Papers

Multimodal Large Language Models (MLLMs) consistently struggle with fine-grained visual counting, yet the underlying causes remain poorly understood. In this work, we present a mechanistic analysis of this failure mode, identifying two critical bottlenecks inherent to the global attention pipeline of MLLMs. First, we reveal an individuation bottleneck stemming from image patchification: because Vision Transformers process patches independently, they struggle to group fragmented geometric features across boundaries into distinct object representations. Second, we identify a collapse in the subsequent counting aggregation process, where representation separation rapidly diminishes as numerosity increases due to attention compression. Identifying and formalizing these twin bottlenecks constitutes our first major contribution. To overcome them, we propose ConvStack, a lightweight architecture that operates directly in the visual token space to explicitly aggregate and inject local spatial structures via zero-initialized residual connections. By explicitly addressing the individuation bottleneck, ConvStack provides unambiguous geometric evidence for downstream aggregation. Remarkably, by fine-tuning exclusively on counting tasks, the model achieves substantial improvements in dense object counting and broader spatial understanding benchmarks, without compromising on general visual capabilities.

Imaginary-time evolution (ITE) is a widely used technique for preparing the ground state of a target Hamiltonian. Building on the recently proposed framework of Double-Bracket quantum ITE (DB-QITE), we introduce a probabilistic variant (PDBQITE) that achieves an exponential reduction in circuit depth. Our algorithm requires only a single ancilla qubit and a quantum circuit that consists of one controlled real-time evolution and one mid-circuit measurement per iteration. The trade-off is an exponential decay of the total success probability with the number of iterations, for which we derive an analytical lower bound that is independent of system size. Our numerical simulations show that the per-step success probability remains close to unity, making it realistic to execute many iterations, thus providing a substantial improvement over prior constructions. We show that PDBQITE outperforms the near-term probabilistic algorithm PITE for ground-state preparation of molecular Hamiltonians, and that a hybrid QAOA-PDBQITE scheme achieves higher approximation ratios than QAOA at the same circuit depth on random regular graphs for the MaxCut problem.

As large language models (LLMs) evolve from standalone models into collaborative agents embedded in physical systems, their reasoning and execution are increasingly distributed across wireless edge nodes. In this setting, wireless networks are experiencing a paradigm shift from only providing data connectivity to supporting the multi-agent reasoning workflow itself. The task performance of such network-constrained LLM-based multi-agent systems (MASs) is jointly affected by the multi-agent reasoning dependencies as well as the underlying network connectivity and edge resources. This coupling gives rise to various technical challenges, including the metric misalignment and message redundancy, state inconsistency and topology mismatch, as well as resource limitation and trust discontinuity. To address these challenges, this article develops a novel joint agent–network design perspective that coordinates decisions on both sides of the system. Specifically, we present the joint design of agent–interaction scheduling and resource allocation, the message selection-transmission co-design, as well as the joint agent–network topology design and workload–resource allocation. Furthermore, we consider the network-verified provenance that is linked with agent-side information-flow control to constrain how received information affects subsequent operations. An illustrative vehicle-to-everything (V2X) case study shows that jointly adapting agent-side interaction decisions and network operations improves task completion under communication and edge-resource constraints, outperforming the conventional agent-only and wireless-only separate designs.

Ding \cite{Ding2016} proposed ten conjectures on binary linear codes arising from Boolean functions. Four of them, namely Conjectures 38–41, were subsequently proved by Göloğlu and Krasnayová \cite{GologluKrasnayova2019}. In this paper, we investigate the remaining six conjectures, namely Conjectures 19, 27, 30, 33, 34, and 37. For Conjectures~19 and~27, we obtain common weight restrictions and several infinite five-weight families. For Conjecture~30, we prove that every admissible code has three, four, or five nonzero weights, and an explicit four-weight example disproves the original “three or five weights” assertion. For Conjecture~33, an infinite five-weight family is obtained. For Conjecture~34, we obtain a general $(2h+1)$-weight upper bound and give an explicit six-weight counterexample, showing that the original “three or five weights” assertion is false in general, where $h$ is a positive integer. The case $h=3$ with $3\nmid m$ is also completely determined. Finally, Conjecture~37 is completely resolved by combining the known results of Ahmadi and Shafaeiabr \cite{AhmadiShafaeiabr2023} with the treatment of the two remaining classes $(a)$ and $(b)$.

