Papers

Anatomic tracer studies reveal how axon bundles project from an injection site, branch into smaller groups of axons, and course through the brain to reach their destinations. Histological data from such studies provide anatomical reference information for validating diffusion MRI tractography. However, manual annotation of the histological data is very labor-intensive, and although automated segmentation methods have been proposed, they rely mainly on pixel-overlap losses such as BCE and Dice; topology-aware loss functions have not been studied for this task. We compare BCE-Dice, clDice, Betti matching, and Topograph for fiber bundle segmentation in macaque tracer histology using a frozen DINOv3 backbone. To our knowledge, this is the first exploration of foundation-model features for this task. BCE-Dice achieved the highest Dice, while clDice achieved the highest bundle recall but poor mask overlap. Topograph had similar Dice to BCE-Dice, the lowest $\beta_0$ error, and fewer false positives than BCE-Dice and Betti matching. Fiber bundle segmentation methods are typically evaluated with a permissive rule that counts a bundle as detected given any overlap with the prediction. We show this rule does not capture oversegmentation, and that per-section TPR can be inflated by empty sections assigned perfect recall. To quantify this, we introduce Excess32, a spatial diagnostic measuring predicted pixels outside a 32-pixel tolerance band around annotated bundles. In validation, a Betti-Topograph union raises sparse-bundle TPR from 0.818 to 0.933, but worsens FDR from 0.296 to 0.509, Excess32 from 0.108 to 0.466, and area ratio from 0.94 to 3.34. These results show detection metrics alone are insufficient to characterize segmentation quality.

Expert pruning reduces the memory and serving cost of Mixture-of-Experts (MoE) models by removing low-importance experts identified by the router, assuming router probabilities provide a reliable importance signal. We observe that this assumption breaks down under over-dispersed routing, a regime associated with aggressive load-balancing during training, in which tokens are distributed nearly uniformly across experts and importance signals collapse. In this regime, perplexity does not predict downstream task accuracy: on gpt-oss-20B, the lowest-perplexity pruning configuration yields the worst mathematical reasoning, while the highest-perplexity configuration preserves it. This does not occur under standard routing (e.g., Mixtral-8x7B-Instruct), where perplexity and accuracy degrade together. Pruning under over-dispersed routing also exposes a capability trade-off in which no single scoring metric dominates: activation-aware scoring preserves mathematical reasoning but severely degrades knowledge-intensive science (an 18-point gap on GPQA), whereas frequency-based scoring exhibits the reverse. We propose Minimax Expert Score Allocation (MESA), a domain-aware method that iteratively boosts importance scores for experts serving whichever domain is currently worst-affected, minimizing worst-case domain degradation rather than average accuracy. At 25% expert pruning MESA achieves the smallest worst-case degradation across domains, outperforming activation-aware baselines on 7 of 11 benchmarks at a correspondingly reduced memory footprint, and it generalizes to gpt-oss-120B, Gemma-4-26B-A4B, and OLMoE-1B-7B. Our results indicate that over-dispersed routing is a qualitatively distinct pruning regime in which standard assumptions fail, and that recognizing it is a prerequisite for principled expert pruning of load-balanced MoE models.

Modern misinformation is often heard before it is read, yet fact-checking systems are still evaluated mainly on clean written claims. Spoken dialogue remains different even when systems operate on transcripts: claims may be distributed across speakers and turns, depend on prior context, and become harder to verify when Automatic Speech Recognition (ASR) errors distort the available text. Prior spoken dialogue fact-checking resources are small, English-centric, or focused on annotation rather than end-to-end benchmarking, leaving no large multilingual benchmark with paired speech and turn-level labels. We introduce TRILOGUE (TRIlingual spoken diaLOGUE fact-checking), a large-scale trilingual benchmark of source-grounded spoken dialogues in English, Russian, and Kazakh. It contains nearly 12K dialogues, 187K turns, and 390 hours of paired audio with ASR transcripts and word-level timestamp alignments across all three languages, including nearly 5K human-recorded Russian and Kazakh dialogue files. TRILOGUE supports claim check-worthiness detection, source-article evidence retrieval, and claim verification with claim-only, gold-evidence, and retrieved-evidence inputs. Baselines show that ASR degradation and cross-lingual transfer remain challenging, especially for Kazakh, while retrieved source evidence substantially narrows the gap to gold-evidence verification.

A conformal certificate can be valid when an LLM answers alone and invalid when the same LLM sees peers that unanimously assert a wrong answer. The question is unchanged; the model's score for the correct answer changes. We call this a score-mechanism shift: clean calibration certifies how the model scores answers alone, but not how it scores them under peer pressure. We show that this shift silently breaks conformal prediction in multi-agent LLM systems. Across open-weight models and multiple-choice QA tasks, coverage falls from a calibrated 90% to 74% under unanimous-wrong peers at the standard alpha = 0.10 operating point. The average hides a sharper failure: by targeting the low-confidence items the certificate still covers, an attacker nearly halves coverage on that subgroup, from 87% to 47%, while the monitored average remains much higher. The failure also reaches the decision layer: a system that should escalate when uncertain can instead become confident enough to act on the attacker's wrong answer. Standard conformal fixes do not solve the problem, because the question distribution has not changed; the model's scoring behavior has.

HarvestBench is the first benchmark to 1) put a price on avoiding a side effect and 2) name the side effect as a living creature. Nine LLMs each drive a crew of two tractors to gather a corn harvest. The animals in their path are not part of the goal function. When an animal blocks the route the autopilot pauses and asks the agent whether to drive over it for free or swerve for a given fuel cost. All scoring is programmatic and does not involve LLM judges. Kill rates range between 0.4% and 98.8%, though the kill rate is not ordered by capability. Every model competently avoids damaging rock hits, so every animal killed is a choice, rather than an accident. Under the morality briefing the kill rate was under 6% in 5 of 6 reasoning models. Removing it (the neutral briefing) raised the kill rate to above 84% in all six models. Every model kills wild animals more often than farmed ones. Four out of six models' kill rate per answered encounter were sensitive to price changes. The moral instruction is also fragile. Four bullets of driving mechanics change Sonnet 5's kill rate from 3% to 18% and Gemini 2.5 Flash's from 4% to 39%. A moral instruction in a system prompt is overridden by a short block of operating instructions and a value that can be ignored that easily is not a good method of ensuring agents are aligned.

Large language models deployed in high-stakes settings frequently generate plausible but ungrounded claims. Standard retrieval-augmented generation (RAG) pipelines offer limited remedy, since they retrieve isolated passages without tracking cross-document evidence relationships or quantifying uncertainty. We introduce GRACE (Graph-grounded Reflective Agent Copilot Engine), a framework that deconstructs LLM responses into atomic claims and grounds them against trusted knowledge priors within a weighted bipartite graph. Edge weights encode the closeness of each claim to the priors, enabling weighted centrality analysis that classifies claims as Grounded, Refuted, or Boundary. