arXiv:2607.13291v2 Announce Type: replace-cross
Abstract: Quantum networking testbeds lack a distinct plane for coordinating distributed measurements and collecting experimental data across heterogeneous devices. To address this gap, we present the Measurement Plane, a dedicated plane that complements the data, control, and management planes rather than replacing or extending their pipelines. The contribution is presented as a distributed framework that organizes measurement functions into four layers: application, experiment coordination, capability, and resource agents. Our design separates user workflows from device-specific control.
We implemented the framework as containerized microservices connected through publish--subscribe messaging, and validated it on a two-node quantum networking setup connected by an optical network. The framework successfully coordinated remote nodes to execute coincidence measurement and polarization entanglement distribution experiments with visibility interference of up to 98 percent. This evaluation demonstrated the effectiveness of the framework for supporting complex, distributed quantum experiments, enabling online measurement and feedback, and significantly reducing manual configuration and execution effort.
Science Journals
arXiv:2607.14087v2 Announce Type: replace-cross
Abstract: We prove a stochastic comparison for Gaussian maxima. Let $R$ be an $m\times m$ correlation matrix satisfying $R-\mathbf{1} \mathbf{1}^{\mathsf T}/m\succeq0$, let $X\sim\mathcal{N}(0,R)$, and let $Z_1,\ldots,Z_m$ be independent standard Gaussian random variables. Then $\max_{1\leq i\leq m}X_i \leq_{\mathrm{st}} \max_{1\leq i\leq m}Z_i$, or equivalently, $\mathbb{P}\{X_i\leq c\text{ for every }i\}\geq\Phi(c)^m$ for every $c\in\mathbb{R}$. This comparison resolves the Weak Simplex Conjecture: among $d+1$ equiprobable equal-energy signals in $\mathbb{R}^d$ transmitted over an additive white Gaussian noise channel, the regular simplex maximizes the probability of correct maximum-likelihood decoding at every signal-to-noise ratio. It also proves the inequality asserted by the Simplex Mean Width Conjecture and gives an exact formula for the largest number of equiprobable messages that can be sent at prescribed energy and error probability by a deterministic no-feedback AWGN code under a per-codeword energy constraint. The proof combines a Gaussian product inequality for log-concave functions with an adaptive tilting argument that makes the inequality applicable to the one-sided threshold events defining the maximum. A lean formalization of this argument is available at https://github.com/abhmul/weak-simplex-conjecture-lean.
arXiv:2607.14089v2 Announce Type: replace-cross
Abstract: Strained membrane resonators have emerged as a promising platform for optomechanical accelerometry; however, the desired combination of low frequency and high $Q$-mass product requires a rethinking of their dissipation dilution engineering. Applying Bayesian optimization to a Si$_3$N$_4$ membrane, we discover a class of sail-like trampoline resonators in which the frequency is decreased by an order of magnitude while preserving the $Q$-mass product. We demonstrate centimeter-scale sails with kHz frequencies, $Q\sim10^7$ and $Q\times\text{mass}\sim$ 10 g. Vertically integrating a 7 kHz device with a nanoribbon, we realize a monolithic cavity optomechanical accelerometer with a room temperature thermal noise of $40\;\text{n}g_0/\sqrt{\text{Hz}}$, sufficient to resolve $\mu g_0/\sqrt{\text{Hz}}$ ambient vibration over a bandwidth of 4 kHz with a displacement imprecision of $10^{-14}\;\text{m}/\sqrt{\text{Hz}}$. Cryogenic arrays of sail membranes may be attractive for new physics searches and distributed quantum sensing experiments.
