arXiv:2606.26254v1 Announce Type: cross Abstract: Quantum computing has seen multiple recent breakthroughs and is getting closer to demonstrations of an exponential advantage over classical computing for certain problems. Programmers will require high-level, general-purpose, executable programming languages to express quantum solutions clearly and effectively, and the field has already produced a wide variety of such languages. This paper presents a language classification framework and uses it to survey ten popular quantum programming languages. The findings include conceptual and experimental comparisons that result in a list of challenges for future language design.
Science Journals
arXiv:2606.26498v1 Announce Type: cross Abstract: This paper addresses model-free continuous-time mean-field control in a setting where the population dynamics evolve continuously according to an unknown McKean-Vlasov stochastic differential equation, while only discrete-time transition data are available. In the model-based formulation, policy evaluation is naturally described by a stationary Hamilton-Jacobi-Bellman equation on $\mathcal P_2(\mathbb R^d)$, but this equation involves the drift and diffusion coefficients of the controlled McKean-Vlasov dynamics, which are not identifiable when only discrete-time data are available. On the other hand, a direct reduction to a time-discrete Bellman equation avoids the non-identifiability issue but loses the differential equation structure. To bridge these two viewpoints, we introduce a Mean-Field-PhiBE (MF-PhiBE), which incorporates discrete-time transition information into a continuous-time PDE on the Wasserstein space. The MF-PhiBE replaces the unknown infinitesimal drift and covariance in the policy-evaluation equation by one-step estimators computed from data, while preserving the generator structure of the McKean-Vlasov HJB equation. We also derive a policy-gradient theorem for entropy-regularized randomized feedback policies, expressing the actor direction through an action-wise infinitesimal advantage and the score of the policy. Combining these two ingredients yields a model-free actor-critic method. We prove a first-order consistency estimate showing that the value induced by an optimal MF-PhiBE policy approximates the optimal continuous-time value with an error of order $\Delta t$. In the linear-quadratic case, we show our approximation achieves second-order accuracy with only one-step data. Numerical experiments on an LQR benchmark and a crowd-aversion problem illustrate the proposed framework.
arXiv:2606.26563v1 Announce Type: cross Abstract: Single-cell studies require analysts to convert raw measurements into specific biological claims through multi-step workflows and integration of metadata, assay context, and auxiliary evidence. Existing AI-biology benchmarks largely measure broad knowledge, executable workflows, or local analysis steps. We introduce scBench-Long, a benchmark for long-horizon single-cell biology in which agents must recover scientific conclusions from raw or near-raw data without prescribed methods. The benchmark contains 21 evaluations spanning melanoma CD8 T-cell reactivity, CD8 RNA+ATAC regulatory inference, human--monkey chimera development, KRAS-driven lung tumor aging, and lethal COVID-19 lung pathology. Tasks cover paired scRNA/TCR sequencing, RNA and chromatin profiling, cross-species transcriptomics, combinatorial scRNA-seq, single-nucleus RNA-seq, immune repertoires, ortholog maps, ligand--receptor resources, and validation evidence. Candidate claims are reproduced, reviewed, and converted into controlled answer vocabularies with deterministic grading and trajectory rubrics. Across 1,068 completed trajectories, the strongest model--harness pair passes 16/63 runs (25.4\%). scBench-Long evaluates whether agents can move beyond local analysis steps and make complex scientific claims that are supported by single-cell data.
arXiv:2606.26577v1 Announce Type: cross Abstract: Self-limiting saturation curves, monotone responses that rise from zero to a plateau, govern gas adsorption, enzyme kinetics, dose-response pharmacology, and the growth per cycle of atomic layer deposition (ALD), yet mapping each curve from a handful of costly measurements is a shared bottleneck. The standard surrogate, a stationary-kernel Gaussian process, enforces no shape constraint: on sparse, noisy data it produces unphysical dips that corrupt both the inferred curve and the uncertainty used to choose the next experiment. We present an active-learning platform built on Bayesian monotonic I-spline regression, in which every posterior curve rises from exactly zero and never decreases, the plateau is set by a measurement at maximum exposure rather than a prior, and the input at any saturation level is read from the posterior by level crossing with no kinetic model assumed. Benchmarked isotherm by isotherm on five kinetically distinct families (Langmuir, dissociative Michaelis-Menten, sigmoidal Sips, logarithmic Elovich, and dispersive Kohlrausch-Williams-Watts), with ALD process development as the working example, the same fixed surrogate recovers every curve to within measurement noise without a single non-monotone posterior draw, and noise-free sweeps show the basis itself is near-exact across each family's regimes, locating its single capacity boundary at the sharpest sigmoidal onset. Driven by ordinary uncertainty sampling, the platform reaches noise-floor accuracy within a 20-measurement budget in every regime, in as few as seven measurements, whereas random sampling succeeds in only two of the five; the predicted pulse times err only on the conservative side, toward longer pulses, near saturation. Because the basis privileges no kinetic form, the platform applies wherever a self-limiting response must be learned from scarce data.
