arXiv:2607.12290v1 Announce Type: cross
Abstract: Audio-language embedding models such as CLAP are widely evaluated on matching present sound events, but rarely on negation. We show this affirmation-only evaluation hides a key limitation: these models fail to encode negated sound concepts, mapping affirmative and negated captions to nearly identical representations. To expose this blind spot, we introduce NegEval-Audio, a framework that converts existing datasets into two negation-aware tasks, Retrieval-Neg and Multiple-Choice Negation (MCQ-Neg), to probe whether models distinguish present from absent events. On AudioCaps and Clotho, performance degrades sharply under negation, with negation-type MCQ accuracy falling far below chance, and the failure persists even for a recent multimodal LLM-based embedding model. While a training-free steering method improves MCQ-Neg, it yields marginal gains for Retrieval-Neg. This indicates that affirmation bias is a fundamental flaw in the representation geometry, necessitating explicit negation-aware training objectives.
Science Journals
arXiv:2607.12415v1 Announce Type: new
Abstract: We study strong numerical approximations for reflected stochastic differential equations in possibly unbounded convex domains with super-linearly growing drift and diffusion coefficients. Under a coupled monotonicity condition and polynomial local Lipschitz assumptions, we first establish the well-posedness of the reflected SDE and derive uniform moment bounds for its solution. We then introduce a coupled tamed Euler--Peano scheme, in which the drift and the squared diffusion coefficient are tamed by a common factor and the resulting Euler--Peano path is corrected through the Skorokhod problem. This common taming factor preserves the drift--diffusion coercivity structure and yields uniform moment estimates for the numerical solution. We prove strong convergence of order $1/2$ for both the constrained state process and the boundary regulator, thereby recovering the standard Euler-type strong order in this reflected setting. Numerical experiments for a reflected stochastic Ginzburg--Landau type system illustrate the constraint preservation of the scheme and support the theoretical convergence rate.
arXiv:2607.12888v1 Announce Type: new
Abstract: Tensegrity form-finding and physical property prediction are fundamental inverse problems in structural mechanics, which aim to determine equilibrium configurations and internal force distributions. These problems are challenging due to strong nonlinearity arising from the coupling between geometry and forces, the need to ensure structural stability, and the enforcement of constraints such as boundary conditions and symmetry. Moreover, traditional methods often lack robustness to noise and outliers. This paper proposes an energy-based learning framework for clustered tensegrity form finding and physical property prediction. The proposed approach incorporates total potential energy minimization and constitutive relations into the training objective, enabling the simultaneous prediction of equilibrium nodal configurations and associated physical quantities, including member forces and force densities. By incorporating energy-based physical losses directly into the learning process, the framework improves physical consistency, robustness, and data efficiency. Numerical experiments on tensegrity structures, including prism and lander systems, show the great potential of the proposed approach and demonstrate its capability for scalable form finding and accurate prediction of structural properties.
arXiv:2607.12433v1 Announce Type: new
Abstract: Diffusion models have recently become the dominant paradigm for monocular depth estimation (MDE). However, they implicitly assume that depth can be recovered as a globally smooth field through iterative denoising, which does not explicitly reflect the piecewise and scale-dependent organization of scene geometry. In practice, geometric structure emerges progressively across spatial scales, where coarse layout, surfaces, and boundaries are constructed in a hierarchical manner. Motivated by this observation, we introduce ARDepth, which formulates depth estimation as structured auto-regressive generation. Instead of recovering depth through global refinement, ARDepth progressively constructs depth representations as spatial resolution increases. To support this generative process, we introduce Scale-Progressive Conditioning (SPC) to inject multi-scale visual features at each generation stage, and Semantic-Aware Guidance (SAG) to provide scene-level semantic priors that enhance global structural consistency. Together, these designs enable the model to capture fine-grained local details while maintaining coherent global geometry. Empirical results demonstrate that our approach achieves strong performance and produces structurally consistent depth predictions across scales, validating auto-regressive generation as a promising alternative paradigm for geometric modeling.