arXiv:2607.14637v2 Announce Type: replace-cross
Abstract: For close binaries and star-planet systems, tidal interactions mediate the energy transfer between the orbital motion and the internal flows of the bodies involved, thus playing a central role in their evolution. For equilibrium tides, the associated energy transfer is commonly modeled through an effective viscosity acting on the tidal flow. However, the scaling of viscous dissipation efficiency with tidal frequency $\omega_\text{T}$ remains debated, particularly when $\omega_\text{T}$ greatly exceeds the convective eddy turnover frequency $\omega_\text{c}$. Previous numerical studies have addressed this issue by subjecting a turbulent convective flow to an oscillating background shear mimicking equilibrium tides. In this work, we adopt a novel three-layered convective box -- designed to represent a stellar convection zone sandwiched between two stable layers -- driven by an external periodic forcing. We quantify tidal dissipation efficiency by the forcing power on the flow in steady state. Our results yield a shallower scaling of tidal power per unit mass with $\omega_\text{T}$ than reported in earlier shear-flow simulations. This scaling is consistent with the prediction by \cite{Terquem2021}, suggesting that the effective turbulent viscosity depends only weakly on $\omega_\text{T}$, although our simulations are restricted to $\omega_\text{T}\lesssim 10\omega_\text{c}$. Moreover, we find no evidence of inverse energy transfer (or ``negative viscosity''), a phenomenon observed in some prior shear-flow simulations. We further investigate the influence of rotation within the same local framework. Slow rotation ($\Omega\lesssim \omega_\text{T}$) tends to enhance the tidal power, whereas fast rotation ($\Omega\gtrsim\omega_\text{T}$) significantly suppresses it. We discuss the limitations of our approach and the broader implications of our findings.
arXiv:2607.15159v2 Announce Type: replace-cross
Abstract: We construct a new family of Calderbank-Shor-Steane (CSS) codes using the generator and parity-check matrices of Low-Density Generator Matrix (LDGM) codes, with row operations applied to both matrices in order to achieve the desired quantum rate. Decoding is performed in an iterative manner, by applying message passing over the associated graph, and discrete Density Evolution (DDE) is used to optimize performance in the depolarizing channel. The proposed construction offers high flexibility and easiness in the design, producing quantum codes that possess excellent error correction capabilities. By properly designing the structure of the code, we are able to control and bound the weight of the stabilizer generators to a small value, which results in codes particularly well suited for fault-tolerant quantum computation. At the same time, these codes achieve very good performance in terms of error correction capability.
arXiv:2607.15700v2 Announce Type: replace-cross
Abstract: Quartet correlations in neutron-rich Te isotopes are investigated within the quartet Bardeen-Cooper-Schrieffer (BCS) framework. Taking $^{100}$Sn as an inert core, we consider two valence protons and valence neutrons occupying the $2d_{5/2} \oplus 1g_{7/2}$ model space, and solve the quartet BCS variational equations with a charge-independent isovector pairing interaction. The effective pairing strength is constrained from empirical neutron pairing gaps in the Te isotopic chain. We find that the valence quartet number increases as the valence neutron number is enlarged from $N_{\rm val}=2$ to $14$. The same increasing behavior is also found for the condensed quartet component. The proton occupation of the $1g_{7/2}$ orbit is strongly enhanced relative to the conventional like-particle BCS reference and is driven close to the degeneracy-weighted limit. These results suggest that additional valence neutrons enhance the quartet admixture in the correlated quartet BCS state, while redistributing the fixed proton weight from pair-like configurations to quartet configurations.
arXiv:2604.03614v2 Announce Type: replace
Abstract: Global optimization of black-box functions from noisy samples is a fundamental challenge in machine learning and scientific computing. Traditional methods such as Bayesian Optimization often converge to local minima on multi-modal functions, while gradient-free methods require many function evaluations. We present a novel neural approach that learns to find global minima through iterative refinement. Our model takes noisy function samples and their fitted spline representation as input, then iteratively refines an initial guess toward the true global minimum. Trained on randomly generated functions with ground truth global minima obtained via exhaustive search, our method achieves a mean error of 8.05 percent on challenging multi-modal test functions, compared to 36.24 percent for the spline initialization, a 28.18 percent improvement. The model successfully finds global minima in 72 percent of test cases with error below 10 percent, demonstrating learned optimization principles rather than mere curve fitting. Our architecture combines encoding of multiple modalities including function values, derivatives, and spline coefficients with iterative position updates, enabling robust global optimization without requiring derivative information or multiple restarts.
