Papers

Poisson sprinkling provides a Lorentz-invariant way of generating causal sets from a continuum spacetime, but numerical work necessarily restricts the sprinkling to a finite region. For observables involving causal-set links, this finite-volume restriction can generate a bias that does not disappear by simply increasing the number of sprinkled points. The origin of the problem is geometric. At fixed proper time from an event in Minkowski spacetime, candidate links populate a non-compact hyperbolic space of rapidities. In dimensions $d+1$ with $d\geq 2$, this space is boundary dominated: the fraction of volume close to the boundary of a large region does not vanish in the infinite-volume limit. Different finite-volume regularizations can therefore produce different limiting normalized statistics. We relate this phenomenon to the failure of naive open-boundary thermodynamic limits on non-amenable spaces, derive explicitly the dependence of the link proper-time distribution on the finite cutoff, and distinguish a well-defined link-intensity profile from the ill-defined notion of a uniformly chosen link. We verify these results with independent Poisson-sprinkling simulations in $d+1$ dimensions with $d=1,2,3$. We then propose an event-centered product-window estimator with a controlled double-scaling limit, and discuss compact hyperbolic quotients as a possible boundary-free alternative.

We construct an explicit quartic polynomial for which Newton's and Halley's root-finding methods both fail to converge to a root on nonempty open sets of initial values. The example can be found through a reproducible computational search within a one-parameter family of quartics designed to exhibit a prescribed Newton two-cycle. We then prove analytically that Newton's method has a superattracting two-cycle and that Halley's method has a distinct attracting extraneous two-cycle, using a two-interval contraction argument. This fills a degree-four gap in the simultaneous Newton–Halley failure problem. The computational approach also identifies another quartic with numerical evidence of simultaneous failure of Newton's, Halley's, and Schröder's methods.

Ensuring fault tolerance in the User Plane Function is essential to maintain service continuity in 5G networks, particularly as latency-sensitive applications become more prevalent. This paper presents a novel mechanism for dynamic UPF failure detection and recovery within an OpenAirInterface-based 5G core environment. The approach introduces an application-layer session restoration mechanism to ensure service continuity and minimize disruptions during UPF failures. Experimental results demonstrate that the proposed mechanism reduces UPF downtime while maintaining low packet loss and enabling rapid service recovery. In tests with TCP and UDP traffic, connectivity was restored within a few seconds, minimizing service degradation. Furthermore, our failure-aware UPF selection mechanism improves long-term system resilience by considering recent failure history. The results highlight the effectiveness of our lightweight framework in improving the overall reliability of 5G networks, making it a valuable contribution to network robustness.

The symmetry of a physical system given by the 122 magnetic point groups (MPGs) leads to only 21 sets of restrictions on any physical tensors. Those, so called, tenor forms were labeled by letters from A to U by R. R. Birss (Symmetry and Magnetism, North Holland, Amsterdam, 1964). A new concept of property-related point groups (PRPGs) is proposed. It is shown that the 21 PRPGs isomorphic to the 21 non-centrosymmetric crystallographic point groups (CPGs) corresponds to the 21 tensor forms. Therefore instead of a symbols from A to U one can use PRPGs as mathematical objects corresponding to the tenor forms. From the Neumann's principle given multipole component is only allowed in a system when a group describing symmetry of the system is a subgroup of the multipole component symmetry. Each tensor can be expressed in the terms of the multipole degrees of freedom, i.e. it corresponds to a subset of all multipole components allowed by given PRPG. Those relations leads to the algebra of tensor forms, where space of allowed tensor components and space of PRPG operators are somehow dual to each other.

