Papers

Motivation: Interactive web tools enabling non-programmers to explore single-cell datasets, embeddings, gene expression, metadata, and marker genes are standard laboratory infrastructure. However, actively maintained tools (e.g., ShinyCell, ScRDAVis, sCIRCLE) rely on R Shiny and require Seurat objects. Python/scanpy, the dominant framework for single-cell RNA-seq and spatial analysis, lacks a lightweight, native option pairing a shareable viewer with a scriptable Python API. Sharing scanpy results currently requires converting to Seurat or sharing code notebooks. Results: We present scPyviewer, a web viewer that ingests AnnData objects directly and reproduces incumbent interaction patterns within the Python stack. In a feature-parity audit against three R Shiny tools, scPyviewer matches or exceeds baseline capabilities (7/7), uniquely offering native AnnData ingestion without conversion and cross-dataset comparison over shared genes and matched cell types. Benchmark tests across three datasets (22,315 to ~313,000 cells) show scPyviewer renders every core view faster (up to 3.6x speedup) using lower memory (5.2x lower on the smallest dataset). On 313,000 cells, scPyviewer completes all views while Seurat exhausts an 8 GB memory limit and fails. Beyond the app, scPyviewer exposes a Python API returning Matplotlib figures and pandas tables for publication-ready output. Availability and implementation: Implemented in Python 3.11 (scanpy 1.11.5, anndata 0.12.19, streamlit 1.59.2, plotly 6.9.0) with a one-command interface to reproduce benchmarks, figures, and launch the app. Source code: https://github.com/xuan13hao/scPyviewer.git

Predicting cellular responses to chemical and genetic perturbations is a central test for AI virtual cells, yet many single-cell models reduce each condition to an average expression shift. This target is misaligned with perturbation biology: rare states, resistant subpopulations, and context-dependent effect sizes may coexist under the same intervention. MRFlow treats virtual-cell prediction as conditional transport between empirical single-cell measures. Control and perturbed cells are embedded in a variational latent state space, greedy low-cost supervision is constructed over detached minibatch states, and a conditional flow-matching vector field transports control populations toward perturbed populations. The field decomposes into a mechanism-structured affine dictionary, which recombines shared latent response operators with a low-rank condition-specific correction, and a spectrally controlled residual field for nonlinear cell-level variation. Genetic perturbations are encoded through a regulatory-graph GCN, chemical perturbations through Morgan fingerprints, and both enter the same conditional dynamics interface. Across eight gene and drug perturbation benchmarks, MRFlow attains the highest composite score while preserving competitive raw performance on differential-expression and distributional metrics. The results support latent distribution transport as a principled and empirically effective formulation for perturbation-aware virtual cell modeling.

Neural audio codecs enable low-bitrate speech compression and tokenization, yet their robustness on accented and multilingual speech remains under-evaluated. We benchmark seven open-source neural codecs (DAC, EnCodec, FocalCodec, LanguageCodec, SemantiCodec, UniCodec, WavTokenizer) on three African speech datasets afrinames, afrispeech dialog, afrispeech multilingual, reporting signal-level quality (NISQA, UTMOS, ViSQOL, STOI, F0-RMSE) alongside two downstream tasks: automatic speech recognition (ASR) and speaker verification (ASV). Our analysis yields four findings. First, signal metrics differ sharply in downstream validity: reference-based structural/intelligibility measures (ViSQOL, STOI) and prosodic error (F0-RMSE) track ASR and ASV degradation far more reliably than neural mean-opinion-score predictors (NISQA, UTMOS). Second, intelligibility and speaker-identity preservation diverge sharply across architectures, and the apparent identity ranking itself depends on the ASV backend. Third, degradation is strongly domain-dependent and largest for conversational dialog. Fourth, the resulting degradation is partly recoverable: parameter-efficient codec adaptation (LoRA, $\sim$1–3% of parameters) on roughly 35 hours of African speech reduces the compression-induced word-error-rate gap, recovering recognition toward the uncompressed baseline (developed in companion work). These results motivate task-aware, domain-representative, and adaptation-aware evaluation of speech codecs as a prerequisite for inclusive deployment.