arXiv:2604.03829v2 Announce Type: replace
Abstract: Mamba is an emerging, complex workload with various short-range and long-range dependencies, nonlinearities, and elementwise computations that are unable to run at near-peak speeds on modern hardware. Specifically, Mamba's complex dependency graph makes fusion across its full operator cascade difficult, leaving substantial inter-operator memory traffic on the table. To address these challenges, we propose Mambalaya, a novel reconfigurable accelerator that leverages fusion to overcome the limitations of Mamba. We use the recently proposed cascade-of-Einsums abstraction to characterize Mamba's full computational structure, then apply the extended Einsum framework to systematically explore inter-Einsum fusion opportunities. This principled approach yields a series of fusion mappings that reduce off-chip inter-Einsum traffic. These mappings are supported by the underlying Mambalaya architecture. Mambalaya achieves a layer performance speedup of 4.9$\times$ for prefill and 1.9$\times$ for generation over MARCA. In prefill-dominated scenarios, it achieves up to 1.5$\times$ over a recent fine-grained, memory-aware fusion accelerator for Mamba.
arXiv:2603.07598v2 Announce Type: replace
Abstract: Chain-of-thought (CoT) reasoning improves problem solving, but long think traces increase inference cost. Existing CoT compression methods usually optimize completion-level length. For structured thinking models, however, a completion contains both a think segment and an answer segment, so completion-level compression can save tokens by compressing not only the CoT but also the answer. We call this failure mode answer drift. We propose Segment-wise CoT Compression with Answer Alignment (SCA), an answer-preserving think-compression method. SCA parses completions into functional segments, routes compression rewards only to successful think tokens, and protects answer tokens through length and distribution alignment to a frozen base model. Experiments show that, across datasets from multiple domains, SCA achieves state-of-the-art-level chain-of-thought compression while preserving the base model's performance and answer alignment. Training data and code are included in the supplementary code and data package.
arXiv:2603.08173v2 Announce Type: replace
Abstract: Quantization has become essential for the efficient deployment of speech processing systems. Although widely studied, most existing quantization methods were developed for vision and NLP architectures, while the specific challenges of audio signals remain largely overlooked. In particular, we show that audio activations can exhibit large calibration ranges, leading to significant information loss when standard calibration techniques are applied. To address this, we propose ESC, an Evolution Strategy-based Calibration method that formulates activation scaling as an optimization problem and solves it using a two-step local-global scheme driven by an evolution strategy. ESC enables unaltered performance under full INT8 quantization and is the first calibration method to achieve near-lossless performance for full INT4 quantization across multiple speech tasks. Integrating ESC with PTQ methods further reduces performance loss, achieving a 1% relative accuracy degradation on the AST model.
arXiv:2604.04788v2 Announce Type: replace
Abstract: Large language models (LLMs) could produce systematically misaligned output, from hallucinated citations to strategic deception of evaluators, yet these phenomena are studied by separate communities with incompatible terminology. We propose a unified taxonomy organized along three complementary dimensions: degree of goal-directedness (behavioral to strategic deception), object of deception, and mechanism (fabrication, omission, or pragmatic distortion). Applying this taxonomy to 50 existing benchmarks reveals that every benchmark tests fabrication while pragmatic distortion, attribution, and capability self-knowledge remain critically under-covered, and strategic deception benchmarks are nascent. We offer concrete recommendations for developers and regulators, including a minimal reporting template for positioning future work within our framework.