Coleman–De Luccia tunneling can exhibit an infinite tower of rapidly oscillating negative modes, complicating the semiclassical evaluation of the path integral. We show that this tower does not represent infinitely many bounce-specific instabilities, but instead is tied to a single underlying gravitational conformal sector. This sector is exposed in two different ways: by a large scalar gradient near the bubble wall and by the turning point of a compact Euclidean geometry. The corresponding conformal sector is already present in fluctuations about the false vacuum. Treating the bounce and the false vacuum in the same well-defined gauge and with a common regulator, we find that their high-frequency negative spectra approach one another. The ultraviolet tower therefore cancels from the relative negative-mode count, leaving a single bounce-specific negative mode. This finite-backreaction spectral cancellation provides a corresponding picture of the result obtained in the small-backreaction limit.

This paper gives quantum query lower bounds for two relational graph problems, triangle listing and explicit multiplicative spanner construction, in the general graph query model, where quantum adjacency, degree and neighborhood queries are all available in arbitrary superposition. Both results are obtained by reductions via intermediate multi-block search problems. We extend the quantum query recording framework by Zhandry (CRYPTO 2019) and Hamoudi and Magniez (ToCT 2023) to handle a recording architecture for bidirectional oracles and give a generic blockwise soundness framework, which for arbitrary families of local accepting projectors, gives an exact operator-norm characterization of their maximum overlap with the subspace of bounded weight records. Using the intermediate search problems, we exhibit a family of $n$-vertex graphs with $\Theta(n)$ triangles on which listing any constant fraction of the triangles requires $\Omega(n^{3/2-o(1)})$ quantum queries. This is the first nontrivial quantum lower bound for triangle listing, a question raised by Jiang and Peng (ICML 2026). Further, we show that, for every fixed $k\ge 7$, constructing a multiplicative $k$-spanner requires $\Omega(n^{1+\frac{1}{2\mu_k}})$ quantum queries, where $\mu_k=k/3+O(1)$. For $k\in\{7,8\}$ the bound is $\Omega(n^{5/4})$, which matches what would be implied by an unproven instance of the Erdős girth conjecture, and for large $k$ the exponent $1+\frac{3}{2k}$ exceeds the $1+\frac{4}{3k}$ implied by the provable high girth dense graphs due to Lazebnik, Ustimenko and Woldar, 1995.

Large multimodal models (LMMs) support diverse visual understanding and reasoning tasks but are often impractical to run entirely on resource-constrained devices. Device-edge co-inference reduces device computation, yet transmitting visual data over bandwidth-limited uplinks can introduce substantial delay. Task-oriented feature compression (TOFC) reduces the payload through feature aggregation and entropy coding. However, continuous-feature coding remains costly, and query-agnostic aggregation may discard task-relevant local evidence. We propose query-guided task-oriented feature compression (Q-TOFC) for device-edge multimodal inference. Q-TOFC employs residual vector quantization (RVQ) to encode each merged feature as a compact sequence of codebook indices, reducing its representation cost and allowing more features to be transmitted. It further incorporates query relevance into feature aggregation and uses a quantization error compensation adapter to mitigate the distortion introduced by discrete quantization. Experiments on seven multimodal benchmarks show that Q-TOFC reduces the visual payload by 53.6% relative to TOFC while maintaining comparable average normalized task performance. End-to-end latency evaluations further demonstrate lower latency under bandwidth-constrained uplinks.