We find traces of fluid dynamics in the self-evolution of an LLM agent, and give the kinetic theory that predicts them. Gated self-evolution is the loop in which an agent rewrites its own skills under a validation gate. Self-evolution research has treated the agent as the unit; we study instead the individual instances inside it. Here the agent is DSH-plugin-based: it runs in production on DeepSeek Harness (DSH), and its plugins satisfy four architectural properties (permutation symmetry, reversibility, acyclicity, typed contracts), which license treating these instances as identical hard spheres; the theory is accordingly scoped to DSH-class plugin populations. On this scope the paper builds three theory layers. The rigorous layer, independent of any analogy, comprises an any-time hitting-time certificate bounding the expected rounds to any prescribed improvement, a resolution law that prices held-out validation budgets, and a separation theorem: the daemon must stay outside the population, because merging evaluator with evaluated voids the certificate. The kinetic layer is a master equation over the plugin x version x task grid with four operators (collision, reaction, external field, gate), where collision is co-activation. Its moment hierarchy, the step that turns a gas into fluid equations, generates the falsifiable statistical signatures. Throughout, the fluid reading is a bounded analogy: momentum is not conserved, so no Navier-Stokes limit exists. The measured layer runs on a faithful minimal instance, a large library of four-parameter skill plugins retrieved one per episode with a co-activation probe, in a one-model, one-task-family WebShop environment; every element maps to the DSH loop by architectural role. Population fluctuation scaling is density-gated: invisible at sparse edit density, it emerges at the predicted rate under tripled density, as directional evidence.

We develop a Newton–Kleinman iteration for linear-quadratic optimal control of infinite-dimensional systems with possibly unbounded control and observation operators. In this setting, the classical algebraic Riccati equation need not be meaningful and therefore cannot serve as the basis of the iteration. We instead linearize a Lur'e identity formulated directly on the system-node domain. Starting from any stabilizing feedback, the resulting iteration is well defined, preserves stability, and yields a monotonically decreasing sequence of cost operators. A resolvent-based normalization makes the construction applicable to unbounded control. Under finite-cost detectability and for a finite-dimensional input space, the iterates converge strongly to the optimal value operator, independently of the fixed normalization parameter. In finite dimensions, the natural zero-state normalization recovers exactly the classical Newton–Kleinman iteration. The formulation leads to a low-rank numerical realization based on shifted system solves and inversions on the input space. We illustrate the method on a linearized cylinder wake with boundary traction control and unbounded observation, using one Newton–Kleinman update per adaptively refined spatial mesh, and on a damped linear elasticity problem with boundary traction control and boundary displacement observation.

Stimulated parametric down-conversion (StimPDC) is the seeded version of the process that underlies much of photonic quantum information. Replacing the vacuum at one input port with a coherent seed leaves the interaction Hamiltonian, the phase-matching function and the two-photon amplitude unchanged, while raising the output to levels accessible with ordinary classical detectors. This review treats the transverse spatial physics of that process. We show that a single relation, in which the idler amplitude is the pointwise product of the pump field with the complex conjugate of the seed, organizes results usually presented separately: the image and coherence transfer experiments of the 1990s; phase conjugation and its use for wavefront correction; the conservation and conjugation of orbital angular momentum, together with the generation of radial modes and geometric intensity patterns; and the processing of vector vortex beams carrying spatially varying polarization. We then consider StimPDC as an instrument rather than as an object of study, covering stimulated emission tomography, the experimental realization of Klyshko's advanced-wave picture, and the transfer of partial coherence, including the conservation of the twist phase of twisted Gaussian Schell-model beams. We close with the quantum regime, reached when the seed approaches the single-photon level or the idler is used as a herald, and with an assessment of open problems and prospective applications.

Large language models (LLMs) are increasingly deployed in public health, finance, and governance, requiring both accuracy and societal value alignment. Despite recent advances, LLMs often perpetuate or amplify bias embedded in their training data, posing challenges to fairness. While self-debiasing encourages an LLM to identify and correct its own biases, relying on a single model's intrinsic knowledge may be insufficient to address deeply ingrained stereotypes. To address this limitation, we introduce Collective Bias Mitigation (CBM), a framework that alleviates bias by learning fine-grained model behavior and fostering knowledge sharing among diverse LLMs. This work is the first to systematically explore the effective selection and organization of distinct LLMs to cultivate fairer LLM responses. Experiments show CBM substantially outperforms standalone baselines (e.g., in the top-7 setting, Committee lowers the age bias score from 0.25 to 0.10). Our Debating and Committee topologies achieve substantial bias reduction, with the latter balancing mitigation effectiveness and inference cost, highlighting the potential of CBM for fairer LLMs.

