Papers

We present LAION-BVD (LAION - Big Video Dataset), a large-scale open video dataset for multimodal learning, which contains 1.3B platform-specific video URLs collected from CommonCrawl. From these, we download 80M videos with a total duration of 13.5 million hours. The dataset is designed for multimodal pre-training across the video, audio, and image modalities. Using content-aware scene detection, we extract clips for which we synthetically generate video and audio captions. Models trained on these data achieve competitive performance on standard video-text and audio-text benchmarks, with consistent improvements as training or model scale increases. Additionally, we explore video frames as an alternative source of image-text data by extracting scene-changing frames. These frames exhibit a visual distribution distinct from standard web image corpora, and models trained on this dataset achieve strong image-text retrieval performance. We release LAION-BVD to the research community. It significantly expands open access to multimodal videos at an unprecedented scale.

Large language models (LLMs) are increasingly deployed as AI analysts to process financial disclosures and support AI-assisted investment decisions. Yet such systems are usually evaluated by what they can retrieve, not whether retrieved information affects their judgments. We identify a retrieval-integration gap in long-context financial analysis. Holding focal-firm information fixed and varying only unrelated context from 2,000 to 128,000 tokens, we find that a risk disclosure's influence on investment judgments falls to the experimental noise floor even as direct retrieval remains accurate. The pattern replicates across model families and judgment tasks and in experiments removing real disclosures from actual 10-K filings. More capable models postpone but do not eliminate the gap. Causal memory interventions show that compressed summaries and source-text lookup jointly transmit disclosures into judgments. Workflow architecture determines whether this transmission succeeds: chunk-and-summarize pipelines evict relevant information, whereas a targeted, structured restatement adjacent to the decision restores its influence. AI analyst performance is therefore jointly determined by model capability and workflow architecture. Retrieval-based evaluations can certify systems whose investment judgments ignore information they demonstrably retrieved.

The transition toward Open Radio Access Networks (O-RANs) is reshaping how cellular infrastructure is deployed, managed, and optimized. This paper investigates the energy-aware joint placement and migration of cloud-native functions (CNFs) in an O-RAN edge cloud. We consider both a Single-CU-UP association model and a slice-aware Multi-CU-UP relaxation, in which distinct slice-flow groups of the same distributed unit (DU) may be assigned to different Centralized Unit User Plane (CU-UP) processing targets under one Centralized Unit Control Plane (CU-CP). For brevity, these scenarios are referred to as Single-CU and Multi-CU, respectively; Multi-CU never denotes multiple CU-CP associations. We formulate the problem as a Mixed-Integer Linear Program (MILP) that minimizes server, transmission, wake-up, and migration energy while satisfying server-resource capacities and one-way delay requirements over the F1 user-plane interface (F1-U) between each DU and its selected CU-UP in a fat-tree edge data center. To improve computational scalability, we also develop a deterministic k-means-based heuristic that approximates the MILP decisions without requiring repeated exact optimization. Over the evaluated 24-hour workload, the theoretical Multi-CU relaxation reduces modeled energy consumption by 5.7% relative to the Single-CU baseline. For the Multi-CU case, the proposed heuristic remains within approximately 9.7% of the proposed MILP, demonstrating a favorable trade-off between energy efficiency and computational tractability.

DNA storage is a promising medium for long-term archiving, but its read performance is limited by coarse-grained random access. Existing random-access DNA storage designs suffer from high read amplification because their sequential layouts co-locate frequently and infrequently accessed data under the same primer pair, where any read must retrieve all associated strands even when only a small fraction is needed. We present HORIZON, a read-efficient allocation policy for DNA block devices that reduces read amplification through activity-aware horizontal placement. HORIZON first introduces a horizontal layout distributing writes round-robin across primer pairs, rather than filling each sequentially. It classifies newly written blocks in the write buffer as active or inactive, tracks recent primer-pair accesses using a sliding-window temperature model, and allocates blocks based on block activity and primer-pair occupancy. Simulations show HORIZON consistently reduces read amplification compared with state-of-the-art schemes across MSR and FIU traces and synthetic filesystem workloads.