arXiv:2606.26701v1 Announce Type: new Abstract: Generalized sparse matrix-matrix multiplication (SpGEMM) is critical in many domains. Existing CPUs and GPUs, as well as specialized accelerators, rely on static dataflows (e.g., inner product, outer product, Gustavson, etc.). Each static dataflow sacrifices some data reuse opportunities and imposes constraints on load balance. To address this inefficiency, we extend the typical SpGEMM dataflows by considering dynamism. Specifically, we add fine-grained dynamic scheduling to optimize reuse and reduce resource contention. We also develop dynamic remapping of partially completed work to improve load balance and parallelism. These ideas are formalized into a specific dataflow called Segment. To demonstrate Segment, we codesign a SpGEMM accelerator called SegFold. SegFold includes a memory controller that identifies fine-grained reuse opportunities in a local window of the stationary input array and exploits them through dynamic work assignment. It also incorporates a merge network that routes inputs to appropriate processing elements (PEs) for computation while dynamically remapping the work assigned to each PE to balance load. Across diverse densities and matrix sizes, SegFold achieves a geometric-mean $1.95\times$ speedup over state-of-the-art SpGEMM accelerators and $5.3\times$ over the best static dataflow configuration, demonstrating that adding dynamism to the dataflow design space unlocks reuse and load-balance gains that no static scheduling choice can achieve in isolation.
arXiv:2606.26591v1 Announce Type: cross Abstract: Selecting a fixed number of representative points from a finite Pareto-front approximation is a fundamental post-processing task in multiobjective optimization. This paper studies this problem for the integral R2 indicator in three objectives, where the indicator is defined as the integral of the lower envelope of weighted Tchebycheff scalarizations over the two-dimensional weight simplex. We provide two complementary algorithmic results. On the positive side, we show that the integral R2 improvement with respect to any fixed baseline is a monotone submodular set function. For the usual ideal-point based R2 indicator, with the ideal point fixed, this yields a direct gap-reduction guarantee: greedy selection closes at least a $(1-1/e)$-fraction of the maximum possible R2 gap between a fixed dominated anchor value and the best cardinality-$k$ value. We also give a tested greedy implementation that evaluates exact integral R2 values by subdivision, with worst-case running time $O(n^6)$. On the negative side, we prove that exact fixed-cardinality subset selection is NP-hard already in three objectives. The hardness proof uses a perspective transformation that maps Tchebycheff-shadow improvements to a weighted anchored-box union problem with density $(x_1+x_2+x_3)^{-4}$, and then adapts the three-dimensional anchored-box construction of Bringmann, Cabello, and Emmerich. Together, these results separate the tractable two-objective case from the three-objective case while identifying a principled approximation route based on submodular optimization.
arXiv:2606.26811v1 Announce Type: cross Abstract: The Collatz map is a simple deterministic transformation whose orbit structure remains highly nontrivial. A recent direction-phase decomposition partitions each orbit into upward and downward steps, and numerical observations indicate that the number of upward phases, $N_{\uparrow}$, follows an approximate Gamma distribution. In this work, we provide a mechanistic explanation for this statistical regularity by modeling the occurrence of upward phases in the odd-compressed, or Syracuse, version of the Collatz map as a homogeneous Poisson process. From the mean-field logarithmic balance and the geometric distribution of $2$-adic valuations, we derive closed-form expressions for the Gamma parameters: the scale parameter $\theta = 2/(2-\log_2 3)^2 \approx 11.61$ is constant, whereas the shape parameter $K$ grows logarithmically with the maximal initial value $X_0=2L+1$. We also analyze the closure conditions for periodic orbits, showing that nontrivial cycles are severely constrained, which supports the plausibility of the statistical framework. Numerical validation for $L$ ranging from $10^5$ to $10^{15}$ confirms the theory with relative errors below $3\%$, and a bias-corrected mean estimate reduces the error to $10^{-3}$--$10^{-2}\%$. These results establish a quantitative link between the arithmetic properties of the Collatz map and Gamma-type statistics, and suggest possible extensions to generalized Collatz-type problems.