The cell cortex and cell adhesion contribute to the mechanics of epithelial cells and hence tissues. Vertex models of epithelia represent the cell cortex and cell adhesion, but these models mostly ignore the other subcellular structures that compose the cytoplasm surrounded by the cortex. Here, we reveal the mechanical consequences of these structures in a minimal vertex model of an epithelial monolayer, in which the single-cell energy includes a cytoplasmic contribution that represents, e.g., the cost of confinement of cytosolic polymers. Strikingly, our simulations show that the buckling transition of the monolayer under compression becomes discontinuous in an intermediate range of values of the associated cytoplasmic modulus. By asymptotic solution of this buckling problem in the continuum limit of the model, we explain this reentrant behaviour in terms of the tricritical points of a sixth-order Landau theory. Our results thus underline how cell-scale mechanics beyond the cortex and cell adhesion can change emergent, tissue-scale behaviour even qualitatively.

Gosper introduced many remarkable formulas involving infinite products. One such formula provides an evaluation of the product $$\prod_{n \geq 1} \frac{1}{e}\left(\frac{1}{3n} + 1\right)^{3n + \frac{1}{2}}. $$ It appears that a complete proof of this evaluation has not appeared in the literature. In this paper, we introduce some techniques to extend Gosper's evaluation. We prove a generalization of Gosper's formula involving three free parameters, which can also be applied to evaluate arbitrary multisections of Gosper's product. Then we investigate higher-order extensions involving higher-degree polynomials in $n$, relative to the polynomials $3n$ and $3n+\frac{1}{2}$ involved in Gosper's product, yielding a generalization of an infinite product evaluation due to Ramanujan and recently studied by Bradley–Thrush [\it Ramanujan J. (2025)]. We also consider connections with infinite products due to Kurokawa and Rovinskiĭ, building on the recent work of the first and fourth authors [\it Ramanujan J. (2025)].

A natural-language requirement leaves questions open, and a model asked to implement it settles them silently: across 600 generated test suites from three models, 42.7% contain no test that distinguishes the competing readings. An acceptance example written before implementing is the usual remedy, but an example both readings satisfy resolves nothing. We insert two steps into that practice: enumerate the requirement's underspecified points by name, then for each construct two throwaway implementations differing only in that point and keep a candidate input only if executing both shows they disagree. The result is recorded as a decision pin: the named point, the confirmed input, and the value the person chose between the two exhibited results. The same record then constrains generation and decides compliance by execution. On a benchmark of 40 tasks with paired reference implementations and two models, a separating input is obtained for 92.5% of decision points and a pin identifying the intended decision for 85-90%. As checks on 703 independently generated implementations, pins agree with the benchmark's classification on 96-97%, with disagreements on three tasks, one where both classifiers erred. As generation constraints, pins are honoured at the pinned input in all 210 generations and are never worse than a prose rule on held-out inputs in 38 cells, though no better than prose stating the same scope. Every compliance failure under a prose rule came from the model deciding the rule's scope itself; one such case silently overturned another recorded decision, was attributed to a single rule by leave-one-out on both models, and was missed by text-level reconciliation but caught by re-running the recorded input. The setting yields too few such conflicts to evaluate a regression step, and we say why.