arXiv:2607.12493v1 Announce Type: cross
Abstract: The NASA InSight mission observed over 2000 marsquakes in the course of its three year mission. These quakes varied in magnitude between 1.5 and 4.5, as well as in spectral content. We present a simple framework to describe the spectral characteristics of all observed marsquakes, based on source process; propagation through the mantle or crust; and local, receiver-side amplification. We assign to each quake an objective measure of its amplitude, as well as the spectral decay created by the duration of the rupture and the dampening of high frequencies due to visco-elastic attenuation. Together, this allows us to obtain characteristic patterns of the whole marsquake dataset, e.g. in terms of event magnitudes, source size, and - for quakes caused by meteoritic impacts - crater size. We show that a significant fraction of all marsquakes - the high-frequency quakes - form a swarm that is likely not caused by tectonic processes in rocks. Our analysis allows separation of the whole marsquake catalogue into three event classes, of tectonic quakes, meteoritic impacts, and swarm events. We finally conclude that the largest marsquake, S1222a, most likely belongs to the group of meteoritic impacts.
arXiv:2607.12681v1 Announce Type: new
Abstract: State space models (SSMs), notably Mamba, have recently emerged as efficient alternatives to self-attention with linear computational complexity. We investigate the integration of Mamba into YOLO26, the latest non-maximum suppression (NMS)-free object detection framework, by proposing MambaPSA, a lightweight Mamba-based replacement for the C2PSA block at the end of the backbone. To complement this study, we additionally insert a bidirectional Vision Mamba (BiViM) module at the P3, P4, and P5 levels of the neck. Experiments on PASCAL VOC 2007+2012 show that MambaPSA reduces parameters by 2.9%, FLOPs by 12.1%, and improves CPU inference throughput by 17.6% (from 17 to 20 FPS) with negligible accuracy change (-0.1 mAP50:95), while the P4 BiViM placement yields the best accuracy gain (+0.9 mAP50:95). These results suggest that SSMs offer a favorable efficiency-accuracy trade-off when replacing attention-based blocks in NMS-free lightweight detectors.
arXiv:2607.12620v1 Announce Type: new
Abstract: Dual-core MCUs are asymmetric: on NXP's MCXN947, the second Cortex-M33 has no FPU, DSP extension, TrustZone, or MPU. We treat the asymmetry as a design input in the Phase 3 dual-core architecture of SynapticOS, an open-source Zephyr-based runtime: the AI runtime (models, NPU/DSP, scheduler) lives on the capable core, and the application core reaches inference only via a message-based OS service -- a remote system call. The transport is a pair of lock-free single-producer/single-consumer rings in shared SRAM: one writer per index, free-running 32-bit counters, ordering by data-memory barriers alone (the platform has no cross-core atomics). Because ring state is shared, a rebooting application core rejoins unaided. Requests carry priority classes, errors and timeouts propagate to the caller, and tensors stage zero-copy in a shared slot -- a 27 KB frame cannot exist twice in 64 KB of RAM. Measured on the FRDM-MCXN947 (both cores 150 MHz): the application core boots in 1,514 us and completes the handshake in 2,554 us, bit-identical over 11 boots; round trips are 15 us typical / 81 us worst-case (50 us budget); pushes cost 25 cycles; a 1,913-serve two-model soak had zero errors (stub-NPU latencies bracket transport, not silicon). An MPU region on the runtime core guards the application core's RAM (fault-injection verified); protection is one-directional -- the application core has no MPU, and ARMv8-M cannot block privileged reads. Two hardware-revealed defects are reported: releasing the second core into erased flash wedges the whole chip and its debug port (now prevented by a ROM-API blank check), and a Zephyr flash-driver Kconfig silently disarmed the devicetree MPU guard (now programmed at runtime). Firmware is 98.9 KB flash (runtime core) and 32.4 KB (application core, 42.6 of 64 KB RAM); 108 tests in 13 suites pass 100%. Apache 2.0: https://github.com/Dimitrios-Kafetzis/SynapticOS
arXiv:2607.12532v1 Announce Type: new
Abstract: Quantum weakest preconditions are a fundamental tool for program verification of quantum programs. Many variations have been reported in the literature. We revisit quantum weakest preconditions from the perspective of expected runtime analysis of quantum programs and introduce a novel pre-expectation framework that enables to reason about the preconditions of quantum programs without the need of an upper bound. This is particularly interesting for quantum programs involving reward statements. The overall goal is to analyze runtime behavior even in the case of programs with potentially infinite expected runtime. This paper presents several ways to do so, e.g., a program transformation such that the expected runtime of a quantum program can be expressed using the weakest pre-expectation calculus with rewards.