Agent deployments increasingly combine language-model inference with retrieval, delegation, tool execution, external-system access, and human approval. Security-relevant deviations can therefore emerge across an evolving process rather than in one isolated input or action. Building on the author's earlier black-box architecture for agentic processes and the subsequent evidence-claim model, this paper proposes an Agentic-Process Intrusion Detection System (A-IDS), an evidence-aware security interpretation layer for runtime intrusion detection whose monitored object is the agentic process itself. A-IDS compares evidence-supported observations with a governed and versioned expectation baseline for workflow state, authorization, communication, and mandatory events. Its conceptual contribution combines dynamically due governed expectations, visibility separated from three-valued matching, explicit unresolved observation states, and bounded findings that separate evidentiary status from operational impact. The model further identifies the monitoring plane itself as an attack surface when adversarial content reaches semantic evidence producers through otherwise legitimate observation paths. Some observations may be produced outside the operational agent's self-report path, but the model does not assume complete observability or universally trustworthy capture. Prompt injection is treated both as an input-security problem and as a possible origin of later process deviations and cross-agent influence paths. A-IDS does not infer malicious intent from anomalous behavior, does not treat an unobserved event as proof of non-occurrence, and does not claim a new anomaly detector, temporal logic, or provenance model. The contribution is conceptual: it does not validate an implementation, demonstrate empirical detection performance, establish causal attribution, or provide an enforcement mechanism.

Biological neurons achieve computational diversity through specialized types: tonic, bursting, adapting, and fast-spiking cells coexist within the same circuit. Artificial neural networks, by contrast, apply a single activation function uniformly to all nodes, which limits what they can represent. We show that this uniformity creates hard limits for evolutionary search: across sparse evolved substrates, monotonic functions fail to solve parity beyond its smallest instance, XOR, while a single oscillatory unit suffices at all tested scales. The gap is one of search and sparsity, not representation: monotonic networks can represent parity with a modest number of hidden units, and gradient descent recovers that solution. We evolve, to our knowledge for the first time in indirect encoding, per-node activation function assignments from an 18-function palette across more than 4,500 experimental runs spanning Boolean logic, regression, and spatial classification. Testing each of the 18 functions individually on Parity-4 reveals a three-tier solvability structure: oscillatory functions achieve 100%, intermediate functions 6.7-80%, and all 9 monotonic functions 0%. This divide is not universal. Recurrence collapses it, and gradient descent inverts it entirely, showing that the barrier is specific to evolutionary search in sparse substrates. What activation functions a network can use, beyond its topology and weights, determines what evolutionary search can solve. Indirect encoding discovers heterogeneous per-node activation assignments unlikely to be chosen by hand.

Open-ended artificial life systems must acquire diverse competencies from a single evolving genotype. Biological brains combine neuromodulation, which reconfigures circuits without changing connections, with diverse neuron types matched to specific computational roles. Can artificial evolution achieve something analogous in indirectly encoded substrates? Using indirectly encoded substrates evolved via CPPNs, we show through more than 10,000 experiments that neuromodulation alone is insufficient: under evolutionary search, monotonic activation functions impose a 75% ceiling on parity tasks that persists regardless of capacity, topology, or population size. This is an evolutionary search barrier, not a representational limit, since Adam gradient descent achieves 100% on the identical architecture. We combine neuromodulation with per-task activation function selection, matching oscillatory primitives to parity tasks and monotonic to threshold tasks, producing multi-behavioral evolved substrates. The result: 100% simultaneous 5-task success across all 30 seeds (median 14 generations). This generalizes across the oscillatory activation class: all four functions reach 100% (30 seeds each). Neither mechanism suffices alone. The barrier extends to higher-arity and asymmetric tasks, while multi-layer depth provides an alternative path. For open-ended evolution, the computational primitive should itself be an evolvable trait. At inference, one evolved genotype expresses many behaviors.

Simultaneously transmitting and reflecting reconfigurable intelligent surfaces (STAR-RISs) extend conventional reflecting-only surfaces by enabling controllable full-space propagation. Yet, once deployed, the physical locations of their elements remain fixed, leaving the surface geometry unable to adapt to users, sensing targets, blockage, or near-field focusing conditions. This article develops a system-level perspective on movable-element STAR-RIS (ME–STAR–RIS) for integrated sensing and communication (ISAC), where the surface jointly reconfigures its electromagnetic response and the physical positions of its elements. We explain how geometric reconfiguration can reshape the effective aperture, spatial correlation, interference nulls, and sensing illumination while preserving STAR-RIS full-space operation. A two-timescale control architecture separates relatively slow element motion from fast beamforming and transmission/reflection control. A 100-realization illustrative case study compares ME–STAR–RIS with an otherwise identical fixed STAR-RIS under a passive coupled transmission/reflection response. The results show a substantially improved sampled communication–sensing tradeoff and, importantly, a rapid saturation of the rate gain with modest element travel. We conclude with representative use cases, implementation constraints, and a research roadmap covering mobility overhead, channel acquisition, mutual coupling, near-field operation, hardware impairments, and learning-assisted predictive control.