Rapid data assimilation is required for real-time prediction and control in digital twins. For many physical systems, the data assimilation task requires the solution of a physics-constrained inverse problem, which is often computationally intractable in real time using traditional physics solvers. This work presents a reduced-basis neural operator approach to enable real-time inverse problem solutions in the digital twin setting. Our approach specifically targets the large class of problems with spatiotemporal dynamics governed by partial differential equations (PDEs) that are parameterized nonlinearly with respect to model parameters $m$, and linearly with respect to inversion parameters $q$. Based on this physical structure, our neural operator approximates the nonlinear map from the model parameters $m$ to the parameter-to-observable operator $\mathcal{F}(m)$ in a reduced subspace. Since the output of the neural operator is the parameter-to-observable operator itself (manifested as a matrix), we refer to this approach as NEural Matrix Operator (NEMO). With NEMO, for new given $m$, we enable a closed-form inverse problem solution for $q$ in a reduced subspace. We apply NEMO in two real-world applications: contaminant transport initial condition identification, and hypersonic vehicle load identification. We show that NEMO delivers high quality inverse problem solutions for data assimilation in real time, with over three orders of magnitude speedup compared to constructing the reduced operator with the PDE solver. Further, NEMO demonstrates comparable inverse performance to a state-of-the-art multiple-input neural operator, while reducing online computational complexity by over an order of magnitude and providing real-time uncertainty quantification.

In a recent breakthrough, Yamakawa and Zhandry (J. ACM 2024) constructed a proof of quantumness in the quantum random oracle model (QROM) in which the quantum prover samples a codeword preimage of a publicly computable function H. They conjectured that given any H, a successful prover must sample their preimage from a high-entropy distribution over possible answers. If true, this would give a certifiable randomness protocol in the quantum random oracle model. As partial evidence for their conjecture, Yamakawa and Zhandry proved the security of their certifiable randomness protocol assuming the Aaronson-Ambainis conjecture. We prove the security of the certifiable randomness protocol of Yamakawa-Zhandry unconditionally, without relying on the unproven Aaronson-Ambainis conjecture, against low query-depth quantum adversaries: specifically, adversaries that make up to o(\log \lambda) adaptive quantum queries to the random oracle.

The rapid expansion of large-scale assessments and the growing adoption of automatic item generation have intensified concerns about incidental content redundancy, where construct-irrelevant elements such as wording or contextual framing become unintentionally repetitive across items. Traditional similarity metrics like BLEU or cosine similarity, often fail to capture the nuanced structural and semantic layers that drive perceived redundancy simultaneously. This study proposes a dual-dimensional framework for Automated Item Similarity Analysis (AISA) powered by Large Language Models (LLMs), operationalizing similarity through Structured Decomposition and Semantic Relatedness. Psychometric validation indicates that LLM-derived metrics align more closely with indicators of construct-irrelevant local dependence and yield more coherent item parameter groupings than traditional text-based measures. The framework is further evaluated through its application in Computerized Adaptive Testing (CAT). Simulations reveal that incorporating LLM-based similarity constraints into item selection improves estimation stability and reduces bias with minimal efficiency trade-offs, outperforming constraints based on conventional metrics. These findings highlight the potential of LLM-powered AISA to support scalable bank curation, content-aware test assembly, and experience-sensitive adaptive testing across diverse assessment contexts.

Large language models are increasingly used for knowledge graph question answering (KGQA), but can fail to correctly ground answers in the underlying graph. Current approaches to LLM-based KGQA either rely on full semantic parsing into executable queries such as SPARQL, which is brittle in practice due to complex schemas or incompleteness of real-world KGs, or on LLM-reasoning and answer generation over KGs, which can be more robust but lacks formal guarantees. In this work, we study a complementary setting in which candidate answers are generated by an LLM-based system and subsequently verified using lightweight symbolic constraints derived from the question. We introduce Constrained Entity Selection under Partial Knowledge (CES-PK), a problem formulation that focuses on eliminating invalid answers and providing symbolic support for valid ones without requiring construction of executable logical forms. To account for incomplete KGs, we employ a three-valued constraint semantics (satisfied, violated, unknown) that avoids incorrect rejections under open-world assumptions. To demonstrate the effects of our method, we instantiate this framework over the Hetionet biomedical knowledge graph and evaluate the impact of type, relation, and exclusion constraints. Experiments show that precision improves by filtering invalid candidates, while recall is preserved due to retaining candidates whose constraints are not explicitly violated. Satisfied constraints provide additional positive symbolic evidence to rank remaining candidates.