arXiv:2606.26710v1 Announce Type: new Abstract: Combined-cycle gas turbines (CCGTs) play a key role in modern power generation, offering both high efficiency and reduced environmental impact. However, their complex thermo-fluid and mechanical interactions complicate fault detection, particularly when labeled fault data are scarce. In this paper, we introduce the Kalman Prototypical Network (KPN), a metric-based few-shot learning (FSL) framework specifically tailored for CCGT fault diagnosis. We model the evolution of class prototypes as latent stochastic states in a dynamic system to reduce episodic variance and improve robustness in embedding representation. Synthetic data sets generated with a high-fidelity Modelica-based dynamic simulation of an offshore CCGT system were used, simulating both normal operation and progressive leak faults under transient conditions. Application of the proposed framework on simulated leak fault detection tasks demonstrate that KPN outperforms conventional FSL methods such as Matching Networks, Relation Networks, and MAML in both accuracy and stability under varying support and query configurations. The proposed framework significantly improves training convergence and generalization by stabilizing class representations, making it well-suited for real-world CCGT fault detection where labeled data is limited.
arXiv:2606.26832v1 Announce Type: cross Abstract: Discovering new fundamental physics requires spotting subtle deviations between theoretical predictions and experimental data. This delicate comparison hinges on the precise knowledge of the integrated luminosity, the measure of how many particle interactions were actually delivered by the collider. Here, we report a landmark measurement of the integrated luminosity by the Compact Muon Solenoid (CMS) experiment for proton-proton collisions at a center-of-mass energy of 13 TeV at the CERN Large Hadron Collider (LHC). By calibrating multiple independent monitors through specialized beam-separation techniques and rigorously validating their long-term stability against well-understood Z boson production rates, we comprehensively map and minimize systematic uncertainties. Combining the findings yields a total integrated luminosity precision of 0.73% for the entire data set. This marks the most precise luminosity measurement ever achieved at a bunched-beam hadron collider. Crossing the sub-percent precision threshold per data taking year fundamentally sharpens our ability to test the standard model and establishes a vital baseline for the upcoming High-Luminosity LHC era.
arXiv:2606.26846v1 Announce Type: cross Abstract: We present a semidefinite programming framework for constructing time-varying Lyapunov densities for nonautonomous dynamical systems on a hypertorus. The formulation leverages Gram matrix representations of hybrid (real-trigonometric) polynomials. In addition, we introduce a novel block decomposition of these Gram representations to confine the blow-up of the resulting density to a prescribed set. The results are then applied to establish the almost global synchronization of a time-varying Kuramoto model and the robust almost-global stability of a parameter-varying nonautonomous system. These examples demonstrate the applicability of the proposed method and validate the theoretical results. All computational results are obtained using an open-source MATLAB implementation, as referenced in the text, thereby facilitating reproducibility of the reported examples.
arXiv:2606.27002v1 Announce Type: cross Abstract: This paper presents a deep learning-assisted methodology for the inverse synthesis of a compact, wideband inverted Doherty power amplifier (PA). Convolutional neural networks (CNNs) and genetic algorithms (GAs) are jointly employed to generate pixelated Doherty combiner networks that integrate load modulation, impedance matching, power combining, and phase compensation into a single structure. As a proof of concept, we design and fabricate a GaN HEMT Doherty PA with a pixelated output combiner. The prototype achieves a measured peak drain efficiency of 51%-63% and a 6-dB back-off efficiency of 48%-54% over 1.9-2.5 GHz. Within the same frequency range, the measured output power is 44+/-0.3 dBm. Furthermore, with digital predistortion (DPD) applied, the prototype circuit demonstrates an adjacent channel leakage ratio (ACLR) better than -53.2 dBc.