We study the well-posedness of small BV solutions to the one-dimensional isentropic Euler system on the half-line under inflow and outflow boundary conditions. These boundary conditions are formulated in terms of admissible trace sets determined by Navier–Stokes boundary layers and zero-speed shocks. For both the inflow and outflow problems, we construct small BV solutions taking values in the subsonic region. These solutions are unique and satisfy the quantitative stability estimate \begin{align*} \|U(\cdot, t)-V(\cdot, t)\|_{L^2} \lesssim \sqrt{\|U(\cdot, 0) - V(\cdot, 0)\|_{L^2}} \, , \end{align*} which holds for any such small BV solution $V$ and any $L^\infty$ entropy solution $U$ satisfying the strong trace property and the inflow/outflow boundary conditions. In particular, we emphasize that $U$ may take values outside the subsonic region and need not have bounded variation or be constructed by the front tracking scheme. This establishes the first weak-BV stability and uniqueness theory for inflow and outflow initial-boundary value problems. A key difficulty lies in verifying that the boundary set is compatible with the $a$-contraction framework, and in proving that the front-tracking limit satisfies the prescribed inflow/outflow conditions.

Collective noise can leave degrees of freedom unaffected, allowing them to store quantum information. Processing this information also requires physically available logical gates. We consider $r$ fundamental qutrits transforming under $U\in SU(3)$ and $s$ dual qutrits transforming under its complex conjugate $\overline U$, with collective action $R_{r,s}(U)=U^{\otimes r}\otimes\overline U^{\otimes s}$. Using the image of the walled Brauer algebra and Lie-algebra arguments, we show that adjacent exchanges within each qutrit type and one singlet-projector interaction between the two types generate any independently chosen family of special unitary gates on the noiseless subsystems by a finite pulse sequence, for every $r,s\geq1$. This establishes semi-universality of this system. We illustrate the result with a numerically optimised sequence of 24 pulses for three fundamental qutrits and one dual qutrit. We then use the encoding and available logical corrections for qutrit teleportation under correlated Weyl noise. Under explicit assumptions on channel composition and which trials are retained, and allowing for baseline implementation errors, we identify noise ranges in which encoded teleportation exceeds the classical fidelity bound of $1/2$, while the unencoded teleportation channel is entanglement breaking. We analyse fidelity on accepted trials separately from the transmission costs of the additional physical qutrits.

We present WAMpy, a Python framework optimized for synthesizing Prolog programs. Unlike general-purpose Prolog systems, WAMpy targets workloads that repeatedly generate and evaluate small candidate programs. WAMpy compiles Prolog clauses into NumPy array-based WAM instructions and supports partial recompilation of hypotheses against fixed background knowledge. Performance-critical routines are accelerated using Numba just-in-time (JIT) compilation. In a benchmark of repeated compilation-and-evaluation workloads, WAMpy improves end-to-end performance compared with SWI-Prolog accessed from Python using Janus.

Quantum covert target sensing aims to detect low-reflectivity targets embedded in the thermal environment while avoiding discovery by an adversary. This covertness requirement constrains the admissible probe energy, thereby imposing a trade-off between covertness and sensing performance. However, how this resource constraint determines the ultimate estimation precision of the target parameter remains largely unexplored. To fill this gap, we establish a unified framework connecting covert target discrimination with quantum-enhanced parameter estimation. We first present perturbative $\epsilon$-covertness-induced bounds on the probe energy, which yield an admissible energy window narrowing as $\mathcal{O}(\sqrt{\epsilon/M})$. Moreover, we introduce the covertness-constrained quantum Fisher information (CCQFI) to quantify the ultimate precision of target-reflectivity estimation and assess its attainability with specific measurements. For Gaussian-distributed coherent-state (GCS) and two-mode squeezed vacuum (TMSV) probes, their CCQFIs scale as $\mathcal{O}(\epsilon)$ and $\mathcal{O}(M)+\mathcal{O}(\sqrt{\epsilon M})+\mathcal{O}(\epsilon)$, respectively. In particular, in the strong covertness regime $\epsilon\to0$, the GCS CCQFI vanishes, whereas the TMSV retains a finite $\mathcal{O}(M)$ contribution exhibiting a pronounced metrological advantage. Besides, photon counting saturates the GCS CCQFI, whereas homodyne detection achieves a higher attainability for the TMSV probe. Our work identifies covertness as a quantitative resource constraint governing the metrological limits of quantum target sensing.