Spatially resolved biology requires representations that preserve biological neighborhood structure rather than only exact cross-modal correspondences. Existing histology–transcriptomics objectives can emphasize instance-level matching even when non-paired spots share molecular or spatial context. We introduce BioKERN, a multimodal spatial representation-learning framework that incorporates biological structure as an explicit, learnable inductive bias. BioKERN constructs a training-time biological kernel by combining transcriptomic similarity and spatial proximity, then uses it to provide graded neighborhood supervision and regularize embedding geometry. Evaluation uses a fixed, model-independent biological neighborhood definition shared by all methods. Across Mouse Brain Visium and Human Liver GSE240429, BioKERN consistently improves biological-neighborhood retrieval over BLEEP in both single- and multi-scale settings. Controlled shared-architecture experiments show that most of the improvement arises from biological-kernel regularization rather than increased model capacity. These results support explicit biological geometry as an interpretable inductive bias for multimodal learning in spatial biology.

Repeated observations of one physical object need not constitute independent channel uses. We model partial nanopore protein rereads as a finite-alphabet channel with canonical content, persistent readout, and pass-local coverage and synchronization. For exact compound-pass data, matched inference approaches the equivalence-class canonical posterior, and sitewise excess Bayes risk admits an action-aware achievable exponent. In an aligned specialization, observation-local redraw can cause linear-in-$K$ growth in true-label negative log-likelihood (NLL). We derive order-$b$ projection-stability bounds and an exact passwise-fusion diagnostic. On a PASTOR-informed semi-synthetic hard-symbol channel, label-blind deterministic-mixture importance sampling (LB-IS) agrees with exact enumeration at $L=7$. At $L=24, K=10$, LB-IS meets every prespecified aggregate absolute marginal-posterior and score-agreement criterion against a fixed high-allocation reference in three selected conditions. Joint agreement holds for the representative and high-NLL conditions, while the near-zero condition remains inconclusive. Exact $L \leq 6$ benchmarks identify order 4 as the smallest tested common cap. At target scale, the reference supports selected unprojected functionals, while neither order 4 nor 5 attains joint agreement, defining a tested finite-memory boundary. Across 16 cells, the order-4 shared branch lowers NLL by 0.033-0.224 nats per residue relative to pass-local.

Neighborhood-based fairness audits evaluate individual fairness by comparing predictions among similar individuals in feature space. Despite their widespread use, little is known about the robustness of the auditing procedure itself. Because these audits rely on nearest neighbor relationships, small perturbations in feature space can alter local neighborhoods and produce different fairness assessments even when model predictions remain unchanged. We develop a geometric framework for analyzing the robustness of neighborhood-based fairness audits under bounded perturbations. Our analysis establishes sufficient conditions for neighborhood invariance, quantifies how neighborhood replacement propagates to audit instability, and introduces audit volatility, a measure of the expected sensitivity of fairness audits under repeated perturbations. Experiments on benchmark datasets support the theoretical analysis and show that the proposed framework explains the observed stability of neighborhood-based fairness audits.

Land use, transportation, and building energy shape one another, yet urban-scale studies typically model each sector in isolation. We present a co-simulation platform that couples the UrbanSim land-use model, the POLARIS agent-based transportation model, and the CityBES urban building energy model into a single integrated workflow, with POLARIS travel skims driving land use and POLARIS agent activities driving dynamic building occupancy. We demonstrate the platform with forecasts through 2045 for the Chicago metropolitan area under a business-as-usual case, a high-telecommuting scenario, and a mileage-based user fee scenario. Both policies produced expected-direction responses that emerged from the model feedbacks rather than being imposed. Telecommuting decentralized activity toward outlying areas and cut 2045 vehicle miles traveled by 12.5%, whereas the mileage fee recentralized activity toward the urban core and cut it by 2.9%. Comparing coupled runs against uncoupled runs that hold land use fixed shows that the land-use feedback contributes over one percentage point to the county-level travel effect in several counties, large relative to the policy effect itself since the mileage fee's county-level effects are only about three percent, so a transportation-only study would materially misstate the sub-regional impact. Both policies raised citywide building energy by about 1%. This is the first platform to integrate land use, transportation, and building energy simultaneously, replacing predefined occupancy schedules and static building stocks with endogenous agent-based occupancy and a forecast-driven building stock. It lets planners evaluate transportation and pricing policies for their joint land use, travel, and energy consequences, and its component models rely on nationally available data, making the approach transferable given local building-stock and calibration data.