arXiv:2606.27055v1 Announce Type: cross Abstract: This paper presents solarsystem, a validated lightweight and dependency-free Python package for planetary positions and solar-lunar event calculations. The package provides heliocentric and geocentric positions for the major planets, selected dwarf planets, the Centaur Chiron, and the Moon, together with sunrise, sunset, moonrise, moonset, and lunar illumination calculations. Additional functionality includes coordinate transformations between commonly used astronomical reference systems. The implemented algorithms employ analytical models that avoid reliance on external ephemeris datasets, resulting in a portable and computationally efficient solution suitable for a broad range of astronomical applications. An optional precession correction model is included, enabling calculations either in a precession-corrected reference frame or in a fixed epoch framework, depending on user requirements. The numerical performance of solarsystem was evaluated against the JPL DE440 planetary ephemerides using the Skyfield framework as a reference. Validation experiments spanning multiple bodies and extended temporal intervals demonstrate good agreement with the reference ephemerides, with mean planetary longitude and latitude deviations of approximately 0.44 and 0.16 arcminutes, respectively. Additional validation of solar and lunar event calculations yielded timing differences of only a few minutes relative to the reference solutions, while lunar illumination estimates differed by approximately 0.2%. The package can be installed directly through PyPI while the source code, documentation, validation notebooks and example workflows are publicly available through the project repository in https://github.com/IoannisNasios/solarsystem.
arXiv:2606.27064v1 Announce Type: cross Abstract: We report independent determinations of the ground-state half-lives of $^{110}$Sn, $^{113}$Sn, and the isomeric states $^{117\mathrm{m}}$Sn (J$^{\pi} = 11/2^{-}$) and $^{123\mathrm{m}}$Sn (J$^{\pi} = 3/2^{+}$), produced via photon activation of natural tin using a TT-300HE Rhodotron accelerator. The activated samples were monitored over several months using a high-purity germanium (HPGe) detector. Time-dependent $\gamma$-ray spectra were analyzed using Gaussian peak fitting for the \SI{280.49}{keV}, \SI{391.697}{keV}, \SI{158.56}{keV}, and \SI{160.34}{keV} transitions, yielding half-lives of \SI{4.165(25)}{h} for $^{110}$Sn, \SI{116.08(94)}{d} for $^{113}$Sn, \SI{13.95(1)}{d} for $^{117\mathrm{m}}$Sn, and \SI{39.95(12)}{min} for $^{123\mathrm{m}}$Sn. Agreement with Nuclear Data Sheets (NDS) recommended values is generally observed for $^{110}$Sn, $^{113}$Sn, and $^{123\mathrm{m}}$Sn, with deviations consistent within combined uncertainties when quantified using standardized differences (z-scores). In contrast, $^{117\mathrm{m}}$Sn exhibits a statistically significant deviation from the evaluated value of \SI{13.76(4)}{d}, with a z-score indicating a discrepancy well beyond expected statistical fluctuations. This result suggests a systematic difference warranting further investigation, with potential implications for applications relying on precise decay data, including calibration, dosimetry, and astrophysical modeling.
Numerical Approximation for Path-Dependent McKean-Vlasov Control with Non-Asymptotic Error Estimates
arXiv:2606.27181v1 Announce Type: cross Abstract: Path-dependent McKean--Vlasov (MKV) control models large interacting populations with history-dependent dynamics and costs. This paper develops a unified approximation-and-learning framework for continuous time path-dependent MKV problem under open-loop controls. First, an Euler discretization scheme with piecewise-constant controls is shown to achieve a non-asymptotic error of $O(h^{1/4})$. Second, we establish a discrete dynamic programming principle and prove value equivalence between open-loop and history-dependent feedback controls, enabling optimization on a reduced filtration. Third, an interacting particle system is introduced to approximate the continuous-time value, yielding an overall error bound of $O(h^{1/4}) + O(M^{-\gamma})$ for $M$ particles and an explicitly given $\gamma > 0$. Finally, we propose a fully implementable neural-network policy-gradient method using pathwise features. Numerical experiments, including a path-dependent linear-quadratic benchmark, demonstrate the effectiveness of the algorithm.