The kernel of a binary code containing the zero word partitions the binary coordinates into blocks, two coordinates lying in the same block when every kernel word takes the same value in both, and the kernel-block rank profile is the multiset of the dimensions of its linear span punctured on those blocks. For the family $H^{t_1,t_2,t_3}$ of $\mathbb{Z}_2\mathbb{Z}_4\mathbb{Z}_8$-linear Hadamard codes, this invariant is known explicitly and gives a complete classification of the family. In this paper, we compute it for the $\mathbb{Z}_2\mathbb{Z}_4$-linear and $\mathbb{Z}_{2^s}$-linear Hadamard families with which those codes are compared, and we prove that it is constant for every nonlinear $\mathbb{Z}_2\mathbb{Z}_4$-linear Hadamard code and for every nonlinear $\mathbb{Z}_{2^s}$-linear Hadamard code $\bar H^{a_1,\dots,a_s}$ with $s\geq2$. The second statement is obtained without any rank formula, by exhibiting coordinate permutations that preserve the code and act transitively on its kernel blocks, which makes the argument uniform in $s$. We also prove a descent theorem: if $\bar H^{a_1,\dots,a_s}$ is nonlinear and $a_1\geq2$, then the code punctured on one kernel block is the $\mathbb{Z}_{2^{s-1}}$-linear Hadamard code $\bar H^{a_1,\dots,a_{s-1}}$. Consequently, the constant local rank equals $rank(\bar H^{a_1,\dots,a_{s-1}})$ and is at least $t-\kappa+2$, where $2^t$ is the length and $\kappa$ the kernel dimension. These results separate every nonlinear $\mathbb{Z}_2\mathbb{Z}_4\mathbb{Z}_8$-linear Hadamard code from every $\mathbb{Z}_4$-linear, $\mathbb{Z}_2\mathbb{Z}_4$-linear and $\mathbb{Z}_{2^s}$-linear Hadamard code of the same length, except for the single infinite family $H^{1,1,t-4}$ and $\bar H^{2,0,t-5}$ with $t\geq5$. The members of this infinite family agree in the rank, kernel dimension and the kernel-block rank profile, but a two-block refinement separates them.

Vision-based anti-UAV systems must function in poor visibility, yet most benchmarks use only clear-sky footage, and previous robustness studies treat adverse weather as a simple present/absent condition. As a result, the impact of fog severity on ground-to-air UAV detection remains poorly understood. This work presents the first severity-controlled fog benchmark for this task: synthetic fog at ten severity levels is applied to the RGB modality of the Anti-UAV300 dataset, comparing a clear-trained YOLOv5m baseline to a fog-aware model trained on both clear and foggy images. Detection performance drops sharply and non-linearly: degradation is front-loaded across light-to-moderate fog (beta approximately 0.05-0.10), with a 96% reduction in mAP@0.5:0.95 from clear to thickest fog, mainly due to lost recall and confidence. On the comparable metric (mAP@0.5), this collapse exceeds the most extreme rain degradation reported in the closest prior benchmark. Fog-aware training boosts detection across all severities (up to +0.320 mAP@0.5:0.95) with only a 9.1% drop in clear-sky accuracy, raising the threshold for reliable detection while not preventing collapse under extreme fog. Since reliability is lost within a narrow visibility range, simple clear vs adverse tests underestimate operational risk.

This paper presents a reproducible, educational study of evasion attacks in image classification and text classification. A compact convolutional network trained on MNIST reached 98.63% clean test accuracy and was evaluated under two white-box attacks. Under FGSM, accuracy fell to 60.20% at $\epsilon$ = 0.15 and 1.72% at $\epsilon$ = 0.30; under PGD it fell to 32.47% and 0.41%, and a bit-depth-reduction defense recovered only part of the loss. In the second experiment, DistilBERT fine-tuned on the SMS Spam Collection reached 98.75% accuracy and a 94.96% F1-score, but a controlled sequence of pre-defined perturbations (character substitutions, whitespace noise, and a benign suffix) produced only modest probability shifts in most displayed examples and no flip from spam to ham. Adversarial vulnerability is strongly modality-dependent: the MNIST experiment is a clear evasion demonstration, whereas the text experiment is a controlled robustness evaluation. Robustness must be tested empirically rather than inferred from clean accuracy.

The arXiv preprint server originally democratized astronomical research by digitizing the circulation of early manuscripts and removing the financial barriers of journal subscriptions. However, a system designed for the lower publication volumes of the 1990s has fractured under modern research output. Today, the sheer volume of unreviewed manuscripts—often tagged as “submitted,”, “under review”, or “comments welcome”—imposes a severe cognitive “tax” on the community. Researchers are forced to track multiple, unvetted versions of the same work, creating a massive redundant workload. This practice also bypasses the blinded peer-review process by establishing premature precedence, and it takes a heavy toll on early-career researchers. The pressure to rapidly post unvetted results stifles independent research, pushing young scientists toward massive mainstream collaborations where individual creativity gets lost. In this Short Communication, I analyze a decade of astro-ph metadata to quantify this trend. To safeguard both researchers' time and research freedom, I propose bifurcating submissions into two separate streams. Level-1 (peer-reviewed science) would host peer-reviewed science (accompanied by a journal acceptance or final DOI), while Level-2 (pre-review drafts) would be reserved for discussion drafts. This distinction would remove the unfair advantage of establishing false precedence through drafts and drastically cut the daily reading burden by allowing researchers to opt out of tracking unvetted versions.