Offline return-to-go (RTG) sweeps can test whether a recommender conditioned on return is controllable, but the intervention is rarely audited. Rewriting every historical RTG token creates an increasingly synthetic context, while rewriting only the current token is more local. We test this distinction in an offline setting with a fixed window. On MovieLens 25M and MyAnimeList 2020 (MAL), we evaluate a Decision Transformer using an RTG locality ladder, a control without RTG, a logged match and score reward check, and a within-trajectory shuffled RTG ablation. On MovieLens, a $K=20$ intervention that covers the full context, applied only to real context positions, shifts the share of Crime predictions by $+23.61 \pm 2.96$ percentage points from the validation 5th to 95th percentile, whereas changing only the current slot shifts it by $+1.77 \pm 1.17$ points. The shuffled RTG model largely removes this response ($+2.08 \pm 1.20$ points at $K=20$). On MAL, the same protocol does not produce a Drama response: $K=20$ changes Drama by $-0.03 \pm 0.07$ points, and $K=1$ by $-0.01 \pm 0.01$. Genre prediction accuracy is numerically close across real RTG, no RTG, and shuffled RTG, and at $K=1$ logged match rates and matched ratings change little. Because dataset and focus-genre selection were exploratory, these magnitudes are descriptive; the cross-diagnostic pattern across locality, shuffled RTG, and the null result on MAL does not establish reward control. We propose four checks: intervention locality, a no-RTG baseline, a reward check, and RTG-content ablation.

We uncover ELR collapse in language model pretraining: learning rate (LR) and parameter norm govern loss dynamics primarily through their ratio, the effective learning rate (ELR). When ELR is matched across runs, their loss trajectories collapse throughout training despite substantially different LRs and parameter norms. Across optimizers, architectures, datasets, and model scales, mean collapse errors are typically a few x 10^-3, below the seed-to-seed variation measured in a representative configuration. Systematic ablations identify normalization design and the timescale of LR-norm variation as key determinants of collapse precision. Controlled interventions further show that weight decay and Hyperball shape loss dynamics primarily through the ELR schedules they induce. Replacing LR with ELR enables a fitted functional scaling law (FSL) to transfer across norm-control methods. The resulting ELR-based FSL also explains delayed acceleration, a recurring effect of norm control. Together, these results establish ELR as a common coordinate linking LR scheduling, norm control, and loss dynamics.

Class-imbalanced node classification on temporal graphs is challenging because majority-dominated temporal propagation progressively assimilates minority representations, while conventional node and neighborhood information provides insufficient discriminative evidence for minority classes. To address these issues, we propose MDTE, a minority-aware diffusion framework that reconstructs stable and discriminative temporal edge-event representations through conditional diffusion denoising. Specifically, MDTE introduces Distribution-Aware Selective Propagation, which combines Local Outlier Factor (LOF)-based propagation filtering with cluster-aware low-frequency propagation. The module preserves informative neighborhood dependencies while mitigating harmful propagation and majority-class information assimilation. It further develops Multi-View Discriminative Fusion, which exploits feature reconstruction and topology prediction to characterize class-wise differences in distribution learning and extracts complementary discriminability signals to guide denoising. Experiments on five real-world datasets demonstrate that MDTE consistently achieves the best performance on minority-class-oriented metrics, improving minority-class recall by up to 23.53 percentage points, minority-class F1 by 8.68 percentage points, and AUPRC by 2.67 percentage points over the strongest baselines.