arXiv:2606.27211v1 Announce Type: cross Abstract: We introduce a framework based on classical quasi group codes to define a class of quantum CSS codes, called quantum group codes, supporting transversal multi-control-$Z$ gates which are both addressable and parallelizable, thus allowing to efficiently implement circuits composed of non-Clifford gates at the logical level. Building on this, we use a lifting procedure of classical AG codes established from class field theory to construct good quantum group codes with improved decoding complexity and logical multi-control-$Z$ gate parallelizability. More precisely, on input a good quantum AG code over the alphabet $\mathbb F_q$ with transversal $\mathsf{C}^m\mathsf Z$ gate, we apply this lifting procedure to its underlying classical AG code and obtain a quantum group code over the alphabet $\mathbb F_{q^2}$ supporting a transversal $\mathsf{C}^m\mathsf Z$ gate as well as addressable and parallelizable $\mathsf{C}^{m-1}\mathsf Z$ gates. In addition, this quantum code admits a quasi-quadratic time decoder with a linear decoding radius. This is to be compared with the previous quantum AG codes which have a cubic-time decoder. Hence, our work implies a decrease of the time complexity of state-of-the-art magic-state distillation protocols by an almost linear factor.
arXiv:2505.05353v2 Announce Type: replace Abstract: Envy-Freeness is one of the most fundamental and important concepts in fair allocation. Some recent studies have focused on the concept of weighted envy-freeness. Under this concept, each agent is assigned a weight, and their valuations are divided by their weights when assessing fairness. This concept can promote more fairness in some scenarios. But on the other hand, experimental research has shown that this weighted envy-freeness significantly reduces the likelihood of fair allocations. When we must allocate the resources, we may propose fairness concepts with lower requirements that are potentially more feasible to implement. In this paper, we revisit weighted envy-freeness and propose a new concept called SumAvg-envy-freeness, which substantially increases the existence of fair allocations. This new concept can be seen as a complement of the normal weighted envy-fairness. Furthermore, we systematically study the computational complexity of finding fair allocations under the old and new weighted fairness concepts in two types of classic problems: Indivisible Resource Allocation and House Allocation. Our study provides a comprehensive characterization of various properties of weighted envy-freeness.
arXiv:2505.09854v3 Announce Type: replace Abstract: As end-user device capability increases and demand for intelligent services at the Internet's edge rises, distributed learning has emerged as a key enabling technology for the intelligent edge. Existing approaches like federated learning (FL) and decentralized FL (DFL) enable privacy-preserving distributed learning among clients, while gossip learning (GL) approaches have emerged to address the potential challenges in resource-constrained, connectivity-challenged infrastructure-less environments. However, most distributed learning approaches assume largely homogeneous data distributions and may not consider or exploit the heterogeneity of clients and their underlying data distributions. This paper introduces Chisme, a novel fully decentralized distributed learning algorithm designed to address the challenges of implementing robust intelligence in network edge contexts characterized by heterogeneous data distributions, episodic connectivity, and sparse network infrastructure or lack thereof. Chisme leverages the affinity between clients' underlying data distributions calculated from received model exchanges to inform how much influence received models have when merging into the local model. By doing so, it enables clients to strategically balance between broader collaboration to build more general knowledge and more selective collaboration to build specific knowledge. We evaluate Chisme against contemporary approaches using image recognition and time-series prediction scenarios while considering different network connectivity conditions, representative of real-world distributed intelligent systems running at the network's edge. Our experiments demonstrate that Chisme outperforms state-of-the-art edge intelligence approaches in almost every case -- clients using Chisme exhibit faster training convergence, lower final loss after training, and lower performance disparity between clients.
arXiv:2506.15681v4 Announce Type: replace Abstract: Recent advancements in vision-language models (VLMs) have leveraged large language models (LLMs) to achieve performance on par with closed-source systems like GPT-4V. However, deploying these models in real-world scenarios, particularly on resource-constrained devices, remains challenging due to their substantial computational demands. This has spurred interest in distilling knowledge from large VLMs into smaller, more efficient counterparts. A key challenge arises here from the diversity of VLM architectures, which are built on different LLMs and employ varying token types-differing in vocabulary size, token splits, and token index ordering. To address this challenge of limitation to a specific VLM type, we present Generation after Recalibration (GenRecal), a general-purpose distillation framework for VLMs. GenRecal incorporates a Recalibrator that aligns and adapts feature representations between heterogeneous VLMs, enabling effective knowledge transfer across different types of VLMs. Through extensive experiments on multiple challenging benchmarks, we demonstrate that GenRecal significantly improves baseline performances, eventually outperforming large-scale open- and closed-source VLMs.