Explainable Artificial Intelligence (AI) enables humans to understand and interpret decisions of AI models. Instead of having a black box, explainability supports humans in understanding AI models' behavior. Existing explainable AI approaches often present explanations on 2D displays using pre-recorded data, requiring users to relate the displayed information back to the physical objects and real world locations involved in a model's decision. Users are forced to decouple data exploration and capture from AI model interpretation. For AI systems that work within physical environments, this separation can make explanations difficult to interpret in context. We propose using Augmented Reality (AR) to enhance the understanding of AI models by enabling spatial explainability information directly in a user's workspace, in real time, as they explore the world. We show how known explainability methods can be applied in AR and provide insights into user experiences with such an application.

We establish sharp bounds for rational points near compact smooth nondegenerate curves in $\mathbb R^n$, for every $n\geq4$. For a nondegenerate curve $\mathcal C$ in $\mathbb R^n$, let \[ \mathcal R_{\mathcal C}(\delta,Q) :=\#\left\{(\mathbf p,q)\in\mathbb Z^n\times\mathbb N: 1\leq q\leq Q,\ \operatorname{dist}(\mathbf p/q,\mathcal C)\leq\delta/q\right\}.\] We prove that \[\mathcal R_{\mathcal C}(\delta,Q) \lesssim_{\varepsilon,\mathcal C} \delta^{n-1}Q^2+Q^\varepsilon \left(Q+\sum_{k=1}^{n-1} \delta^{k^2/(k+1)}Q^{(2k+1)/(k+1)}\right).\] A key new ingredient is a wave packet method developed by Gan–Maldague–Oh.

Charge deposition in the particle-in-cell (PIC) method can be viewed as kernel density estimation on a spatial mesh. This allows the smoothing bandwidth to be selected through statistical criteria such as mean integrated squared error or cross-validation, but it also changes the electric field acting on the particles. Therefore, a selected bandwidth can alter the growth or damping of physical modes. We develop a dynamics-aware bandwidth-selection procedure that combines a density-error objective with constraints on specified linear modes. These constraints follow from the plasma equilibrium and the combined transfer of deposition, filtering and field gathering within the model rate-tolerance boundary. We demonstrate the need for equilibrium information through a controlled two-stream experiment. The controlled two-stream scan varies the beam drift while holding the initial spatial density, thermal samples, wavevector and fixed-filter transfer unchanged. Across four resolved equilibria approaching marginal stability, this filter reduces the measured growth rate by $0.7\%$ to $9.1\%$. Equilibrium-dependent interior settings satisfy the prescribed $10\%$ rate tolerance. We then apply the constrained selection rule to 85 Gaussian candidates, that reduces density and grid-field mean-square errors by $7.40\%$ and $3.70\%$ on independently evaluated snapshots. A separate projected-linear calculation at that exact width gives a ratio of mean individual growth rates of $0.9556$, with paired $95\%$ interval $[0.9453,0.9640]$, consistent with the prediction $0.9513$. Analysis of finite-marker loading and ensemble averaging explains how the measured rates relate to the response model. Together, these results demonstrate a filter-design procedure that combines statistical error reduction with subsequent validation of the selected linear dynamics.

In 2012, Mirtabatabaei and Bullo studied heterogeneous Hegselmann–Krause models of opinion dynamics, where different agents may have different confidence or influence bounds. They conjectured that opinion vectors always converge, a fundamental problem that remains open. In support of this main conjecture they gave some partial results and made some auxiliary conjectures. Here we give counterexamples to two of these auxiliary conjectures, and even to one theorem in their paper. We explain why these disproved conjectures and false theorem nevertheless likely remain valid for almost all initial opinion vectors. In the last section (Section 7), we describe the role of AI in the production of the results in this paper.