Recent work has applied Mamba style state space models (SSMs) to video anomaly detection, yet existing approaches still rely on buffering clips or windows internally, lack a theoretical account of how temporal memory relates to detection latency, and benchmark efficiency only through GPU throughput rather than the edge hardware these methods are intended to target. We introduce a strictly causal streaming anomaly detector whose fixed size state is updated in O(1) time and memory per incoming frame, with no lookahead and no clip buffering. Its temporal core is a diagonal linear state space recurrence with an input and state dependent decay gate, trained self supervised through causal next embedding prediction on a frozen visual backbone. We derive a closed form relationship between the recurrence decay spectrum and both detection delay and the shortest anomaly it can reliably capture, then validate empirically on UCSD Ped2 and CUHK Avenue. The settling delay bound predicted from the learned base decay (57 to 59 frames) sits far above the measured detection delay (1.6 and 18.4 frames), showing that the event boundary gate, not the base decay, governs responsiveness. We further report end to end latency and throughput measured directly on Apple M3 Pro hardware, 0.74 ms and 0.77 ms per frame (over 1300 FPS), rather than simulated GPU numbers. With an untuned initial configuration the method reaches 67.9 percent and 70.2 percent frame level AUC on Ped2 and Avenue, trailing prior non causal SSM baselines in accuracy. Ablations over decay rate, state size, and gating reveal that the gate contribution is dataset size dependent, hurting accuracy on the smaller Ped2 training set but helping on the larger Avenue one. Closing this accuracy gap and extending evaluation to a third, larger benchmark are immediate next steps.

We present Crase, a bounded and inspectable alternative to deep research agents for scholarly search. Instead of an open-ended search loop, Crase queries a search engine once for seed papers, expands them along their 1.5-hop citation neighborhood, prunes citation edges whose claims lack entailment support, and ranks the remaining papers with a recency-aware random walk. This makes the candidate set, the reason each paper is kept, and the stopping condition explicit and fixed before inference. On LitSearch and one further benchmarks over a 500K-paper arXiv corpus, Crase outperforms deep research agents built on proprietary models by up to 3$\times$ recall@50 at roughly a third of the cost.

Model cards are structured documents that summarize key information about machine learning models to improve transparency, usability, and accountability. However, they often lack a consistent structure, and many models provide no model cards, making comparison and interpretation difficult. This paper presents two contributions. First, we propose MCTidy, an LLM-based approach that reorganizes existing model cards into a standardized template to improve clarity and comparability. Second, we introduce MCGenie, an LLM-based system that generates model cards directly from model repository data. We apply MCTidy to 48 Hugging Face model cards and evaluate information retention, section alignment, hallucination, and stability. Our findings show high information retention with minimal textual loss, accurate section assignment, rare hallucinations primarily in descriptive sections, and strong stability across runs. We assess MCGenie by generating model cards for the same 48 models and assessing semantic similarity, factual correctness, and sensitivity to input resources. The generated model cards achieved high semantic similarity (mean around 0.9); over half were fully correct, and most remaining errors were minor. Generation quality depended strongly on the availability of supporting resources, particularly associated papers. Overall, our findings demonstrate the potential of LLM-based methods to enable scalable, standardized model card documentation.

We present StarHarness, a framework for evolving environment-specific agent harnesses while keeping model weights fixed. The evolved harness can include prompt and task framing, tool interfaces, skills, MCP-backed providers, subagent structure, and agent-loop configuration. StarHarness constructs a compact evolution pool by stratifying tasks according to baseline failure behavior, separates proposer-visible search tasks from proposer-hidden selection tasks, and reserves held-out tasks for evaluating generalization. Across ITBench SRE, EnterpriseOps-Gym ITSM, and AutomationBench Finance, harness evolution improves full-benchmark performance by 20-35 percentage points over the default harness after 4-12 accepted changes per environment. These gains persist on tasks excluded from evolution and transfer without re-evolution across GPT and Qwen model families. Trace analysis links the improvements to interface repairs, environment conventions, and operational knowledge that compresses search, with fewer false-positive diagnoses and shorter trajectories in several settings. StarHarness therefore offers a practical way to reduce persistent model-environment mismatch in tool-rich enterprise tasks.