arXiv:2604.15699v2 Announce Type: replace
Abstract: Graph self-supervised learning can reduce the need for labeled graph data and has been widely used in recommendation, social networks, and other web applications. However, existing methods often underuse high-frequency signals and may overfit to specific local patterns, which limits representation quality and generalization. We propose Frequency-Corrupt Based Graph Self-Supervised Learning (FC-GSSL), a method that builds corrupted graphs biased toward high-frequency information by corrupting nodes and edges according to their low-frequency contributions. These corrupted graphs are used as inputs to an autoencoder, while low-frequency and general features are reconstructed as supervision targets, forcing the model to fuse information from multiple frequency bands. We further design multiple sampling strategies and generate diverse corrupted graphs from the intersections and unions of the sampling results. By aligning node representations from these views, the model can discover useful frequency combinations, reduce reliance on specific high-frequency components, and improve robustness. Experiments on 14 datasets across node classification, graph prediction, and transfer learning show that FC-GSSL consistently improves performance and generalization.
arXiv:2604.16643v2 Announce Type: replace
Abstract: Rapid progress in the field of machine-learning for weather prediction has led to the emergence of algorithms whose forecasting skill can exceed that of traditional physically based models. This development represents an opportunity to improve the quality of forecasting services provided by operational centers, particularly given the speed at which machine-learning based models generate predictions. Despite the clear promise of these systems, questions remain about the ability of the current generation of machine-learning models to generate physically consistent predictions of the full suite of required forecast fields under all conditions. Answering these questions will require careful comparisons between the well-understood physically based models, current state-of-the-art machine-learning models, and the hybrid models that combine elements of these two archetypes. The Weather Prediction Model Intercomparison Project (WP-MIP) is a World Meteorological Organization-supported initiative whose initial goal is to create a centralized database of physically based, machine-learning, and hybrid model forecasts to enable a distributed assessment and evaluation effort. The first instance of WP-MIP focuses on global deterministic predictions using both center-specific and common initializations to facilitate sensitivity studies. Forecasts contributed by institutions across six continents will be used to develop AI-ready verification techniques that highlight the strengths and weaknesses of each class of prediction system, with the goal of establishing best-practice guidance to model developers and national weather centers. The broad engagement of the operational and forecast-evaluation communities in WP-MIP will ensure that the project results are highly relevant to the development and deployment of next-generation weather prediction systems.
arXiv:2607.16114v2 Announce Type: replace-cross
Abstract: It is well-known that combinatorial circuits are modeled mathematically by string diagrams in monoidal categories. Given a gate set $\Sigma$, the circuits over $\Sigma$ can be thought of as string diagrams in the free monoidal category generated by $\Sigma$. In this model, circuit semantics are then given by monoidal functors out of this free category. For quantum circuits, this functor is often valued in the category of unitary matrices. This model suffices for concrete quantum circuits, but fails to describe parameterized families of quantum circuits, such as those which arise in the analysis of ansatz circuits. In this paper, we introduce an approach to parameterized circuit semantics, which is based on enriched category theory. We first introduce an abstract categorical construction, and use this to gain new insights on controlled operations and quantum communication. We then study the special cases of Cartesian monoidal parameters and monoidal closed parameters, both endowing the parameterized semantics with useful constructions. We conclude by showing that the monoidal closed case can be used to unify two perspectives on quantum control.
arXiv:2607.17201v1 Announce Type: cross
Abstract: In this work we study the Best Policy Identification (BPI) problem in online, tabular Reinforcement Learning. This is an active sequential hypothesis testing problem in which the learner's objective is to identify an optimal policy in a Markov Decision Process (MDP) with high confidence, while minimizing the expected sample complexity to do so. We consider an online setting with deterministic rewards, where the agent must strategically navigate through the MDP in order to effectively explore. Previous works in the literature have provided asymptotically optimal methods for BPI, such as the Navigate and Stop (NaS) algorithm and its variants, however existing analysis remains asymptotic. In this work, we fill that gap by providing the first non-asymptotic sample complexity guarantees for NaS, showing that its sample complexity depends not only on the characteristic time, but also on the connectivity of the underlying MDP, the curvature of the optimal characteristic time, and other instance-dependent quantities. We identify these additional attributes and make explicit their contributions to the overall sample complexity.
