function under a matroid constraint, for every $k\ge2$. The algorithm combines a top-two label rule with a nested-chain potential and maximum-weight matroid bases. Its analysis reduces the approximation guarantee to a one-dimensional time allocation: a uniform local lower bound leads to a linear differential equation, concavity closes the remaining interval, and two elementary budgets complete the proof. A strict saving in the time budget absorbs sampling and discretization errors, yielding the exact $0.48$ guarantee in the value-oracle model.

Sparse autoencoders (SAEs) have emerged as a promising approach to mechanistic interpretability by decomposing language-model activations into sparse, potentially interpretable features. However, differences in design assumptions, evaluation objectives, and experimental settings complicate conclusions about their effectiveness. We present a critical survey of 138 studies on SAEs in natural language processing. Our contributions are fourfold: (1) we propose a task-grounded taxonomy connecting dictionary design, feature operations, and NLP functions; (2) we examine how sparsity allocation, geometry, data coverage, and inference affect feature recoverability; (3) we assess evidence for semantic interpretability, mechanistic faithfulness, and behavioral utility; and (4) we synthesize findings across multilingual processing, retrieval, reasoning, classification, and behavioral regulation, highlighting the roles of objectives, baselines, supervision, and intervention pipelines. Through comparative analyses of published results, we reconcile apparent disagreements, identify unresolved questions, and propose directions for more rigorous evaluation. An anonymous research register https://anonymous.4open.science/r/sae-research-register-C331/register.html accompanies the survey.

Batch black-box optimization is important for expensive systems where many candidates can be evaluated in parallel, yet only a few rounds of evaluation are affordable. In this regime, effective batch construction must identify multiple promising regions while avoiding redundant evaluations, and conventional Bayesian optimization can require increasingly difficult batch-selection procedures as the batch size grows. We introduce Flow Matching for Thompson-Sampling Refinement (FMTR), an amortized approach that learns offline a continuous, history-conditioned proposal distribution from which variable-size batches can be sampled. FMTR is trained offline by conditional flow matching on Thompson-sampling argmaxes refined toward high-value regions via a horizon-dependent Langevin pushforward map, adapting the exploration-exploitation tradeoff over the optimization horizon. At test time, the model parameters remain fixed and batch generation requires neither task-specific surrogate refitting nor numerical acquisition optimization. FMTR achieves strong performance against batch BO and in-context baselines on synthetic and real-world benchmarks, while substantially reducing the computational cost of batch selection.

Current video object removal methods suffer from prohibitive latency (minutes per video) and non-causal processing that prevents real-time interaction. We present StreamErase, enabling sub-second latency with up to 26 FPS streaming generation on a single GPU. Our approach first augments a text-to-video model with mask input to generate high-quality videos that erase target objects and synthesize coherent backgrounds, but does not perform inference on the fly. As such, we progressively transfer the erasing capability from a powerful bidirectional foundation model to an efficient unidirectional streaming eraser, enabling long-horizon erasing without sacrificing visual fidelity. Several key challenges arise when generating videos over long, potentially infinite horizons: (1) bridging the domain gap from training on finite length and extrapolating to long horizons, (2) sustaining high quality by preventing error accumulation, and (3) maintaining fast inference, without increasing computational cost as the context grows. To address these challenges, we carefully designed sliding-window causal attention with attention sinks. By incorporating self-rollout with attention sinks and KV cache rolling during training, we simulate inference-time extrapolations with a fixed context window, enabling constant-speed generation of arbitrarily long videos. Finally, we establish a dedicated benchmark, LongErase, for streaming long-video object removal. Our model achieves strong performance compared to existing video object erasing models and generalizes well to real-world video scenarios, uniquely enabling infinite-length streaming. With StreamErase, users can input a mask and see the erasing results unfold in real-time, delivering a truly interactive experience.

Vision-language models (VLMs) are moving onto edge devices, but little is known about how much energy they draw on single-board computers with NPU or GPU accelerators, or where that energy goes. We measure total-system energy, latency, and thermal behavior of Qwen2-VL-2B-Instruct and its text-only sibling on a Raspberry Pi 5 with a Hailo-10H NPU and on a Jetson Orin Nano Super, on both CPU and accelerator, and separate the fixed energy of each request from the energy that grows with output length. Because the accelerators use different quantization and serving software, we compare deployment stacks rather than silicon. Three findings stand out. First, NPU offloading makes the Pi a much more efficient VLM platform: it cuts energy per generated token by about two-thirds, leaves the host CPU largely free, and avoids the thermal throttling that the CPU backend reaches in sustained use. Second, an image adds both fixed and per-token energy, but at typical response lengths generating the answer costs far more than processing the image. Third, the Jetson GPU favors short responses through its lower fixed cost per request, whereas the Pi NPU favors longer ones through its lower cost per output token; their energy curves cross near 40 output tokens.