We introduce masked auto-correlation, a new primitive for the cryptanalysis of symmetric-key primitives, together with a quantum attack pipeline built on it. For a permutation $f$, output masks $\alpha,\beta$, and an input difference $w$, masked auto-correlation (MAC) measures the correlation between the masked outputs $\alpha\cdot f(x)$ and $\beta\cdot f(x\oplus w)$. The associated masked differential-linear (MDL) approximations strictly generalize several classical techniques; ordinary linear cryptanalysis, differential-linear cryptanalysis, and the differential-linear connectivity table all arise as special cases. Our central object of study is the problem of finding mask pairs with large masked cross-correlation – those that yield powerful distinguishers – which we call MAC Fishing. We give a constant-query quantum algorithm that samples such pairs according to their squared correlation, and we prove an exponential classical lower bound of $\Omega(N/\log N)$ queries, by adapting the hardness of Fourier Fishing. To our knowledge this is the first result pairing a quantum upper bound with a classical lower bound for the core task of identifying high-correlation approximations, making quantum algorithms an absolute necessity. Building on this, we analyse the distribution of masked auto-correlation for random permutations, and then construct capacity-based distinguishers and key-recovery attacks, both classically and with a quadratic quantum speed-up using amplitude estimation. We validate our claims with experiments on reduced-round mini-AES.

Recently, the rapid advancement of multimodal domains has driven a data-centric paradigm shift in graph ML, transitioning from text-attributed to multimodal-attributed graphs. This advancement significantly enhances data representation and expands the scope of graph downstream tasks, such as modality-oriented tasks, thereby improving the practical utility of graph ML. Despite its promise, limitations exist in the current neural paradigms:(1) Neglect Context in Modality Alignment: Most existing methods adopt topology-constrained or modality-specific operators as tokenizers.These aligners inevitably neglect graph context and inhibit modality interaction, resulting in suboptimal alignment.(2) Lack of Adaptation in Modality Fusion: Most existing methods are simple adaptations for 2-modality graphs and fail to adequately exploit aligned tokens equipped with topology priors during fusion, leading to poor generalizability and performance degradation.To address the above issues, we propose LION (c\underline{LI}ff\underline{O}rd \underline{N}eural paradigm) based on the Clifford algebra and decoupled graph neural paradigm (i.e., propagation-then-aggregation) to implement alignment-then-fusion in multimodal-attributed graphs. Specifically, we first construct a modality-aware geometric manifold grounded in Clifford algebra.This geometric-induced high-order graph propagation efficiently achieves modality interaction, facilitating modality alignment.Then, based on the topology-aware Clifford components of aligned tokens, we propose adaptive holographic aggregation. This module integrates component-wise energy and propagation-scale information with learnable parameters to improve modality fusion. Extensive experiments on 9 text-image MAG datasets demonstrate that LION significantly outperforms SOTA baselines across 3 graph and 3 modality downstream tasks.

Reliable search requires more than acquiring external evidence. An agent must also recognize and recover from errors as its trajectory unfolds. In-trajectory feedback provides a mechanism for such recovery by diagnosing where the search has drifted and redirecting subsequent reasoning steps. This is particularly important in long-horizon search, where an early directional error may receive no immediate corrective signal and can compound across later steps. Making such feedback learnable, however, creates a coupled problem: the agent must learn when to request and use feedback, while the critic must learn corrections from outcome-confounded rollouts as the agent's failure patterns evolve. We introduce CAFE (Coupled Agent–Feedback Evolution), a framework in which a shared-parameter model alternates between search-agent and critic roles. CAFE initializes feedback-conditioned recovery from trajectories built around the base agent's own failures, then couples online and offline optimization. During online RL, a comparative feedback estimate uses a prompt-level call–skip success gap to shape request returns, while feedback-aware advantage shaping reweights token advantages before and after feedback. Offline, rollout-derived preference optimization learns feedback from matched successful and unsuccessful trajectories. On seven agentic search benchmarks, CAFE outperforms the evaluated RL-based search agents on average, retains its gains across all six out-of-domain benchmarks, and reduces answer-level hallucinations. One-sided ablations show that improving only the agent or only the critic eventually plateaus, whereas alternating the two updates continues to improve performance. These findings suggest that a self-improving search agent needs feedback that co-evolves with the policy it guides.