Federated learning enables collaborative model training across institutions without centralizing data, yet high annotation costs, domain shifts, and class imbalance remain major obstacles, especially when irrelevant out-of-distribution (OOD) samples dilute the labeled data. Existing active learning methods target uncertainty or diversity within in-distribution (ID) data and overlook unknown samples in federated clinical settings. We propose FIDAL, an open-set federated active learning framework that combines calibrated global-local evidential uncertainty, support-set diversity weighting, and adaptive OOD rejection. The rejection gate thresholds a foundation-model Gaussian-coverage signal per client and per round with Otsu's criterion, so that highly informative ID samples are queried while irrelevant outliers are excluded without any hand-tuned threshold. Evaluated on three multi-center medical imaging benchmarks (dermatology, histopathology, and mammography with organically occurring artifacts) in realistic open-set scenarios, FIDAL outperforms detector-based open-set methods by up to about 12 percentage points of balanced accuracy and is the only method on the accuracy-ID purity Pareto front of all three benchmarks. At an equal query budget it spends at least 1.3 times fewer annotations on OOD samples than every accuracy-matched baseline, saving an estimated 7-29 hours of expert reading on the mammography benchmark. By labeling only a fraction of the data pool, it matches or exceeds fully supervised performance across modalities. These results highlight the value of integrating uncertainty, diversity, and OOD rejection in open-set federated active learning for medicine.

Vision-language models are increasingly used for driving-scene understanding, yet the semantic relations expressed in their outputs are often difficult to verify against the underlying traffic situation. This paper introduces a deterministic multi-dataset predicate framework that derives driving-scene semantics from measurable geometric, kinematic, temporal, map, and traffic-control evidence. Dataset-specific interfaces are used only to recover the required scene information, while predicate definitions remain unchanged across nuPlan and nuScenes and are materialised in a common Predicate Knowledge Graph. Quantitative semantic validation against manually annotated predicate relations on 200 scenarios from each dataset yields macro F1 scores of 0.94 on nuPlan and 0.93 on nuScenes, with an average cross-dataset difference of 0.02 across the shared predicates. The Predicate KG is further evaluated using a frozen LLaVA-OneVision-7B model on the nine NuPlanQA subtasks. Predicate grounding achieves the highest accuracy among the evaluated visual-input conditions in seven of nine NuPlanQA subtasks, including Traffic Light (53.2% to 71.5%), Situation Assessment (76.2% to 86.1%), and Action Recommendation (82.9% to 89.0%). Weather/Lighting remains essentially unchanged (89.4% vs. 88.8%), consistent with the absence of corresponding predicates, while Predicate KG only input outperforms metadata-only input in eight of nine subtasks. The results show that deterministic predicates provide a consistent and traceable semantic representation and, under oracle grounding, can reduce visual dependence for reasoning tasks covered by the predicate vocabulary.

Clinical trajectory models are usually evaluated by next-event accuracy on observed histories. Simulation is different: models must condition on their own generated events, allowing errors to compound. Although this problem is well known in sequence modelling, it has not been systematically quantified for clinical trajectory simulators. We developed EDSim-Bench to evaluate this failure mode using 425,028 MIMIC-IV-ED stays, with external replication on MC-MED, and release the evaluation protocol and scoring code. Starting from held-out visit prefixes, models generate the remainder of each visit and are evaluated on termination, event composition, timing, conditional fidelity, and occupancy forecasting, with a train-only order-3 n-gram as a reference baseline. Despite next-event accuracies within 0.001, three neural architectures behaved very differently under rollout. Across seeds, one Transformer recipe ranged from 0.43 to 0.96 in termination score and from 4.2- to 137-fold the divergence of the n-gram; no prefix-trained neural model approached the n-gram on termination or event composition. Inference-time interventions improved termination but did not jointly recover composition and timing. Supervising every eligible sequence position rather than only the final prefix position was associated with one to two orders of magnitude lower divergence across Transformer, GRU, and LSTM models, with the pattern persisting under model scaling, temporal shift, and external-site evaluation. Nevertheless, even the best model generated visits approximately half as long as observed, and model rankings reversed on occupancy forecasting, a downstream quantity relevant to bed management. These results show that next-event accuracy is insufficient to evaluate clinical trajectory simulators and motivate closed-loop evaluation across seeds, rollout criteria, and downstream tasks.