Real-world videos provide rich demonstrations of manipulation, but turning them into reusable robot skills requires visually aligned environments, executable physical interactions, and mechanisms for learning from experience. We introduce Real2Gym, an agentic Real2Sim2Real framework that turns human and robot demonstrations into interactive simulation gyms and brings skills acquired in simulation to physical robots. The Real2Sim module reconstructs editable scenes, aligns objects and cameras with the input, validates demonstrated or retargeted actions through native physics execution, and generates task-conditioned variations with action-feasibility checks. Within these environments, the agent generates executable code for manipulation stages, observes their outcomes, and distills successful attempts and failures into reusable task procedures, object-relative motions, and recovery strategies. Through a shared perception-and-control interface, these skills guide subsequent execution in simulation and on real robots, with motions adapted to current observations and no updates to the underlying model weights. Extensive evaluations demonstrate that Real2Gym enables high-fidelity simulation environment reconstruction, outperforming GPT-6 Astra Direct Mode by 16.7% in success rate with approximately 74.9% fewer policy-execution tokens across these environments, while exceeding it by 33.3% in physical robot execution success rate across four tasks on a real Franka robot.

The Galactic microquasar GRS 1915+105 is one of the most interesting X-ray binary systems, exhibiting an exceptionally diverse range of observational phenomena. It displays the most complicated and variable X-ray light curves and carries superluminal radio jets. Its X-ray and radio fluxes change on very different timescales, from seconds to minutes to days and even longer. In recent years, its jet has also displayed large position angle shifts. We find that most of the observational phenomena can be well explained within our newly developed variable turbulent ADAF (TADAF) model. The accretion disk is composed of an outer thin disk and an unstable inner TADAF torus whose size is variable. The previously suggested 'disappearance' and regain of the inner disk is understood as the transition between the thick TADAF and the thin disk. Some types of QPOs are understood in terms of the radius change of the TADAF torus. The jet position angle shift is understood as the direct sideways frame-dragging on the jet by the Lense-Thirring effect from a tilted, spinning black hole. The slower the outshooting jet, the larger the angle shift. We note that such an effect also applies to other sources, such as V404 Cygni and the AGN M87. This mechanism may also be applicable to the famous 'precessing' jet in SS 433.

Near 57 keV, destructive interference makes argon-40 almost transparent to neutrons. The 0.4% of argon-36 and argon-38 in natural argon then supply over 90% of its opacity, amplified 230-fold. Published transmission data rule out the argon-36 evaluation FLUKA uses by default, even with every other isotope made transparent, and bound its cross section in the window to 2-11 b. In a DUNE-scale detector, switching to the data-consistent evaluations moves the fiducial fraction of argon-41 calibration captures by 9-13% while the total yield changes by only 2%: the effect is on where captures occur, not how many.

Multimodal biomimetic underwater robots (BURs) can conduct underwater tasks suitable for the environment. Combining the characteristics of aquatic organisms enables the swimming and leggedlocomotion required for underwater exploration. Locomotion control mechanism relies on rule-based behavior selection and the designer's discretion. This limits the robot's ability to acquire new behavioral capabilities to the predetermined range of behaviors. To address these challenges, we propose a mechanism and control system that enables the expression of multimodal locomotion capabilities from the same multi-jointed structure. A mechanism equipped with four leg-fins each having four axes is used. This controller achieves nonlinear behavior based on sensor modalities, rather than relying on predefined conditional rule-based on locomotion functions. This system was validated through both physical and simulation testing based on multiple sensor data and behavioral patterns. Utilizing a potential function in multimodal locomotion control was verified to enable transitions between two or three behaviors. Implementing the control method as a multimodal controller is expected to enhance its application in underwater exploration. Our project page is at https://tasada038.github.io/multi-jointed-bur/.