Automated malaria diagnosis from stained blood-smear microscopy is dominated by deep convolutional neural networks that are accurate but computationally expensive, poorly interpretable, and rarely validated with patient-level rigor. We present EMFE (Efficient Mathematical Feature Extraction), a five-feature framework for classifying single red-blood-cell images as parasitized or uninfected using Gray World color normalization, adaptive green-channel thresholding, morphological spot detection, and classical machine learning. Using the NIH LHNCBC malaria dataset (27,558 images from 200 patients), we evaluate Random Forest, Histogram Gradient Boosting, and Support Vector Machine classifiers under patient-grouped nested cross-validation (K_outer=20, K_inner=3), ensuring that cells from each patient remain within a single fold. The optimized Random Forest achieves 94.6% pooled out-of-fold accuracy (95% CI [93.6, 95.7]), corroborated by an untouched 40-patient holdout test (94.3%) and a patient-level permutation test (p<0.001, 1,000 permutations). Ablation experiments quantify the contribution of individual features and pipeline stages. Hardware-matched comparisons with retrained DenseNet121, ResNet50, and MobileNetV2 models assess the accuracy-efficiency trade-off. Synthetic perturbations characterize three failure modes, while explainability analysis identifies spot saturation as the dominant discriminative feature. Patient-level aggregation further quantifies sensitivity-specificity trade-offs and false-positive accumulation. These results demonstrate a statistically rigorous, interpretable, and computationally lightweight alternative to deep learning, while explicitly quantifying its limitations.

Clinicians read chain-of-thought (CoT) rationales as evidence of medical reasoning, but whether the visible chain plays that role is rarely tested. General-domain CoT-faithfulness probes ignore clinical cost, and medical LLM evaluations treat the chain as a black box. We close this gap with a medical perturbation audit: a 30-operator battery edits both the chain and the question with clinically motivated operators (severity reversal, negation flip, demographic swap, evidence ablation), paired with a chain-update times answer-flip joint analysis that classifies each model by its failure mode. Applied to 14 LLMs on four medical QA benchmarks, three independent tests converge: the Chain-Decoupling Rate (CDR; chain does not register the edit and the answer does not flip) is 72.9% panel-wide on clinically meaningful destructive edits, chain corruption leaves accuracy unchanged, and removing CoT prompting does not reduce accuracy. Two board-certified clinicians re-annotate N=197 perturbed questions; 98.5% leave the gold defensible. The pattern holds across medical and reasoning fine-tuning and scale; on the closed-source tier, where the chain text is unavailable, the answer-side signals are consistent with the same decoupling. Our framework and CDR provide a reusable yardstick for auditing whether medical CoT is faithful or merely documentation.

Color discrimination is a fundamental aspect of visualization as it influences how people interpret visual encodings. Many visualization guidelines are informed by perceptual studies, yet relatively few have been replicated. Acknowledging that the interaction between human perception, visual tasks, and display technology can change over time, we replicate two crowdsourced color discrimination studies conducted 10 years earlier. Specifically, we replicated a visualization-focused color discrimination task (N=144) and a more general perceptual discrimination task (N=394). In both studies, our results reproduced the original perceptual effects. We further use the replication to investigate whether color-related practice influences color discrimination. Specifically, we extended our replication studies by adding questions about participants' engagement with color practices. We then examined whether diverse color-related practices (e.g., artistic hobbies, knowledge of color theory, and cosmetic makeup use) influenced color discrimination. We found no significant difference between participants who reported engaging in color-related practices and those who did not, suggesting that design guidance regarding color discrimination may generalize across viewers regardless of their regular color practice.

The increasing heterogeneity of multi-agent systems poses significant challenges for jointly training a global model across agents. At the same time, cooperative inference between agents has long been recognized as a powerful mechanism for distributed decision making over networks. Motivated by these observations, we propose a collaboration framework for distributed binary classification over multi-agent networks, where a set of independently trained agents, potentially differing in architecture, feature space, or modality, coordinate their actions during test time to form collective predictions. This coordination is achieved by exchanging local decision statistics through a distributed learning protocol. We develop a theoretical and experimental study of this independent training and cooperative inference paradigm, and examine its performance under different communication budgets and distributed learning rules. We establish classification error guarantees under sufficient, finite-round, and finite-precision communication, together with PAC-style generalization bounds. These results capture the influence of model heterogeneity, network topology, combination policy, and communication constraints on prediction accuracy. Taken together with the experimental results, they reveal both the price of independent training and the benefit of collective prediction for the proposed distributed decision making framework with models learned from data.