arXiv:2607.12881v1 Announce Type: new
Abstract: Vision Transformers (ViTs) have demonstrated remarkable performance in computer vision tasks. However, their self-attention mechanism often diffuses focus across background regions, relying on spurious correlations rather than object-relevant cues. Inspired by inhibitory mechanisms observed in biological vision systems, we propose the Inhibited Self-Attention (ISA), a novel self-attention that integrates inhibitory signals to enhance feature selectivity and suppress spurious responses. In contrast to conventional self-attention, which relies solely on positive attention values due to softmax normalization, our approach retains and utilizes negative attention scores to suppress irrelevant features and sharpen focus on objects of interest. Experiments across multiple datasets, including ImageNet-1k and COCO, and several robustness benchmarks demonstrate that ISA enhances object-centric selectivity, reduces shortcut reliance, and improves out-of-distribution generalization. Our analysis of relevance maps confirms that ViTs with ISA exhibit sharper, more localized focus on object-relevant regions while reducing distractions from non-relevant (background) features, enabling more reliable models. We release our code at https://github.com/prdvanderwal/inhibited-self-attention
arXiv:2607.12603v1 Announce Type: cross
Abstract: Recent advances are enabling quantum emitters to interact with photonic crystals, whose electromagnetic modes exhibit complex dispersion relations, spatial mode structure, and polarization textures. However, modeling light-matter behavior in these systems faces a persistent trade-off: electromagnetic approaches based on Maxwell-equation solvers provide realistic vectorial descriptions but are difficult to integrate with quantum many-body and non-perturbative methods, whereas simplified quantum-optical lattice models are tractable but typically rely on scalar and spatially independent light-matter couplings that miss essential features of these structured photonic environments. Here, we introduce a constructive framework to derive quantum-optical lattice descriptions that overcome this trade-off. Combining symmetry-constrained tight-binding constructions with numerically computed photonic band structures and field profiles, our method yields minimal, symmetry-enforced lattice Hamiltonians that reproduce the target photonic dispersion while retaining the mode-resolved (position- and polarization-dependent) structure of the light-matter coupling. We show that these models recover Green's-function-based emitter dynamics in the perturbative regime, while providing access to non-perturbative quantum dynamical simulations beyond emitter-only descriptions. As a proof of principle, we apply the framework to a two-dimensional photonic crystal and show that it captures polarization-dependent directional emission inaccessible to scalar models, while enabling the analysis of non-Markovian light-matter dynamics and entanglement. Our results provide a practical bridge between classical electromagnetic simulation tools and quantum-optical many-body and non-Markovian modeling in photonic crystal settings.
arXiv:2607.12730v1 Announce Type: new
Abstract: Smart-building load forecasters are often trained offline on dense, multivariate, high-frequency data, but deployment may provide only hourly, feature-limited inputs. Missing features must then be reconstructed, and their errors can propagate through the model. If this input uncertainty is not reflected, prediction intervals may become miscalibrated, affecting demand-response scheduling. Our work examines where uncertainty should be placed once inference inputs are reconstructed. We develop a unified one-day-ahead probabilistic forecasting framework that aligns temporal resolution, reconstructs the unavailable inputs, and derives causal features, and we compare a modular post-hoc residual-quantile scheme with an integrated in-model quantile-learning scheme. The comparison uses three mid-scale Deep Learning (DL) backbones: recurrent, hybrid recurrent, and attention-based Temporal Fusion Transformer (TFT) models, under identical inputs, forecasting horizon, preprocessing rules, and training budgets. Results show that uncertainty placement is backbone-dependent. Integrated quantile learning is most reliable with the TFT, yielding 2.2-3.6% MAPE and 28-83W RMSE on the labeled test window, while producing intervals about 5x narrower than the modular intervals at the closest-to-nominal coverage level. Diebold-Mariano tests support the TFT ranking and the mixed behavior of the recurrent backbones. A reconstruction-sensitivity test shows that reconstructed inputs increase the Quantile Score (QS) by 106% while interval width remains nearly unchanged, indicating that the model does not automatically absorb reconstruction-induced uncertainty. Robustness checks against non-DL baselines and seasonal hold-out weeks support this ranking. Our results expose the limits of post-hoc residual quantiles when inference depends on reconstructed inputs.