Data-intensive services in the Computing Continuum must balance analytics quality, resource usage, and cost across heterogeneous nodes with limited and uneven capacity. This balance becomes especially difficult when resource scaling reaches capacity limits, because changes in demand and cluster pressure must then be absorbed without violating client-defined quality ranges. Existing orchestrators mainly adapt resources, placements, or replicas, while analytics requirements such as coverage, sample, and freshness remain fixed. This article presents ARGOS, the Adaptive Reinforcement Learning-Driven Governance for Orchestrated Services, an end-to-end controller that formulates multidimensional elasticity as a per-request Markov decision process over analytics quality and cluster pressure, supported by capacity-aware admission. ARGOS is evaluated under controlled workloads and time-varying multi-tenant arrivals on a heterogeneous cluster. Across the controlled scenarios, the deep reinforcement learning policies consistently outperform the non-learning baselines and approach the independently tuned best-fixed reference. A separate live evaluation reports improvements over the static midpoint under realistic and saturated arrivals, with no recorded CPU or memory violations but remaining coverage violations. These results support deep reinforcement learning as an adaptive mechanism for multidimensional elasticity when resource scaling alone is insufficient.

The ability to navigate toward sound sources extends a robot's reach beyond its visual field, enabling response to auditory events in unknown environments. To equip robots with this capability, existing methods couple acoustic and visual information through joint audio-visual learning in acoustic simulators. However, acoustic simulation is both low-fidelity and expensive, producing a domain gap that prevents reliable real-world deployment, while the discrete action spaces inherited from grid-based simulators introduce an additional kinematic gap on physical robots. To alleviate these issues, we propose BCNav, a decoupled framework that separates the acoustic module from the learned navigation policy using direction-of-arrival (DOA) estimation: an estimator provides a scalar bearing to the sound source, so the navigation policy only processes depth images and a bearing angle, two inputs whose domain gaps are well characterized. We collect shortest-path demonstrations with calibrated bearing noise injection and train the policy via imitation learning to output continuous velocity commands directly executable on ground robots. We demonstrate the method in simulation and on a physical robot, navigating unknown environments without any acoustic fine-tuning, prior mapping, or real-world audio data collection. Code is available at https://github.com/york1to/bcnav.

We study the problem of recovering the governing ODE of a dynamical system from unstructured, high-dimensional observations such as images. Existing methods for ODE discovery typically assume direct measurements of the variables, or do not provide theoretical guarantees on the learned variables and equations. While Causal Representation Learning (CRL) methods provide guarantees on identifying variables from high-dimensional observations up to component-wise diffeomorphisms, we show that in general these variables cannot be used directly as input to equation discovery methods, which typically assume that the variables will lead to sparse equations. So we introduce SParse Equivalent Equation Discovery AutoEncoder (SPEED-AE), a framework that combines a pretrained CRL method with a component-wise autoencoder that learns transformations of variables that are amenable to sparse ODE discovery. We show that for polynomial ODEs, this additional step allows us to restrict the identifiability of each variable from polynomial to monomial diffeomorphisms. Experiments on Lotka-Volterra, Lorenz, and a two-pendulum system show that SPEED-AE improves on the disentanglement of the CRL methods and that it recovers ODEs that are closest to the ground truth, while achieving state-of-the-art forecasting performance.

Long-horizon information-seeking agents often accumulate noisy or misleading context, causing early mistakes to persist and making recovery increasingly difficult. We introduce an autonomous search harness in which the agent manages its own search process through three states: Rubric, Answer, and Verify. The agent first defines criteria for a valid answer, searches under these criteria, and then independently verifies the result before deciding whether to terminate or continue searching. It is further equipped with a Seal Memory tool that enables active context management. Training this behavior with reinforcement learning, however, can induce Seal Collapse, resulting in unstable training and preventing the agent from reliably learning when and how to use its memory tools. We solve this with a simple strategy that trains only the final segment after context management. Our 35B model achieves 72.83 on BrowseComp, outperforming comparable open-source systems, and consistently improves over the base model across BrowseComp-ZH, xbench, DeepSearchQA, WideSearch, financial investigation, and product search. Ablations show that autonomous compression outperforms automatic compaction and validate our RL design.