This paper develops a theorem-level pipeline in symmetry-quotient settings: quotient linear stability implies quotient flatness, quotient flatness implies input smoothness, and input smoothness yields generalization under local covering assumptions. Flatness is often associated with generalization, and Stochastic Gradient Descent (SGD) is frequently viewed as implicitly biased toward flat solutions. However, standard flatness measures are typically defined in the raw parameter space and are therefore not invariant under function-preserving symmetries such as positive rescaling. We develop a symmetry-aware theory of quotient flatness, quotient linear stability, input smoothness, and generalization on quotient spaces of neural-network parameters. For square loss and models equipped with function-preserving group actions, we define quotient flatness as the trace of the Hessian of the empirical loss on the regular quotient manifold. We show that quotient flatness controls input smoothness through a quotient-space analogue of the flatness-to-smoothness argument. We also prove that one-step mean-square quotient linear stability of the linearized SGD dynamics implies an explicit quotient-flatness bound in terms of the batch size and learning rate, and extend this analysis to higher-order tensor moments. Finally, under local covering and boundedness assumptions, we derive population generalization bounds in terms of quotient flatness and, consequently, in terms of quotient linear stability.

In this study, a physics-enhanced neural operator framework is proposed to enhance the generalization prediction ability of the cooling layout of a turbine endwall with variable number of film holes. Specifically, inspired by the film cooling superposition principle, we propose a film cooling prediction model, namely superposition-based deep neural operator (SDNO), that divides the endwall temperature field prediction into two stages. In the first stage, the cooling layout of a turbine endwall is divided into several sub-parts with randomly assigned film holes, and a Transformer-based neural operator network, namely Calculate Net, is designed to predict the temperature field of each sub-part. Then, in the second stage, another neural operator network, i.e., Super Net, is trained to combine the temperature fields predicted by Calculate Net for each sub-part and obtain the superposed temperature field of the full cooling layout. Additionally, instead of directly taking the film cooling contours as pixel plots, a signed distance function (SDF) which is sensitive to the variable locations of cooling holes, is designed to encode the location information of cooling holes. Furthermore, the proposed endwall film cooling prediction model is trained with the samples that changing the number of film holes from 1-5 with variable locations. Then, the trained prediction shows excellent generalization prediction ability, which can accurately predict the film effectiveness of the cooling layout with 10-20 film cooling holes that are unseen in the training samples. The proposed SDNO also improves prediction accuracy relative to the fully supervised baseline. With the above, the effectiveness of our proposed prediction model has been well demonstrated.

Electroencephalography (EEG) analysis is evolving from short-segment classification toward long-horizon interpretation that demands iterative evidence accumulation, multi-step reasoning, and coordinated use of specialized signal-processing tools. Although large language models (LLMs) have recently shown promise as autonomous agents for EEG analysis, existing EEG agentic evaluations remain fragmented, covering limited tasks over narrow temporal horizons with inconsistent protocols, and providing no comprehensive assessment of agents' reasoning, tool-use, and workflow construction capabilities. To address this gap, we propose EEGAgentBench, a unified benchmark for systematically evaluating LLM agents on short- and long-horizon EEG analysis. EEGAgentBench spans six representative EEG applications ranging from knowledge question answering to sleep staging. It encompasses signal durations from 2 seconds to nearly 23 hours, with prediction targets ranging from class labels to event intervals and epoch-level sequences. This design supports unified evaluation across knowledge reasoning, short-horizon interpretation, long-horizon event detection, and sequential understanding. The benchmark further provides 10 deterministic EEG analysis tools that expose only task-relevant signal measurements. Agents must therefore select tools autonomously, accumulate evidence iteratively, and construct multi-step workflows. For evaluation, we benchmark 29 frontier LLMs from 15 model families. Results demonstrate that EEGAgentBench effectively distinguishes agent capabilities beyond model scale and inference cost, while revealing substantial limitations of current LLM agents in long-horizon EEG analysis, particularly in sustained evidence accumulation and multi-step reasoning.

Large language models (LLMs) are proposed as synthetic respondents for pilot testing, but their usefulness depends on whether they supply the evidence assessment development requires. We calibrated rating scale models on 14 digital-use skill items from 6,245 adults and used the human item parameters to evaluate responses generated for 1,300 demographically matched personas. LLM responses had high internal consistency ($\alpha \approx .94$) but used the lowest category in 0.2-0.3% of responses versus 13.7-21.5% for humans, and no persona chose it on every item. These gaps changed the response-category judgment on both subscales and the PC targeting judgment; on the human-calibrated scales, 12 of 14 items had infit below 0.70, indicating responses more predictable than the Rasch model expects. Preregistered changes to the prompt, category order, and persona information did not restore the human lower range. High internal consistency is insufficient evidence that LLM responses can replace human pilot data.