Batching can change a quantized language model's logits even when token positions and causal masks are unchanged. We present a method for executing known queries in parallel while retaining the numerical behavior of a scalar deployment fixed before optimization. The method derives physical KV-prefix and backend arithmetic requirements, groups compatible queries, and keeps independent projection work batched. On CUDA, it packs independent four-column reductions into a broadcast batch without widening the reference-sensitive reduction. On Vulkan, it preserves projection activation arithmetic and extends a scalar attention path across query groups. Controlled interventions and frozen-input studies cover Llama, Qwen and SmolLM2. CUDA packing reduces complete resident-service time by 4.00% on four fresh inputs. Vulkan batched attention reduces ordinary-request time by 32.34% versus per-query attention and 52.87% versus scalar execution on twelve fresh inputs. Equally optimized whole-graph boundary cuts match local grouping on aligned natural prefill and outperform it at some wider Vulkan query sizes. At deliberately crossing CUDA appends, local grouping gives a $1.307\times$ speedup over whole-graph cuts; this conditional result is not a workload-average speedup. These results identify how numerical compatibility constrains parallel execution and how width and alignment determine the benefit of local cuts.

We study two variants of location-domination in infinite triangular grids which allow locating up to $\ell$ vertices simultaneously. A locating-dominating code $S$ is a dominating vertex set such that every vertex outside of it has a unique neighborhood within the code $S$. As this neighborhood is unique, we may use it as an identifier for locating the considered vertex outside of the vertex set. With the variants we study, we may also locate sets of vertices instead of only single vertices. In particular, we obtain the exact minimum densities of these types of sets in the infinite triangular grid for every value of $\ell$.

GPU kernels generated by large language model (LLM) agents can remain less efficient than expert implementations, but runtime alone does not reveal how the gap relates to design discovery and implementation. We introduce D2K-Bench, a diagnostic benchmark of 26 tasks and 85 workloads that measures how effectively agents translate expert design guidance into efficient GPU kernels. The guidance covers L1: high-level algorithmic insights, L2: dataflow design, and L3: low-level optimization tricks, including dependencies among these levels. Pairwise runs with and without guidance share task descriptions, workloads, tools, hardware, and a 350-turn budget. Complementary assessments examine independently proposed designs and the design properties implemented in generated code. Across five models on NVIDIA B200 GPUs, guidance raises correctness over 130 model-task pairs from 93.1% to 98.5% and increases the Performance Score over all 26 tasks from 1.46 to 1.95. For the three frontier models with correct submissions on all 26 tasks in both runs (GPT-6-Astra, Claude-Opus-4.8, and GPT-5.6-Sol), geometric mean speedup increases from $1.69\times$ to $2.49\times$. Across all five models, the mean combined implementation score increases from 57 to 70 out of 100. These results show the value of expert design guidance while identifying design properties that remain unimplemented.

We present the Neuro-Physical Inverter (NPI), a modular, uncertainty-aware framework for geophysical inversion that couples ensemble-based conditioning with constrained residual learning, demonstrated in the 1D magnetotelluric (MT) setting as a controlled testbed. The framework operates in two stages. An Ensemble-Conditional Gaussian Process (EnsCGP) conditions a prior ensemble of resistivity models on the observed response, producing a physically admissible reference ensemble. A residual-learning neural network then predicts targeted corrections to this reference, trained on synthetic data and fine-tuned per station for field application through a physics-coupled objective. Because an ensemble is conditioned, refined, and propagated through both stages, every estimate carries an associated ensemble spread. Synthetic experiments show that NPI systematically reduces ensemble-mean error without destabilizing the ensemble. Applied to broadband MT data from the Gabbs Valley geothermal region (Nevada, USA), NPI reduces the across-station mean misfit over the mid-period band while retaining comparable ensemble spread. The propagated ensemble yields a factor of uncertainty that serves as an operational measure of constraint within the assumed model class. Both stages are dimension-agnostic in formulation, and the design principles established here are intended to scale to higher-dimensional parameterizations.