We study how actuator rank and control horizon jointly determine worst-case control performance for finite-dimensional linear systems under a fixed total actuator-gain budget. The design variable is the input covariance $X=BB^\top$. Scalar actuators give rank-one matrices $X=bb^\top$; dropping the rank constraint yields their convex hull and a semidefinite benchmark. We identify regimes in which low-rank designs necessarily fall short and others in which they attain the benchmark. At small time, the relaxed maximizer is unique and full rank. If $A$ is cyclic, then, for every $k

Latent Diffusion Models (LDMs) typically adopt a two-stage pipeline: an auto-encoder (AE) is first pre-trained to define a latent space, then a diffusion model is trained to perform denoising within it. Such a two-stage design introduces a representation mismatch, as the latent space is optimized for reconstruction rather than adapting the denoising dynamics. We reveal that the LDM itself is an AE, and consequently present LDM-is-AE, an end-to-end one-stage LDM training framework that eliminates the need for a separately trained tokenizer. Our key observation is that the LDM backbone actually performs a latent-to-feature-to-latent transformation at each denoising step, which can be interpreted as an internal decoding–encoding process. Leveraging this structure, we split the DiT backbone into two reciprocal components, DiT-E (i.e., DiT Encoding) and DiT-D (i.e., DiT Decoding), and impose image-space supervision on the intermediate features across all timesteps. Our model encourages the internal representation to align with the image domain throughout denoising, thereby establishing an explicit latent-to-image-to-latent path. At the zero-noise timestep, our model further performs an image-to-latent-to-image mapping, corresponding to an auto-encoding process. As a result, LDM-is-AE jointly learns latent representations and denoising dynamics in an end-to-end manner, yielding a diffusion-native latent space tailored to the generation process. Experiments demonstrate that LDM-is-AE exhibits highly competitive generation performance, achieving an FID of 1.80 and 1.90 on 256x256 and 512x512 class-conditional image generation, respectively.

Extreme-mass-ratio inspirals (EMRIs) provide a promising probe of the dark matter environment surrounding massive black holes through their long-duration gravitational-wave signals. In this work, we investigate the influence of self-interacting dark matter (SIDM) on EMRI dynamics and gravitational-wave phase evolution. Motivated by the collisional Boltzmann description of SIDM, we construct a leading-order effective correction to the conventional Chandrasekhar dynamical-friction force, controlled by the momentum-transfer cross section $\sigma_T$ and the dimensionless collisionality parameter $n_\chi\sigma_T r=r/\lambda_{\rm mfp}$. We then compute the resulting orbital evolution, waveform modification, and accumulated gravitational-wave dephasing. In the enhanced-drag regime considered here, increasing the SIDM self-interaction strength increases the effective dynamical friction, accelerates the inspiral, and produces a progressively larger phase deviation relative to the collisionless dark matter case. We further distinguish the total dark-matter-induced dephasing relative to vacuum from the additional phase correction generated specifically by dark matter self-interactions. Our results demonstrate that the cumulative gravitational-wave phase of EMRIs can be sensitive to the microscopic momentum-transfer properties of dark matter, suggesting that future space-based detectors such as LISA may provide a complementary probe of SIDM in the vicinity of massive black holes.

Time series modeling increasingly demands high-quality supervision, yet target observations remain scarce - exogenous inputs are broadly available, but target measurements are often unavailable due to cost, infrastructure, or accessibility constraints. Can models trained on observed locations reconstruct target time series where measurements have never been collected? We term this zero-shot time series reconstruction. A naive approach - directly mapping exogenous inputs to targets - can yield predictions at unobserved locations, but without target signals, such models fail to capture the intrinsic dynamics of the target variable, producing overly smooth outputs that underestimate extremes. This reveals systematic errors that call for explicit modeling and calibration. We propose ZeroDiff, which constructs an informed prior from exogenous variables alone, then learns to calibrate reconstruction errors through diffusion - training on observed locations and generalizing to unobserved ones. Experiments across diverse real-world datasets demonstrate significant improvements over existing approaches. Our code is available at https://github.com/YingdaFan/ZeroDiff-ICML2026.