arXiv:2607.17086v1 Announce Type: cross
Abstract: Any viable interpretation of quantum theory needs to account for the Born rule, from which the theory gets its probabilistic empirical predictions. In this paper, we give an overview of possible approaches to this problem in the context of the Many Worlds interpretation. We argue that, for structural reasons, none of them can possibly succeed. More precisely, we argue that the Many Worlds interpretation must obtain the Born rule by proceeding either axiomatically, deductively, or inductively, and that all three of these approaches run into general, fundamental obstructions.
arXiv:2502.08829v2 Announce Type: replace
Abstract: Federated learning (FL) with non-IID data often degrades client performance below local training baselines. Partial FL addresses this by federating only early layers that learn transferable features, but existing methods rely on ad-hoc, architecture-specific heuristics. We first conduct a systematic analysis of layer-wise generalization dynamics in FL, revealing an early-emerging transition between generalizable (safe-to-federate) and task-specific (should-remain-local) layers. Building on this, we introduce Principled Layer-wise Federated Learning (PLayer-FL), which aims to deliver the benefits of federation more robustly. PLayer-FL computes a novel federation-sensitivity metric efficiently after a single training epoch to choose the optimal split point for a given task. Inspired by model pruning, the metric quantifies each layer's robustness to aggregation and highlights where federation shifts from beneficial to detrimental. We show that this metric correlates strongly with established generalization measures across diverse architectures. Crucially, experiments demonstrate that PLayer-FL achieves consistently competitive performance across a wide range of tasks while distributing gains more equitably and reducing client-side regressions relative to baselines.
arXiv:2603.17068v2 Announce Type: replace
Abstract: Structured 3D representations such as keypoints and meshes offer compact, expressive descriptions of deformable objects, jointly capturing geometric and topological information useful for downstream tasks such as dynamics modeling and motion planning. However, robustly extracting such representations remains challenging, as current perception methods struggle to handle complex deformations. Moreover, large-scale 3D data collection remains a bottleneck: existing approaches either require prohibitive data collection efforts, such as labor-intensive annotation or expensive motion capture setups, or rely on simplifying assumptions that break down in unstructured environments. As a result, large-scale 3D datasets and benchmarks for deformable objects remain scarce. To address these challenges, this paper presents an affordable and autonomous framework for collecting 3D datasets of deformable objects using only RGB-D cameras. The proposed method identifies 3D keypoints and robustly tracks their trajectories, incorporating motion consistency constraints to produce temporally smooth and geometrically coherent data. TrackDeform3D is evaluated against several state-of-the-art tracking methods across diverse object categories and demonstrates consistent improvements in both geometric and tracking accuracy. Using this framework, this paper presents a high-quality, large-scale dataset consisting of 6 deformable objects, totaling 110 minutes of trajectory data. Project page: https://roahmlab.github.io/trackDeform3D-core-tracking/
arXiv:2603.17154v2 Announce Type: replace
Abstract: We study the problem of coded information retrieval for block-structured data, motivated by DNA-based storage systems where a database is partitioned into multiple files that must each be recoverable as an atomic unit. We initiate and formalize the block-structured retrieval problem, wherein $k$ information symbols are partitioned into two files $F_1$ and $F_2$ of sizes $s_1$ and $s_2 = k - s_1$. The objective is to characterize the set of achievable expected retrieval time pairs $\bigl(E_1(G), E_2(G)\bigr)$ over all $[n,k]$ linear codes with generator matrix $G$. We derive a family of linear lower bounds via mutual exclusivity of recovery sets and develop a nonlinear geometric bound via column projection that holds for every linear code. For codes with no mixed columns, this yields the hyperbolic constraint $s_1/E_1 + s_2/E_2 \le 1$, which we conjecture to hold universally whenever $\max\{s_1,s_2\} \ge 2$. We analyze explicit codes, such as the identity code, file-dedicated MDS codes, and the systematic global MDS code, and compute their exact expected retrieval times. For file-dedicated codes, we prove MDS optimality within the family and verify the hyperbolic constraint. For global MDS codes, we establish dominance by the proportional local MDS allocation via a convex-ordering argument for hypergeometric distributions, simplifying and extending prior work to the asymmetric case. Finally, we characterize the limiting achievability region as $n \to \infty$: the hyperbolic boundary is asymptotically achieved by file-dedicated MDS codes, and is conjectured to be the exact boundary of the limiting achievability region.