Multimodal stroke recurrence prediction requires effective integration of heterogeneous clinical and imaging data, yet modality imbalance often causes models to over-rely on dominant modalities and underutilize complementary information. While self-supervised pretraining and selective parameter freezing are commonly employed to improve representation learning and fine-tuning stability, their effect on modality contributions and cross-modal behavior in multimodal medical models remains largely unexplored. In this work, we investigate whether image pretraining on 3D CTA scans reduces modality imbalance and improves cross-modal integration for stroke recurrence prediction, a clinically critical task we recently addressed. To this end, two multimodal neural networks are pretrained in a self-supervised manner and subsequently fine-tuned using two distinct freezing strategies. Their performance and modality utilization are compared against both the baseline model from our previous work and models trained entirely from scratch in this study. Our results demonstrate that self-supervised pretraining enables more effective utilization of the multimodal image-tabular dataset, outperforming both the prior baseline and all non-pretrained models. Notably, the best-performing Vision Transformer based neural network successfully overcomes unimodal collapse. Synergy analysis reveals significant interactions between vision and both gender and CHD, suggesting clinically relevant patterns for stroke recurrence. Overall, our findings demonstrate that self-supervised pretraining and strategic fine-tuning support more balanced modality utilization and enable meaningful cross-modal interactions. Code is publicly available at https://github.com/ChristianGappGit/SSL_Pretraining.

Adaptive demodulation reference signal (DM-RS) placement in 5G New Radio (NR) trades pilot overhead against channel-tracking robustness, making a single static density suboptimal across heterogeneous mobility and propagation regimes. This paper studies a lightweight per-UE controller intended for O-DU-side distributed application (dApp) execution. The controller consumes eleven deployable uplink-CSI statistics and selects among six standards-derived PDSCH DM-RS configurations. Rather than optimize oracle-class accuracy, we train an 11-128-64-32-6 multilayer perceptron with a utility-aware objective that preserves the throughput structure of all candidate actions. In a large-scale trajectory-disjoint evaluation covering 294 channel/speed/SNR conditions and 7,056 observations, the proposed U-SoftCE policy significantly improves throughput over a training-selected best fixed pattern by 1.30, 0.86, and 1.13 percentage points on in-distribution, CDL channel-OOD, and 300-km/h mobility-OOD splits, respectively; paired 95% bootstrap intervals exclude zero in all three cases. An independent bit-accurate Sionna NR LDPC validation over 36 held-out conditions and 50 transport blocks per candidate retains a +1.21-point gain with a 95% cluster-bootstrap interval of [0.27, 2.13]. A native C++ implementation uses 12,070 dense parameters and requires 5.59 microseconds mean inference time, while a FlexRIC-based emulated O-DU prototype demonstrates sub-10-ms closed-loop operation. These results support utility-aware objectives for compact adaptive PHY controllers while exposing the limits of exact class-accuracy optimization.

Many stochastic systems in operations and economics exhibit feedback between their long-run state distribution and the transition law governing their dynamics. In this paper, we develop a computational framework for stationary equilibria in such measure-dependent Markov systems when this feedback operates through a finite-dimensional aggregate. We show that the original stationary-equilibrium problem can be reduced to a finite-dimensional self-consistency equation, separating steady-state analysis of the underlying Markov system from equilibrium computation. We use properties of the resulting self-consistency map to guide the choice among fixed-point iteration, relaxed fixed-point iteration, and minimization of the fixed-point residual. The last approach requires derivatives of the self-consistency map, which are typically unavailable in closed form. We therefore develop finite-time infinitesimal perturbation analysis estimators for these derivatives, with error bounds that separate Monte Carlo error from finite-time bias. We illustrate the framework through a strategic $G/G/c$ queue and an opinion-dynamics model, showing how different structural properties lead naturally to different equilibrium-computation methods.

Designing polymers with targeted properties requires navigating vast chemical spaces from limited labeled data. Here we introduce PolyLatentFlow, a framework based on continuous-time flow matching in latent space for unconditional and conditional polymer generation, together with LlamaUni, a multimodal representation combining polymer sequence and 3D structural information. In unconditional generation, PolyLatentFlow with LlamaUni produced the largest yield of valid candidates novel relative to PolyInfo among the evaluated unconditional generators while maintaining high diversity. For $T_g$ conditioning, generated property distributions shifted systematically across a 200 {\deg}C target range. In multi-property tasks, molecular representations showed similar surrogate target fidelity but differed markedly in validity, training-set replay, and structural proximity to labeled polymers. PolyLatentFlow with LlamaUni consistently combined high validity with low replay and achieved the largest per-attempt yield of nonreplayed target hits for CO$_2$/N$_2$ conditioning. These results demonstrate latent space flow matching for polymer inverse design and identify molecular representation as a key determinant of target control and exploration beyond labeled chemistry.