For a controllable single-input linear system in dimension $n$, the controllability Gramian eigenvalues, ordered decreasingly, are known to have the successive small-time orders $T, T^3, \ldots, T^{2n-1}$. We refine this leading-order hierarchy by explicitly computing the next-order coefficient of every eigenvalue and the first-order variation of its eigendirection and of the nested spectral subspaces. The resulting formulas have intrinsic expressions in the orthonormal Krylov basis: eigenvalue corrections describe the action of the dynamics along each Krylov direction, while variations of the spectral subspaces describe coupling to the next direction. We also establish local joint real-analytic dependence on $(T, A, b)$ near $T=0$ for the eigenvalues divided by their leading powers of $T$, consistently oriented eigenvectors, and spectral projectors. This provides convergent expansions with remainders uniform on compact families of controllable pairs. These results yield refined asymptotics for the worst-case minimum energy required to reach a unit target from the origin, the Gramian determinant and condition number, and the Ornstein–Uhlenbeck Gaussian profile in moving principal coordinates. The energy equals the reciprocal of the smallest Gramian eigenvalue, whose eigendirection identifies the most energy-demanding targets for sufficiently small times. We further derive an exact Gramian-weighted energy identity for the forward Fokker–Planck equation and sharp anisotropic short-time smoothing asymptotics using, respectively, the Lyapunov equation and an exact Fourier norm formula combined with the graded Gramian factorization. For symmetric dynamics, the spectral data reconstruct $A$ and determine $b$ up to its global sign.

Large language models trained on vast corpora inherently risk memorizing harmful content that may later re-emerge in their outputs. To mitigate this issue, existing unlearning methods typically rely on training-based parameter updates, such as gradient ascent and its variants, to delete targeted content while preserving other knowledge. However, balancing the competing goals of forgetting and retention makes hyperparameter choices for these methods particularly difficult, often requiring repeated tuning to obtain a strong model that still leaves substantial room for improvement and transfers poorly across models and datasets. To address this challenge, we investigate whether unlearning runs exhibit exploitable structure in weight space, and observe that models from different runs still lie in a shared evaluation-performance basin. This suggests that stronger models may be recovered through an unlearning-tailored soup strategy, reducing the need for repeated tuning for further improvement or new settings. Motivated by this, we propose UnlearningSoup, a unified framework that provides two strategies: EfficientSoup uses binary-search-based interpolation to quickly discover a well-performing model in the early stage, where repeated tuning would otherwise make strong model selection costly. PerformanceSoup uses reweighted souping to efficiently unlock the remaining performance potential in the later stage, where repeated tuning becomes increasingly inefficient. Extensive experiments across diverse datasets and models show that UnlearningSoup delivers 2.4x to 3.3x efficiency gains in hyperparameter selection, while consistently improving performance across settings.

Thin chromium nitride (CrN) films were deposited by RF magnetron sputtering in a reactive atmosphere of Ar-N$_2$ on fused silica and MgO (001) substrates at temperatures between 400 $^o$C and 800 $^o$C. The influence of nitrogen content and substrate temperature on structural, electrical, and optical properties was systematically investigated. The overstoichiometric CrN films exhibited a reduced resistivity below 6 m$\Omega \cdot$cm with stable p-type behavior, while increasing the substrate temperature above 600 $^o$C induced a transition to n-type conductivity. Spectroscopic ellipsometry in the UV-Vis-NIR and infrared spectral ranges revealed a strong dependence of the dielectric function and the absorption edge on film stoichiometry and deposition conditions. Structural characterization using X-ray diffraction and atomic force microscopy confirmed substrate-dependent crystallinity and grain growth. Density functional theory calculations showed that cation and anion vacancies strongly modify the electronic structure and optical response, explaining the experimentally observed conductivity transitions. The results demonstrate that the deposition parameters provide an effective means to tailor CrN thin films via defect engineering.