arXiv:2605.12111v2 Announce Type: replace
Abstract: We study a sequential resource allocation problem motivated by adaptive network recruitment, in which a limited budget of identical resources must be allocated over multiple rounds to individuals with stochastic referral capacity. Successful referrals endogenously generate future decision opportunities while allocating additional resources to an individual exhibits diminishing returns. We first show that the single-round allocation problem admits an exact greedy solution based on marginal survival probabilities. In the multi-round setting, the resulting Bellman recursion is intractable due to the stochastic, high-dimensional evolution of the frontier. To address this, we introduce a population-level surrogate value function that depends only on the remaining budget and frontier size. This surrogate enables an exact dynamic program via truncated probability generating functions, yielding a planning algorithm with polynomial complexity in the total budget. We further analyze robustness under model misspecification, proving a multi-round error bound that decomposes into a tight single-round frontier error and a population-level transition error. Finally, we evaluate our method on real-world inspired recruitment scenarios.
arXiv:2607.09992v2 Announce Type: replace
Abstract: Modern ML serving increasingly lets learned, unverified components (routers, latency-SLO admitters, admit ladders) decide a tenant's quality of service; when one is wrong, the assured SLO can silently break, and the Kubernetes layers beneath (Kueue, DRA, the Gateway-API Inference Extension, GAIE) add cross-layer surprises. Rather than trust the learner to be right, we bound the damage a wrong one can do: a small trusted guard wraps the untrusted learner (learned proposes, the guard disposes). A tenant's assured-SLO obligation splits into two parts with different epistemics. Its safety projection, a per-class, per-window assured floor (with an optional drop rule, doom-sound only under an assumed service lower envelope), is a controllable obligation a guard enforces at runtime, holding it regardless of a learned admitter that is arbitrarily wrong within a bounded proposal interface. The admission floor is enforced structurally; given the stated assumptions, the service floor follows as a conditional response-time implication. Its aggregate obligation (the population tail-latency percentile) has no per-request enforcement point, so we treat it as a statistical residual and screen it. On real 2xV100 the guard (a Simplex-style assured-floor gate plus assured-first priority) holds assured-class miss 0.0 across every tested miscalibration of a learned admitter that, unguarded, misses 0.86-0.94; against a live deployment of the GAIE Flow Control, an injected mapping fault (emulating an untrusted mapper) flips the same assured requests from miss 0.0 to 1.0 (a mechanism-level trust-boundary test, not a head-to-head), while our guard reserves by the true class. As a Frontiers submission we evaluate the stance on commodity 2xV100 and a serving simulator, scoping datacenter scale, real-model Flow Control, and a closed worst-case theorem as the agenda.
arXiv:2606.07939v2 Announce Type: replace
Abstract: Research on artificial intelligence and work assigns each occupation a single exposure score. We build an instrument to see what those scores average over: a decomposition of 1,961 O*NET work activities into 15,817 atomic micro-actions by a consensus multi-agent LLM pipeline, clustered from text alone into seven semantic classes. Projecting exposure indicators onto these classes reveals two extreme poles, tool-mediated physical execution and planning-and-design, separated by a gap far larger than random partitions of the same data produce (permutation $P < 10^{-4}$; Cliff's $\delta = 0.80$ under our tech-risk index and $0.90$ under GPT-4 task ratings). The poles flank a broad central band that carries most work and is only weakly more compressed than chance. The poles are stable across clustering resolution, sentence encoder (under a common partition), and indicator, yet which pole is most exposed has inverted since 2013: the two extremes swap identity between the Frey-Osborne computerisation era and the LLM era, and at the occupation level an occupation's 2013 automatability declines as its linguistic content rises ($\rho = -0.40$, $n = 618$). We release the instrument and its outputs. The durable object for forecasting is the structure of work itself, not any era's exposure ranking.