In scientific simulation, regular grids, unstructured meshes, and particle-based formats are chosen to represent field data for computational efficiency, geometry/adaptive flexibility, and following motion/deformation, respectively. Each of these field data formats is often handled through separate data-specific processing pipelines. We present a unified sample-based Gaussian encoding method that represents these data forms under a single fixed-budget formulation. The method initializes and refines Gaussian primitives directly from the input samples while preserving a prescribed primitive count and encoded size to achieve a desired level of data reduction. Across structured, unstructured, and particle data, the sample-based formulation improves reconstruction accuracy with measurably fewer primitives in comparison to prior formulations, achieving up to 4.8 dB higher PSNR with an approximate 44x reduction in primitive count. For time-varying data, warm-starting from the previous timestep reduces the optimization required to reach independently trained reconstruction quality. Together, these results demonstrate a unified fixed-budget Gaussian encoding framework for structured, particle, unstructured, and time-varying scientific data with predictable storage, higher reconstruction accuracy, and improved temporal encoding efficiency.

Medical evaluation is shifting from static option-based questioning to realistic clinical scenarios with open-ended output modes. Grading these at scale naively, however, is expensive, and rubric-based evaluation has become the dominant scalable alternative. We ask what happens when the rubrics themselves are not airtight, and whether such flaws can be detected and corrected. We apply RIFT, a global rubric failure taxonomy, to two clinical benchmarks (HealthBench Professional and LiveMedBench), and find failure modes are meaningful: on HealthBench Professional an LLM judge flags 29.6% of criteria as non-atomic and 65.4% as misaligned/rigid. Then, we show that these flaws are meaningful and not simply cosmetic. As an example, rewriting bundled criteria of the form "at least one of / all of the following" as equally weighted children and regrading identical responses shifts scores by up to 15.9 percentage points on affected conversations, with disjunctive bundles inflating scores and conjunctive bundles deflating them. We also find that RIFT generally under-detects bundling on clinical rubrics, flagging 3.3% of LiveMedBench criteria as non-atomic where surface-form analysis finds structure in 25.8%.

We study the rational quadratic differential equation \[ \frac{\mathrm{d}y}{\mathrm{d}x} = \frac{ay^2+by+cx}{y^2}, \qquad a,b,c\in\C, \] under the non-degeneracy assumptions \[ c\neq 0,\qquad 2ay+b\not\equiv 0. \] Equivalently, after clearing the denominator, we consider the polynomial vector field \[ \dot{x}=y^2,\qquad \dot{y}=ay^2+by+cx. \] We give a complete, directly checkable classification of its Darboux curves. If $a=0$, no non-constant Darboux polynomial exists. If $a\neq 0$, a non-constant Darboux polynomial exists if and only if \[ c=-ab\qquad\text{or}\qquad c=-2ab. \] In these two cases the unique irreducible Darboux polynomials, up to non-zero constant multiples, are respectively \[ y-ax,\qquad y^2-2bx. \] Consequently every non-constant Darboux polynomial is a non-zero constant multiple of a positive integral power of the corresponding irreducible factor. We then place this classification in the framework of Riccati (R-)integrability and rational potentials. Excluding the exceptional branch $c=-2ab$, the analytic integrability theorem identifies $c=-ab$ as the branch admitting a global Riccati-type analytic first integral. For $c=-2ab$, we compute the first four transverse variational groups along the transformed Darboux divisor over the rational function field. Their dimensions are $1,2,3,4$; the fourth is the full group of invertible fourth-order transverse jets. We prove that this exceptional branch admits no Riccati first integral and hence is not R-integrable. Finally, we relate the branch $c=-ab$ to the Quartic Inverse Riccati (QIR) class and discuss an invariant-based classification problem for quartic Abel equations.

Graph-RAG systems often assume pristine data quality, overlooking the severe impact of perturbations in cloud-native databases. This paper proposes a three-layer decoupled diagnostic framework to orthogonally attribute system errors to reasoning loss, Knowledge Graph (KG) defects, and Cypher generation errors. Evaluated on a spatio-temporal ecological KG of the Southeastern Tibet region with eight defect types, results reveal that data integrity, rather than algorithmic reasoning, is the dominant performance bottleneck, with structural defects degrading system accuracy from 0.93 to 0.39. Crucially, we observe a masking-like phenomenon termed the Parametric Knowledge Masking Effect (PKME), suggesting LLMs compensate for broken retrieval paths using internal memory. This shrinks apparent query generation errors by over 70 percent, obscuring actual storage deterioration and increasing the risk of false negatives for automated monitoring. This work provides a quantitative foundation for auditing and optimizing data integrity in cloud-based information fusion systems.