Diffusion models can synthesise contrast-enhanced CT (CECT) from non-contrast CT (NCCT), avoiding contrast administration and its environmental and patient-access costs. However, visually realistic images are not necessarily anatomically correct, and the pixel-intensity and feature-space similarity metrics used to assess generation quality do not directly measure anatomical correctness. In this work, we investigate whether uncertainty can serve as a proxy for semantic correctness in diffusion-based medical image synthesis. We study NCCT-to-CECT synthesis using AortaDiff, a multitask diffusion framework that jointly generates CECT images and lumen segmentations. The segmentation output provides an explicit representation of the generated vascular anatomy, enabling segmentation-derived errors to be used as a quantitative measure of generation correctness. Six methods spanning weight (Ensemble, HyperDiff, BayesDiff), architecture-perturbation (MCDropout), generative-stochasticity (RDS) and input-perturbation (TTA) uncertainty are compared at the pixel, region and image levels, and for detection of clinically relevant out-of-distribution (OOD) cases. Uncertainty proves informative at all three spatial scales, remains informative on an external multi-centre dataset under distribution shift, and supports OOD detection. MCDropout stands out among the six: it ranks among the leading methods at every scale, generalizes well on the external dataset, and can be enabled at inference on any model already trained with dropout, so reliable uncertainty comes at no extra training cost. Uncertainty reliably flags severe failures but discriminates poorly among already high-quality images. These findings support uncertainty as a practical and computationally economical signal for quality filtering, reliability assessment and OOD detection in NCCT-to CECT synthesis.

Long-horizon LLM agents are typically trained with sparse outcome rewards, making trajectory-level objectives too coarse to distinguish the contribution of individual decisions. Step-level credit assignment provides finer-grained supervision, but its estimates can be unreliable because observed returns also depend on subsequent actions, environment transitions, and trajectory length. We propose AdaStep, an Adaptive Step-credit weighting method that controls how strongly each group-derived local advantage modifies the trajectory-level signal. We formulate this weighting as a mean-squared-error estimation problem for the latent step advantage and, under an explicit conditional sampling assumption, derive an optimal per-state shrinkage coefficient. The coefficient admits a signal-to-total-variance interpretation: it preserves local credit when return variation is attributable to the selected action and suppresses it when variation is dominated by downstream randomness. AdaStep requires only lightweight scalar computation, with no critic, additional rollouts, or extra model inference. Experiments with three model backbones on ALFWorld, WebShop, and ScienceWorld show consistent improvements over baselines at low computational cost.

This paper studies the complexity of convex optimization using lazy second-order oracles (Doikov, Chayti, and Jaggi, ICML 2023), where an algorithm queries gradients every iteration and Hessians once per $m$ iterations. Under this setting, we show a lower bound of $\Omega(m+ m^{1/7} \epsilon^{-2/7})$ on the number of total iterations to find an $\epsilon$-solution using a novel block zero-chain construction. Then we propose a novel method that achieves a new upper bound of $\tilde{\mathcal{O}}(m+ m^{1/7} \epsilon^{-2/7})$, which significantly improves the prior one (Chen, Liu, Luo, and Zhang, COLT 2026) of $\tilde{\mathcal{O}}(m+ m^{13/21} \epsilon^{-2/7})$ and is tight up to logarithmic factors.