arXiv:2603.25468v2 Announce Type: replace
Abstract: This paper investigates how eight disciplinary research data repositories from the geosciences and social sciences navigate metadata conflicts - conflicts in implementations of the same standard and inter-standard conflicts - and how these conflicts affect the completeness of DataCite metadata. It combines results from analyzing DataCite metadata records, structural differences between three disciplinary metadata schemas and the DataCite Metadata Schema, and a direct comparison of two metadata records describing the same dataset. The results show that both conflicts in implementations of the same standard and inter-standard conflicts contribute to incomplete DataCite metadata. In addition to inherent differences between the metadata schemas, workflows and conscious decisions by the repositories also contribute to these conflicts. Some of these conflicts could be resolved by updated metadata crosswalks, emerging initiatives for retroactive collaborative metadata enrichment, the implementation of metadata application profiles or schema updates. Other conflicts can't be resolved or are strongly connected to the repository mission. The results also highlight that metadata completeness is multifaceted, and assessment requires careful consideration of context.
arXiv:2605.28021v2 Announce Type: replace
Abstract: Out-of-distribution (OOD) detection is essential for deploying machine learning models in open-world and safety-critical scenarios, where test inputs may deviate from the training distribution and overconfident predictions on unknown samples can lead to unreliable decisions. Outlier Exposure (OE) has emerged as a promising OOD detection paradigm by introducing auxiliary outliers during training to enlarge the margin between in-distribution (ID) and OOD samples. Existing OE-based methods typically enlarge this margin by employing uniform labels to maximize the entropy of OOD samples over ID categories. However, we theoretically show that uniform labels inevitably disregard the relations between OOD samples and ID categories, termed the over-softening effect, leading to a suboptimal margin bound. Our theoretical analysis further reveals that explicitly exploiting such relations can instead yield improved OOD detection performance. Motivated by this insight, we propose \underline{A}daptive Confidence \underline{OE} (AOE), a simple yet effective method that leverages temperature scaling to recalibrate outlier labels. Specifically, AOE generates adaptive soft targets from temperature-scaled model predictions for OOD samples, where the learnable temperature smooths the prediction distribution without fully erasing class-wise relational information. By supervising OOD samples with these adaptive soft targets, AOE preserves the semantic proximity between OOD samples and ID categories while encouraging the softened targets to approach a high-entropy distribution, thereby suppressing overconfident OOD predictions and enlarging the separation margin. Extensive experiments across diverse benchmarks demonstrate the effectiveness of AOE.
arXiv:2607.18215v1 Announce Type: cross
Abstract: Non-Abelian gauge fields play a key role in describing the behavior of particles whose motion is coupled to internal degrees of freedom, such as their spin. Here, we experimentally realize a tuneable non-Abelian gauge field in an active mechanical lattice by using pairs of oscillators to encode a local pseudo-spin for each site, with inter-site spin-dependent couplings engineered via real-time measurement and feedback. We experimentally extract Wilson-loop observables in our set-up and hence demonstrate that we can create a genuinely non-Abelian gauge field. We then exploit the controllability of our mechanical lattice to engineer non-reciprocal hoppings to explore non-Hermitian non-Abelian gauge potentials. For a two-dimensional (2D) lattice, we demonstrate that the non-Hermiticity can manifest in direction-dependent Wilson loops for a single plaquette, while for a one-dimensional (1D) system, we show that a non-Abelian gauge potential can switch the localization of non-Hermitian skin modes between opposite ends of a chain. Our work establishes active mechanical lattices as a flexible and programmable platform for probing non-Abelian gauge fields and exploring their interplay with non-Hermitian dynamics.