arXiv:2508.08005v4 Announce Type: replace Abstract: The Maximum Clique Problem (MCP) is an NP-hard problem with wide-ranging applications in fields such as bioinformatics, network science, and social computing, yet no single algorithm consistently outperforms all others across diverse graph instances. This underscores the critical need for instance-aware algorithm selection, a domain that remains largely unexplored for the MCP. To address this gap, we propose a novel learning-based framework that integrates both traditional machine learning and graph neural networks. We first construct a benchmark dataset by executing four state-of-the-art exact MCP solvers on a diverse collection of graphs and extracting structural features. An evaluation of conventional classifiers establishes Random Forest as a strong baseline and reveals that connectivity and topological features are key predictors of performance. Building on these insights, we develop GAT-MLP, a dual-channel model that combines a Graph Attention Network (GAT) to encode local graph structure with a Multilayer Perceptron (MLP) to model global features. Experiments demonstrate that GAT-MLP outperforms all baselines, and our selector significantly outperforms the Single Best Solver. Our results highlight the effectiveness of the dual-channel architecture and the promise of graph neural networks for combinatorial algorithm selection, achieving 90.43% accuracy in choosing the optimal solver. Code and models are available at: https://anonymous.4open.science/r/GAT-MLP-7E5F.
arXiv:2508.16159v2 Announce Type: replace Abstract: Meta-learning aims to uniformly sample homogeneous support-query pairs, characterized by the same categories and similar attributes, and extract useful inductive biases through identical network architectures. However, this identical network design results in over-semantic homogenization. To address this, we propose a novel homologous but heterogeneous network. By treating support-query pairs as dual perspectives, we introduce heterogeneous visual aggregation (HA) modules to enhance complementarity while preserving semantic commonality. To further reduce semantic noise and amplify the uniqueness of heterogeneous semantics, we design a heterogeneous transfer (HT) module. Finally, we propose heterogeneous CLIP (HC) textual information to enhance the generalization capability of multimodal models. In the weakly-supervised few-shot semantic segmentation (WFSS) task, with only 1/24 of the parameters of existing state-of-the-art models, TLG achieves a 13.2\% improvement on Pascal-5\textsuperscript{i} and a 9.7\% improvement on COCO-20\textsuperscript{i}. To the best of our knowledge, TLG is also the first weakly supervised (image-level) model that outperforms fully supervised (pixel-level) models under the same backbone architectures. The code is available at https://github.com/jarch-ma/TLG.
arXiv:2606.26746v1 Announce Type: new Abstract: Collective perception messages (CPMs) introduce significant packet size variability due to dynamic object inclusion and periodic security overhead. While 5G NR-V2X Mode 2 typically employs semi-persistent scheduling (SPS) designed for periodic traffic with relatively stable packet sizes, the impact of realistic CPM-driven size fluctuations on distributed resource allocation remains insufficiently understood. This paper presents a comparative system-level evaluation of NR-V2X Mode 2 scheduling strategies under variable-size CPM traffic reconstructed from real-world perception datasets. We analyze dynamic scheduling and multiple SPS-based approaches, including padding-based allocation, aggressive size-driven reselection, and modulation and coding scheme (MCS) adaptation. Results show that packet size variability can significantly degrade reliability when scheduling stability is compromised. In particular, dynamic scheduling and aggressive reselection increase collision probability due to frequent resource reallocations. In contrast, SPS with padding converges to a stable resource allocation and provides robust performance, while MCS adaptation achieves the highest average packet reception ratio but with uneven reliability across packet types. The findings demonstrate that, under realistic CPM traffic, stability of resource usage is more critical than instantaneous load optimization, and provide design guidelines for CPS deployment over NR-V2X Mode 2.