Neural network conformal theories were first introduced in \cite{Halverson:2024axc} and exact two-, three- and four-point correlators were computed from an ensemble of infinite-width neural networks. In particular, the four-point correlator of any scalar operator satisfied crossing-symmetry and conformal block decomposition across different channels were in agreement without any constraint on moments. An infinite tower of degenerate operators with the same scaling dimension were found. In this note, we take a closer look at the structure of correlators and relation between operators. We also layout a recipe to construct the algebra of the theory.

We analyze exponential backoff, an algorithm used to share a single communication channel in a distributed manner between several users, in a similar way as many networking standards such as 802.11. We demonstrate that this algorithm has a meta-stable behavior, in the sense that the system oscillates over long time-scales between meta-stable configurations where only a subset of the sources effectively access the channel. In other words, meta-stability creates a form of implicit admission control. We provide both theoretical tools and numerical experiments to understand this phenomenon further. Our experiments are fully reproducible and the code is publicly available as a Python package available at https://pypi.org/project/slotted-aloha-simulator/.

Aerial seed broadcasting can reach remote restoration sites, but provides limited control over seed placement within the soil. This paper presents a geometry-assisted, tri-functional morphing robot that combines aerial access, ground locomotion, and controlled-depth seed embedding in a fly-drive-plant architecture. After landing, the platform reconfigures into a four-wheeled planting configuration: an electronically coupled dual-motor drive folds the rear arms outward to form ground wheels, while a descending front tray engages the propulsion motors with a drill gear train. Reusing the propulsion motors for drilling eliminates a dedicated drill drive. The planting sequence forms a hole, dispenses a seed, and allows the vehicle to reposition on the ground or return to its flight configuration. Ground mobility supports repeated planting without requiring a separate flight between adjacent sites. A companion controller issues reconfiguration, tray, and seed-gate commands, while a dedicated autopilot handles flight control. The morphing and planting mechanisms are validated using a hardware prototype that demonstrates ground repositioning and seed embedding. This proof of concept establishes a hardware basis for aerial-ground seed embedding through coordinated reconfiguration and actuator reuse.

Non-rigid point-cloud registration aims to find the corresponding target point for each point on a deforming source surface. Point-level matching keeps the full target cloud available, but correspondence becomes ambiguous when different regions have similar local geometry. Regional or coarse-to-fine methods provide broader spatial context, but an incorrect regional match can exclude the correct correspondence before dense matching. We propose CoCo-Reg, which uses regional patches to enrich dense point features without allowing patch predictions to restrict the final point-level search. CoCo-Reg constructs farthest-point-sampled patches, exchanges geometric information within and between source and target, supervises patch similarity using identity-corrected point overlap, and projects the resulting regional information back to dense point features. The final registration stage still scores the full target cloud before global point-level candidate selection. On 726 held-out ModelNet10 objects across nine deformation levels, two established learning-based baselines obtain mean correspondence errors of 0.1993 and 0.1921, whereas CoCo-Reg obtains 0.0547. Relative to its point-level baseline, this is a 72.6% reduction. CoCo-Reg achieves lower correspondence error on 92.3% of paired test objects and reduces the mean fraction of points with error above 0.1 from 47.3% to 17.3%. Chamfer distance and HD95 decrease in the same direction, and CoCo-Reg remains lower across all tested deformation levels. These results support using regional context for dense non-rigid correspondence without imposing a hard patch-level restriction on the final search. Because evaluation uses one checkpoint per method, the reported gains characterize the complete systems rather than the isolated causal contribution of an individual component. Code will be made publicly available.