Accurate segmentation of coronary arteries in X-ray angiography videos is essential for quantitative coronary analysis and image-guided interventions. However, accurate segmentation remains challenging because coronary vessels are thin and exhibit low contrast, while the presence of catheters, guidewires, and complex anatomical background structures can further interfere with vessel delineation. Existing U-Net- and Transformer-based models provide strong baselines, but their shared feature-adaptation pathways may be insufficient for heterogeneous angiographic appearances. In this paper, we propose a prompt-free mixture-of-experts (MoE) feature adapter for binary coronary artery segmentation. Built upon parameter-efficient Vision Transformer adapters, the proposed method uses multiple lightweight experts with input-dependent top-$k$ routing to adaptively refine vessel-related features while limiting active computational cost. Experiments on MOSXAV and external evaluation on XACV show that the proposed method outperforms representative baselines and improves cross-dataset generalisation. These results suggest that MoE-based adapter learning is effective for robust coronary artery segmentation in X-ray angiography videos.

Traditional image similarity metrics such as Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), and the Structural Similarity Index Measure (SSIM) rely on pixel-level comparisons and often fail to capture perceptually meaningful differences between images. In contrast, latent representations learned by deep neural networks encode high-level semantic information that is more closely aligned with human visual perception. This paper proposes a convolutional autoencoder-based framework for quantifying image differences using cosine similarity in latent space. The learned compact embeddings enable robust differentiation between visually distinct images under variations in illumination, pose, and background. Extensive evaluation on dog-cat images and additional cross-domain datasets demonstrates clear class-wise clustering and strong inter-class separability in the latent space, with 98.4% of dog-cat image pairs exhibiting similarity scores below 0.5. Further validation using the TID2013 dataset shows that latent-space distance correlates positively with human Mean Opinion Scores (MOS), demonstrating sensitivity to perceptually relevant image distortions. The proposed approach provides a computationally efficient and semantically grounded alternative to conventional pixel-based similarity metrics, with potential applications in content-based retrieval, perceptual quality assessment, and semantic similarity analysis.

Distinguishing machine-generated text (MGT) from human-written text (HWT) becomes increasingly important due to potential misuse. However, most supervised detectors often degrade out-of-domain (OOD) and require large, diverse training sets. In this work, we analyze the linearity and quality of MGT representations and show that simple linear probes outperform a wide range of detectors while being substantially more sample-efficient. We first show that MGT and HWT latent representations are linearly separable in low-dimensional space, and provide a plausible explanation for this separability through systematic differences in their representation quality. Motivated by these insights, we train two variants of simple linear probes and evaluate them across 4 benchmarks against 16 baselines. Probes consistently improve OOD detection (+11 AUC), requiring solely ${<}100$ samples to reach near-peak performance. We show that this transferability arises because probes recover a shared latent MGT direction that generalizes across diverse settings. Finally, we demonstrate that probing vectors capture a continuous spectrum of “machineness”, highlighting their potential for fine-grained estimation of AI-edited text. Overall, our work provides insights into latent-space differences between MGT and HWT and demonstrates the potential of linear probes as as robust and sample-efficient MGT detectors. We release our code on~\href{https://github.com/gerritq/mgt_probes}{github}.

Peak power limits restrict concurrent tasks and may increase assembly-line cycle time. To the best of our knowledge, this study is the first to minimize cycle time for the simple assembly line balancing problem type 2 (SALBP-2) with a fixed number of workstations and a fixed limit on total instantaneous power. An exact satisfiability (SAT) method finds feasible schedules, searches systematically for shorter cycles, and proves optimality when possible. Two reproducible formulas define looser and tighter power limits for 72 cases based on the standard SALBP library, with Gurobi and CPLEX providing commercial MIP and CP comparisons. Relative to standard SALBP-2 optima, the two limits increase the best-known cycle time by 16.49% and 61.02% on average. The best-performing SAT configurations find a feasible solution for every case, solve more cases to optimality than each commercial solver, and when both prove optimality, are almost always faster under the reported settings.