A folder of competent skills is not yet a reliable system. Skills rarely fail alone; they fail at the seams of a collection. As an agent's skill library grows, procedures compete for attention, aliases double-load, boundaries blur, and poorly sized skills turn routing errors into task failures. We introduce SkillSeam, a method that audits the relationships through which individual skill files become a system. It maps each collection-level principle to a failure mechanism, its strongest observable, and a controlled perturbation test. From one sealed skill system, SkillSeam perturbs six design principles: persistence gradient, system coherence, regime gating, orthogonal coverage, flow, and granularity discipline. Crucially, each principle is evaluated through the channel its failure mechanism predicts rather than through accuracy alone. Flattening the persistence hierarchy increases loaded-skill tokens by 60%; a dangling anchor raises total tokens by 64% and shifts accuracy by -3.1pp; with skill count and context size held fixed, replacing an unrelated control with a synonymous alias raises noncanonical routes from 0/32 to 15/32 and flips half of matched paraphrase pairs; in a candidate-ownership audit, overlapping lanes raise reported ownership conflicts from 0/16 to 14/16; bland triggers drive routing conflicts from 3/32 to 30/32 and inflate loaded-skill tokens 3.7x; and one granularity mis-mix produces the largest accuracy drop, -12.5pp. These outcomes turn six pieces of authoring advice into testable system properties without treating every probe as confirmation. We release the byte-differenced variants, task slices, rollups, and a one-screen design checklist so that other skill systems can measure the same failure channels.

3D point-cloud observations are inherently ambiguous in complex, cluttered manipulation scenes, where target objects may be partially occluded or tightly intermingled with visually similar distractors. As a result, standard 3D diffusion policies often struggle to localize and exploit task-relevant geometry as scene complexity grows. We propose Attention-DP3, a spatially object-aware 3D diffusion policy that injects object-level geometric cues via attention while keeping the DP3 diffusion backbone unchanged. Our pipeline performs open-vocabulary 2D segmentation on RGB images, then lifts predicted target masks into 3D using calibrated camera geometry to obtain object-centric geometric priors. We incorporate these cues through Tri-field Attentional Conditioning, which constructs three complementary fields: (i) a targetness field to anchor the target object, (ii) an intra-target saliency field to emphasize task-relevant geometry within the target, and (iii) a backgroundness field to suppress distractors and clutter. Experiments on Adroit, DexArt, MetaWorld, and the real-world SO101 platform show consistent improvements over DP3, achieving state-of-the-art performance across benchmarks. Notably, as distractor objects increase, DP3 drops sharply, whereas Attention-DP3 remains stable and outperforms DP3 by up to 31% under heavy clutter. The code is publicly available at https://github.com/zhangzhongbo2213/Attention-DP3.

We study fifteen properties of binary relations that have established uses in modal logic, order and preference theory, and relation algebra. The organising question is pragmatic: once some properties of a relation are known, which further properties follow, which combinations force degeneracy, and which properties remain independent? We give a proved Horn basis of elementary and compound entailments, explicit countermodels for non-entailments, and a relation-algebraic translation of all fifteen properties. The modal discussion includes the usual Scott–Lemmon correspondences and the logic of transitive dense frames studied by Ghilardi and Mints. Exhaustive finite enumeration and targeted model search are used for discovery, while the catalogue is certified in Lean. A kernel-checked coverage calculation ranges over all 2^15 = 32,768 antecedent sets and all fifteen possible consequents. Its 51-rule Horn manifest is coupled definitionally to Lean proofs of semantic soundness, and its 49 witness rows are coupled definitionally to Lean-certified finite or infinite relations. Consequently, over non-empty domains, the Horn closure is complete for positive entailment among the fifteen selected properties, and the consistency classification is complete for their positive combinations.

How a reservoir encodes input history jointly in density and geometric degrees of freedom remains insufficiently understood. We address this mechanistic gap with a continuous density on an internal manifold coupled to an independently evolving Chern-Simons (CS) connection through a conserved current. On the same internal manifold, we compare reciprocal transport, instantaneous transverse reconstruction, local nonlinear feedback, and coupled CS evolution. These controls isolate how transport and independent connection evolution contribute to memory. We identify a current-generated exact connection component that the tested instantaneous reconstruction and scalar local closure miss. Resetting this component at fixed density and Gauss-law curl changes the future density response within the chosen gauge and drift law and the small effect persists under numerical refinement. We then distinguish dynamical memory from its readout accessibility. Although the four models have similar scalar-memory scores and density traces of pulse order, their circulating and longitudinal traces differ. With all measured features, coupled CS improves pulse-order decoding by \(6.5\pm1.9\) percentage points over the instantaneous-transverse control in paired seedwise comparisons, while remaining comparable to local feedback. Coupled CS shows less lag than the instantaneous and local-feedback controls on the first cycle, but more on subsequent cycles, revealing a dependence on preparation. We thus connect a field-theoretic mechanism of history retention to controlled tests of its dynamical role and computational use. The four-model framework provides a way to determine what an evolving transport field contributes when density snapshots and aggregate performance alone leave the memory mechanism unresolved.