Rare but consequential failures can persist in learned locomotion policies for legged robots even when average task performance is high, in part because standard curricula primarily adapt task difficulty rather than the distribution of safety-critical experience. We introduce Predictive Safety Curricula (PSC), a framework for allocating locomotion training experience using learned predictions of future safety cost. PSC trains a distributional safety critic from policy rollouts and uses its predictions to prioritize both terrain contexts and previously encountered randomized events. The resulting curriculum modifies the training distribution while leaving the task reward and policy-optimization loss unchanged. We evaluate PSC in controlled rough-terrain locomotion and in production locomotion systems. PSC improves reliability relative to standard terrain progression, advantage-based replay, and learning-progress curricula, with the largest gains on difficult terrain and under degraded observations. The same allocation principle transfers to two production locomotion stacks. On ANYmal-D hardware, PSC reduces shank-collision incidence by $63\%$ relative to the learning-progress curriculum across three matched training seeds, with a reduction in every seed. On a production stair-climbing platform, PSC eliminates observed shank collisions in the evaluated hardware trials. These results show that learned predictions of future safety cost can provide an effective signal for allocating training experience toward rare failure modes and improving locomotion reliability.

Distributed optimization in edge networks is constrained by heterogeneous client computing capabilities and limited communication resources. We propose the Windowed and Quantized Group-Based Alternating Direction Method of Multipliers (WQ-GADMM) to coordinate group updates under limited activation capacity and reduce communication costs. Clients are grouped by estimated computation time. Each window activates a limited number of groups per round, and the cloud updates the global model after all groups have updated once. The method quantizes both downlink and uplink model exchanges to reduce communication costs and allows bounded model staleness and inexact proximal local updates. For smooth nonconvex objectives, we establish an average squared Karush-Kuhn-Tucker residual bound under the stated assumptions and parameter conditions. The bound consists of a term that decreases with the iteration count and a quantization-dependent error term. Experiments on MNIST and CIFAR-10 show that 12-bit communication reduces communication volume and simulated wall-clock time while maintaining test accuracy comparable to full precision. The 12-bit configuration also maintains complete group coverage and achieves shorter mean group inter-completion gaps than the evaluated baselines.

We investigate numerically fluid flow through a thin porous layer, which is modelled as a lower-dimensional interface separating two bulk domains. At this interface, we impose effective transmission conditions involving homogenized coefficients rigorously derived via two-scale methods in previous works. These coefficients are determined by cell problems encoding the microscale properties of the porous layer. In a first step, we analyze qualitatively how the effective coefficients depend on the pore geometry. Based on these computations, we then present numerical simulations of the macroscopic models incorporating the effective interface conditions. Although replacing the thin porous layer by an effective interface substantially reduces the computational cost, the resulting nonstandard transmission conditions pose additional challenges for the numerical implementation. We employ a Taylor–Hood-type mixed finite element discretization with discontinuous pressure and tangential velocity across the interface and prove stability as well as error estimates for the resulting scheme. Finally, for microscale geometries that are still computationally resolvable, we compare the results of the macroscopic models with those obtained from direct simulations of the microscopic models.

Visually plausible articulated assets may still fail during contact interactions or exhibit inaccurate motion. We present FACT (Fidelity-Aware Construction of Articulated Twins), an agentic framework that progressively constructs articulated twins to improve geometry, contact, and dynamic fidelity. The agent drives an evidence–diagnosis–revision loop on a shared editable representation, selecting measurements and model edits using quantitative feedback, while numerical tools execute and validate the updates. It reconstructs editable articulated geometry from images through feature planning, targeted measurements, and diagnostic refinement. On this reference, it repairs collision proxies through task-aware local repartitioning before fidelity-constrained compression. Finally, it constructs response models from passive-response videos, using simulation residuals to guide model revision and constrained physical parameter fitting. Experiments show that FACT improves geometric reconstruction over baselines, enables more reliable interaction with simpler collision proxies, and better reproduces held-out physical responses than direct parameter inference.