In this paper, we give a family of the arithmetic progressions in which the numbers of odd unimodal and odd strongly unimodal sequences are divisible by 2 and by 4. In particular, we answer a question of Bringmann and Lovejoy on the existence of such arithmetic progressions for odd strongly unimodal sequences modulo 4.

Tactile sensing is becoming central to robust, adaptive, and dexterous robotic manipulation. Beyond its traditional role in contact detection and reactive control, touch is increasingly used to learn transferable representations, condition manipulation policies, and support embodied physical reasoning. Yet existing surveys largely emphasize tactile hardware and perception, leaving the rapid shift toward learning-based tactile manipulation insufficiently synthesized. This paper reviews tactile-guided skill learning from a learning-centric perspective, organized by the functional role of touch across the learning and control pipeline. We examine tactile representation and policy learning, tactile foundation models, predictive and adaptive control, and generalization across objects, tasks, sensors, embodiments, and simulation-to-real transfer. Across these directions, we identify common design principles and remaining challenges in learning transferable physical representations, integrating semantic reasoning with contact-level control, and scaling tactile learning across heterogeneous data and hardware. Finally, we outline emerging directions toward tactile foundation and world models that can predict, reason about, and adapt physical interactions. Together, this review characterizes the transition of touch from a task-specific feedback signal toward a core modality for general-purpose embodied intelligence. An accompanying database of the reviewed literature and associated resources is available at https://sites.google.com/view/touch-in-robot-learning-review.

We present eMOOD, an entropy-guided evolution of the Multidimensional Optimal Order Detection (MOOD) framework for high-order finite-volume methods on unstructured meshes. A novel troubled-cell indicator based on numerical entropy production replaces the classical relaxed discrete maximum principle. Using entropy generation as a physically motivated measure of non-smoothness, the detector targets shocks, contact discontinuities and under-resolved flow features with a compact stencil. Compared with conventional MOOD detection, eMOOD reduces flagged cells by up to 75%, improving efficiency and reducing memory-access overhead. The detector is used within both the classical MOOD cascade of progressively reduced-order schemes and a novel MOOD-CWENO hybrid, which retains inexpensive linear reconstruction in smooth regions and activates nonlinear high-order reconstruction only where needed. Implemented in the open-source parallel solver UCNS3D, the methodology is assessed on the Shu-Osher shock-entropy-wave interaction, double Mach reflection, supersonic forward-facing step, shock-helium-bubble interaction and triple material point problems. These benchmarks encompass strong shock interactions, shock-induced instabilities, multi-material interfaces and vortex-dominated flows, testing robustness, shock capturing and accuracy on unstructured meshes. Results show that eMOOD preserves the accuracy and robustness of reference CWENO and MOOD formulations while achieving speed-ups of up to 20% without loss of solution quality. Its compact detection stencil, reduced memory footprint and improved localisation offer a more scalable and efficient framework for massively parallel high-order compressible-flow simulations. UK Ministry of Defence (copyright) Crown owned copyright 2026 AWE.

Ionizing radiation produces correlated errors in superconducting qubits, posing some challenges to quantum error correction. In addition, it causes background events in other superconducting cryogenic devices, such as kinetic inductance detectors and superconducting nanowire single-photon detectors. Cosmic muons are a component of this background that can only be partly shielded against at ground level, but can be tagged by an independent detector inside the cryostat, so that the affected data can be vetoed, or the muon tags incorporated into error correction. We present a low-cost, compact muon detector consisting of two plastic scintillator modules, each read out by a silicon photomultiplier (SiPM), operated in coincidence. We characterize the SiPMs from room temperature down to 10 mK, including the breakdown and upper-limit voltages that define the operating bias range. Operated as a muon counter, the detector measures a coincidence rate that, as expected, is independent of temperature within uncertainties and agrees with analytical and Monte Carlo predictions. Below approximately 8 K, each muon event produces two pulses, one narrow and one broad, an effect that, to our knowledge, has not been reported before. Coincidences with an independent muon detector outside the cryostat confirm that muons are still counted reliably in this regime. By studying the appearance of double pulses as a function of temperature and magnetic field up to 4.5 T, we find a behavior consistent with the phase diagram of a type-II superconductor. This links the double pulses to a superconducting material in the SiPM, likely an adhesion layer or diffusion barrier introduced during manufacturing. The mechanism by which the superconducting state produces the second pulse is not yet understood.