arXiv:2607.12673v1 Announce Type: cross
Abstract: Evolved sequences can be used to infer the rules of evolution. Orally transmitted folk melodies are evolved sequences whose similarity to protein sequences (one-dimensional, drawn from a limited alphabet) invites application of bioinformatics methods to study cultural evolution. A major obstacle is that melodies encode rhythm, which breaks some assumptions of standard sequence-alignment algorithms. We develop a rhythm-aware alignment method and apply it to \num{40000} Irish dance tune variants, enabling the first large-scale automated melodic alignment. Four canonical bioinformatics analyses -- mutability, substitution matrices, positional conservation, and covariance -- reveal patterns distinct from those of molecular evolution, revealing the forces that shape each domain: biochemical and biophysical constraints for proteins; memory, motor, and social biases for melodies. Together the results show that bioinformatics provides a powerful framework -- conceptual as much as algorithmic -- for studying cultural evolution. Although the cultural transmission of music has been discussed for centuries, here we show how to analyze it at large scale.
arXiv:2607.12635v1 Announce Type: new
Abstract: OLEDs are increasingly applied in illumination and displays because they offer excellent color quality, are mechanically flexible, and are self-emissive. However, their usage is limited by low external quantum efficiency (EQE) and efficiency roll-off at high driving voltages. These limitations, together with demands for smaller pixels and device sizes in emerging technologies, motivate innovations that increase efficiency and allow replacing external optical elements with embedded solutions. Here, we demonstrate enhanced outcoupling as well as directional and polarization control of OLED emission, based on collective surface lattice resonances of plasmonic nanoparticle arrays that are embedded in the active layers of four different state-of-the-art OLED structures. Both square arrays and more complex lattices producing flat bands are demonstrated to guide the light to directions and polarizations determined by their optical modes. We show that by the design of the array geometry and the OLED structure, spectral and angular enhancement of the electroluminescence (EL), up to 30 %, can be achieved. Our results verify that surface lattice resonances of nanoparticle arrays offer a robust and versatile embedded solution for tailoring the OLED emission, as well as exciting prospects for efficiency increase if combined with narrow-spectrum emitters.
arXiv:2607.12641v1 Announce Type: new
Abstract: Wireless aerial virtual reality (VR) aims to provide immersive access to large-scale scenes, but high-resolution view generation and delivery are jointly constrained by limited bandwidth, latency, and power. 3D Gaussian Splatting (3DGS) can reduce the payload by rendering views from compact pose information, yet its geometry errors may cause severe VR quality degradation. Existing channel-aware or pixel-level resource allocation schemes fail to capture such geometry-sensitive distortion. To address this issue, this paper proposes GeoFovea-GS as a geometry-aware cross-layer framework for communication-efficient wireless aerial VR. A foveated geometry-aware distortion metric is developed to characterize photometric rendering error, geometric inconsistency, and view-dependent perceptual importance in a unified form. Based on this metric, the joint selection of pose-only 3DGS rendering and image/tile correction transmission is formulated as a cross-layer optimization problem under wireless constraints. A lightweight value-of-information scheduler is further developed to allocate communication resources to regions that are both geometry-critical and perceptually important. Experiments on real-world 3DGS scenes demonstrate that GeoFovea-GS achieves superior immersive rendering quality with substantially reduced transmission cost.
arXiv:2607.12780v1 Announce Type: cross
Abstract: Quantum circuit optimization for fault-tolerant computing requires exact functional equivalence while minimizing expensive non-Clifford resources such as T gates. We study this problem using a compact 44.8M-parameter encoder-decoder transformer with structured circuit tokenization, evaluating on parameterized circuits (2-6 qubits) and Clifford+T circuits (3-6 qubits). On parameterized circuits, a hybrid approach -- structure from the transformer, angles from classical optimization -- achieves median fidelity 1.000 on 3-6 qubit circuits. On Clifford+T circuits, where all gates are discrete and no post-processing is possible, the model learns valid syntax and accurate T-Count statistics, yet exact equivalence degrades sharply with target length -- from 88% on circuits with <=9 gates to near zero beyond 26 gates.
We trace this failure to autoregressive drift: early-token divergence cascading irrecoverably through left-to-right decoding. Two levers partially mitigate the drift: inference-time strategies that generate multiple candidates and select via equivalence verification raise exact-match rates from 7% to 22.5%, while scaling training data by 2.5x pushes them to 39.5%. Yet the degradation with target length persists -- even with more data, exact equivalence drops from 94% on short circuits to under 4% beyond 26 gates.