arXiv:2509.07295v4 Announce Type: replace Abstract: Unified multimodal models (UMMs) unify visual understanding and generation within a single architecture. However, conventional training relies on image-text pairs (or sequences) whose captions are typically sparse and miss fine-grained visual details, even when they use hundreds of words to describe a simple image. We introduce Reconstruction Alignment (RECA), a resource-efficient post-training method that leverages visual understanding encoder embeddings as dense "text prompts", providing rich supervision without captions. Concretely, RECA conditions a UMM on its own visual understanding embeddings and optimizes it to reconstruct the input image with a self-supervised reconstruction loss, thereby realigning understanding and generation. Despite its simplicity, RECA is broadly applicable: across autoregressive, masked-autoregressive, and diffusion-based UMMs, it consistently improves generation and editing fidelity. With only 27 GPU hours, post-training with RECA substantially improves image generation performance on GenEval (0.73 $\rightarrow$ 0.90) and DPGBench (80.93 $\rightarrow$ 88.15), while also boosting editing benchmarks (ImgEdit 3.38 $\rightarrow$ 3.75, GEdit 6.94 $\rightarrow$ 7.27). Notably, RECA surpasses much larger open-source models and applies broadly across diverse UMM architectures, establishing it as an efficient and general post-training alignment strategy for UMMs.
arXiv:2606.18408v2 Announce Type: replace-cross Abstract: We characterize two nested Walsh parity-check filters implemented on Quandela's Belenos cloud photonic processor in a single-photon eight-mode spatial register. The modes are indexed by the vertices of the cube $Q_3$. The filters realize the classical $[8,7,2]$ single-parity-check code, the zero-sum neutral subspace $\mathcal{N}$ and the $[8,4,4]$ extended Hamming code, the parity-checked subspace $\mathcal{S}\subset\mathcal{N}$ with one DC and three face-parity syndrome channels. These are first-quantized path/mode encodings of classical codes: the experiment verifies leakage suppression and syndrome routing, not error correction or protection against photon loss, and all probabilities are conditional on postselected single-photon detections. Across more than 340,000 detections, neutral inputs show residual DC-port leakage of $0.02\%$-$1.1\%$ (mean $0.6\%$), corresponding to $\approx21\times$ suppression relative to the ideal $0.125$ DC-capture baseline and $31.6\times$ relative to the measured non-neutral control. Injected DC contamination gives a monotonic soft error signal, and the three face-parity syndrome channels route to their predicted ports with $94$-$99\%$ selectivity. A sector-preserving unitary core keeps leakage far below non-neutral controls over one to three applications, with differences dominated by calibration and compilation systematics rather than gate-cycle physics. We quantify these limits, including fixed-pattern separator bias, $\pm 0.02$ calibration offsets, and compilation scatter near the $10^{-3}$ leakage level, and report a Hong-Ou-Mandel degradation episode in which suppression vanished and recovered after recalibration.
arXiv:2604.08448v2 Announce Type: replace Abstract: AfriVoices-KE is a large-scale multilingual speech dataset comprising approximately 3,000 hours of audio across five Kenyan languages: Dholuo, Kikuyu, Kalenjin, Maasai, and Somali. The dataset includes 750 hours of scripted speech and 2,250 hours of spontaneous speech, collected from 4,777 native speakers across diverse regions and demographics. This work addresses the critical underrepresentation of African languages in speech technology by providing a high-quality, linguistically diverse resource. Data collection followed a dual methodology: scripted recordings drew from compiled text corpora, translations, and domain-specific generated sentences spanning eleven domains relevant to the Kenyan context, while unscripted speech was elicited through textual and image prompts to capture natural linguistic variation and dialectal nuances. A customized mobile application enabled contributors to record using smartphones. Quality assurance operated at multiple layers, encompassing automated signal-to-noise ratio validation prior to recording and human review for content accuracy. Though the project encountered challenges common to low-resource settings, including unreliable infrastructure, device compatibility issues, and community trust barriers, these were mitigated through local mobilizers, stakeholder partnerships, and adaptive training protocols. AfriVoices-KE provides a foundational resource for developing inclusive automatic speech recognition and text-to-speech systems, while advancing the digital preservation of Kenya's linguistic heritage.
arXiv:2606.27050v1 Announce Type: new Abstract: We analyze the maximum burst size achievable in all-optical satellite networks across different constellations. With a 100 Gbps uplink capacity, a WDM-based optical burst switching network supports burst sizes of up to 500 MB in high-altitude LEO constellations and 600 MB in low-altitude LEO constellations.