Video creation spans text-to-video (T2V), image-to-video (I2V), and condition-based generation, yet video diffusion models remain costly because they repeatedly evaluate large backbones during sampling. Distribution matching distillation (DMD) reduces this cost, but its reverse Kullback–Leibler (KL) objective can provide unstable or incomplete guidance when the student and teacher distributions have limited overlap. VDOT addressed this issue by adding optimal transport distillation (OTD), whose explicit coupling supplies geometric directions for condition-based generation. Balanced OTD, however, performs full-mass matching between the spatial tokens of each corresponding student–teacher frame pair. This assumption weakens for T2V and I2V, where one condition admits many valid outputs and spatial content need not align across different realizations. We present VDOT++, a unified distillation framework that applies the same training recipe separately to generators for the three task families. It makes OTD robust to output diversity through an asymmetric unbalanced formulation that allows unreliable student tokens to carry less mass while maintaining coverage of the teacher tokens. An $\ell_1$ ground cost further replaces mean-based aggregation with a more mode-preserving weighted median that limits the influence of distant transport targets. The two changes respectively determine whom to match and how the selected targets should be aggregated. We additionally combine distribution matching and adversarial refinement through sequential backward passes, and exploit the decoupled score networks for cross-scale distillation, where larger score networks improve a compact generator. Experiments on UVCBench, VBench, VBench-I2V, and the VACE benchmark show that the resulting four-step generators are competitive with many-step teachers and strong few-step baselines across all three task families.

Hybrid quantum classical neural networks integrate parameterized quantum circuits (PQCs) with established deep learning architectures, but their performance depends strongly on the choice of quantum circuit architecture, a choice that remains largely manual. Most existing approaches rely on hand-designed or fixed circuit ansätze, requiring circuit structure, gate composition, and qubit connectivity to be specified in advance with no guarantee that they suit the task. This limitation is especially acute in image classification, where quantum circuits must transform features extracted by classical networks while remaining compact enough for practical training, requirements that generic, task-agnostic ansätze are unlikely to satisfy simultaneously. We extend EXAQC, an evolutionary framework for automated quantum circuit discovery, to image classification. EXAQC evolves PQCs as intermediate processing modules while retaining classical feature-extraction and prediction layers. On MNIST, Fashion-MNIST, and CIFAR-10, EXAQC achieves 98.42%, 90.62%, and 85.47% accuracy, respectively, while using comparable gate counts to other quantum architecture-search methods. Against classical networks, evolved hybrid models maintain comparable accuracy with substantially fewer trainable parameters, reaching 85.68% on CIFAR-10 with over 25$\times$ fewer parameters than a 10-layer CNN. Encoding choice also matters: rotation-based encodings (RX, RY, U3) outperform amplitude encoding by 22-25 points on CIFAR-10. These results demonstrate that automated circuit discovery yields compact quantum modules that can replace larger classical components in vision architectures while retaining competitive accuracy.

Repository-level RTL repair must localize a failure that spans files, modules, and clock cycles, then propagate the fix consistently. Existing methods reason over source code, which reveals possible behaviors but not the failed execution, and cannot tell whether a local fix was propagated consistently. We therefore present an evidence-guided framework with three modules. Failure grounding converts a problem statement into a reproduced failing run and a failure anchor. Waveform-guided localization uses a localization toolbox to narrow the observed violation into a candidate mechanism and an evidence trail. Consistency-aware repair and validation then expand that seed into a coordinated patch and replay the same scenario to check that the violation disappears. We conducted experiments on HWE-Bench and achieved better performance than the baseline.

Microscaling (MX) formats are emerging as a hardware-supported approach to efficient training and inference. They combine low-precision elements with shared block scales, but their impact on vision models remains underexplored. We systematically investigate post-training MX quantization across vision models and tasks. An analysis of direct conversion identifies three sources of error: block-scale representation, the poor alignment of some small convolutional weight tensors with nonuniform element grids, and the underuse of signed codes by nonnegative activations. These findings motivate VisionMX, a post-training MX quantization method that optimizes bounded weight rounding and applies a foldable affine correction to activations. We evaluate VisionMX across image classification, object detection, semantic segmentation, and low-light image enhancement using several MX-style formats. It improves on direct conversion and the evaluated post-training quantization baselines, with the largest performance recoveries in architectures most sensitive to MX conversion