The contrast between settings is our central finding: when approximate outputs can be rescued by post-processing, the transformer succeeds; when exact discrete correctness is required, autoregressive drift limits reliability, with both inference-time search and data scaling as effective levers while training-side fine-tuning and model-level diversification are not.
arXiv:2607.12813v1 Announce Type: cross
Abstract: We investigate the role of parity-violating interactions between supernova neutrinos and chiral molecules in nearby interstellar molecular clouds as a potential source of biomolecular homochirality. We introduce neutrino interactions into the autocatalytic chemical reactions in a far-from-equilibrium noise-induced system. These interactions create a directional bias between L and D enantiomers in the racemization reactions, which is amplified by autocatalysis and stochastic fluctuations. We solve the stochastic equations within the Ito sense to obtain the dynamics of the probability distribution of the enantiomeric excesses, offering an astrophysical scenario for the delivery of homochirality seeds to Earth by meteorites. In spite of the weak interactions of supernova neutrinos, our framework introduces an amplification mechanism to yield a considerable enantiomeric excess of more than $10\%$, in agreement with the latest chemical analysis of the meteorites. Moreover, we scan over the parameter space of the model, inferring from the observational values in order to explore the window of opportunity to generate the initial seeds of homochiral states in interstellar molecular clouds.
arXiv:2607.11975v1 Announce Type: new
Abstract: Current machine unlearning methods predominantly rely on global, coarse-grained intervention strategies. They lack precise pilot signals to guide the unlearning process and fail to provide differentiable guidance across different unlearning tasks. Due to the varying memorization strengths of samples during original training, such a uniform strategy leads to two problems: some samples are over-unlearned, which harms model utility; while others are under-unlearned, leaving residual information that can be exploited by privacy attacks. In this paper, we propose GSUO, a guidance-signal-aware unlearning optimization framework that designs task-specific fine-grained guidance signals to steer the unlearning process and is applicable to both random-subset and class-wise forgetting tasks. Extensive experiments demonstrate that GSUO outperforms 14 baselines in terms of both unlearning effectiveness and generalization, while achieving high efficiency and significant speedups, validating its effectiveness for reliable machine unlearning.
arXiv:2607.12150v1 Announce Type: new
Abstract: As the number of missions to cislunar space increases, the population of space objects in this region is expected to grow, making efficient uncertainty propagation essential for space situational awareness (SSA). This is complicated by the cislunar domain's vastness, chaotic dynamical environment, and limited availability of measurements. This paper presents an adaptive multi-fidelity uncertainty propagation method that dynamically adjusts the included perturbing forces based on position in cislunar space, minimizing computation time while maintaining a prescribed modeling accuracy. The proposed adaptive method is then integrated into a multi-target tracking framework to reduce the computational cost of track prediction without sacrificing accuracy, which is important for managing the growing number of objects in cislunar space. The effectiveness of the approach is demonstrated in simulated test cases relevant to upcoming cislunar missions and SSA applications, resulting in a significant reduction in computational cost compared to a non-adaptive approach while achieving equivalent or superior accuracy.
arXiv:2607.12165v1 Announce Type: new
Abstract: Recently, the societal implementation of high-performance image classification models has expanded rapidly. While these models require vast amounts of training data to improve performance, securing sufficient real images is often impractical. As a means to compensate for this shortage, the use of synthetic data is becoming widespread. However, synthetic images are not necessarily equivalent to real images for training purposes. This study systematically analyzes the differences between two types of synthetic images created by different generation methods and real images from three perspectives: high-dimensional feature space, low-level statistics in color space, and the model training process. Furthermore, it experimentally verifies how synthetic data should be utilized by considering realistic data mixing scenarios. This enables the proposal of an evaluation and application strategy for performing preliminary assessments on synthetic images of unknown quality and safely incorporating them into training. This research aims to contribute to enhancing the reliability and safety of image classification models utilizing synthetic images.
arXiv:2607.12152v1 Announce Type: new
Abstract: We present a DLM-anchored hybrid physics/ML framework for brownfield optical links that accurately predicts per-channel power, OSNR, and GSNR. Calibrating span/ILA boundaries via DLM yields OSNR/GSNR errors of no more than 0.39/0.43 dB across single-channel and OSaaS provisioning.
arXiv:2607.12648v1 Announce Type: new
Abstract: In this paper we introduce Hermite spectral approximation for functions with endpoint singularities using exponential transforms, including single exponential (SE), double exponential (DE) and error function (EF) transforms, and present a comprehensive convergence analysis for these approximations without and with scaling. In the case without scaling, we show that these methods converge at some root-exponential rate. In the case with scaling, we derive optimal scaling factors for each of exponential transforms and show that the convergence rate of Hermite spectral approximation can be significantly improved. Numerical comparisons with sinc method are present and it is shown that Hermite method has comparable or superior accuracy performance when using the same number of terms. Extensions to quadrature and rootfinding algorithm are also discussed.
arXiv:2607.12176v1 Announce Type: new
Abstract: Ambivalence and hesitancy (A/H) undermine digital behaviour-change interventions, and recognizing them automatically from video is the goal of the ABAW A/H challenge on the BAH dataset. We describe our system for the 11th edition of the challenge: a calibrated, equal-weight ensemble of three fusion models over frozen face, audio, text, and pose embeddings, which reaches 0.7358 macro-F1 on the public test set. This year's private test, released on a disjoint set of 30 new participants, is scored on five allowed submissions; we report the configuration and rationale of each of our five submissions, and, where already available, the private-test score obtained. Our first submission, an exact replica of the calibrated ensemble tuned only on public validation, scored 0.7361 macro-F1 on the private test, matching our public-test estimate almost exactly and confirming the pipeline generalizes to unseen participants without leakage.
arXiv:2607.12195v1 Announce Type: new
Abstract: Semantic memory retrieval can be conceptualized as navigation through conceptual space. We compared semantic search dynamics between humans and three large language models (GPT-4o, Gemini-2.5-Pro, Claude-Sonnet-4.5) using verbal fluency data. By applying trajectory-based NLP metrics to the items generated by 82 human participants and LLM output across eight temperature settings, we quantified three complementary dimensions: entropy (step size predictability), distance to next (successive semantic steps), and distance to centroid (global dispersion). Humans exhibited higher entropy, larger semantic steps and broader dispersion than all LLMs, indicating more variable and exploratory search. Temperature tuning produced only partial alignments, as individual metrics matched between humans and LLMs at specific settings, but no configuration reproduced the complete human profile (in all dimensions). These findings suggest that human semantic search implements a distinctive balance between local exploitation and global exploration that current model architectures fail to reproduce.
arXiv:2607.12696v1 Announce Type: new
Abstract: Sparse Mixture-of-Experts (MoE) models have become an important approach for scaling Large Language Models (LLMs), but their inference efficiency depends strongly on expert activation patterns. Speculative decoding (SD) accelerates autoregressive generation by verifying multiple draft tokens in parallel, yet existing draft selection strategies primarily optimize acceptance likelihood. In large-scale MoE models, however, selecting draft tokens also determines the union of experts activated during verification. We observe that confidence-driven SD can introduce \textit{expert scattering}: high-probability draft tokens may route to disjoint experts, increasing expert-weight memory traffic and reducing the speedup from speculation. Motivated by this observation, we revisit draft-tree selection under the non-uniform memory-cost structure of MoE inference. We propose \textsc{EcoSpec}, a cost-aware speculative decoding framework that incorporates predicted marginal expert activation cost into draft selection. With a lightweight expert predictor and a dynamic expert buffer, \textsc{EcoSpec} favors draft paths that preserve high acceptance likelihood while reusing experts already covered by the current verification set, without modifying the target-model verification rule. We evaluate \textsc{EcoSpec} on three large-scale MoE models, including DeepSeek-V3.1 (671B), Qwen3-235B-A22B, and GPT-OSS-120B, across reasoning, coding, question-answering, and dialogue benchmarks. \textsc{EcoSpec} consistently reduces active expert footprints and improves end-to-end decoding speed, achieving up to $1.62\times$ speedup. These results show that accounting for expert activation cost is important for efficient speculative decoding in large-scale MoE models.
arXiv:2607.12253v1 Announce Type: new
Abstract: This paper presents lessons learned from parallelizing the legacy software known as Advancing Front Local Reconnection (AFLR) as a black box. The parallel procedure utilizes (i) a data decomposition scheme where each subdomain is refined in parallel using the sequential mesh generation code and (ii) a runtime system for work-load balancing. Results on the mesh refinement operation show that the parallel method's stability (output mesh quality) is good and that the parallel method outperforms the serial AFLR by about 11 times when utilizing 16 CPU cores. However, full stability (i.e., generating the same quality as the serial method) and potential scalability is limited due to the constraints set by the black box's input boundary requirement. Satisfying this requirement for each subdomain not only adds overhead but also causes the parallel method to generate a different output mesh volume than that generated by the serial method. The complexity of such a state-of-the-art code requires that it be modified to a non-trivial extent in order to remove these constraints. These results suggest that the parallelization of black-box legacy codes like AFLR may not be practical and instead encourages an approach that is originally designed without such constraints to efficiently leverage the concurrency offered by large-scale architectures.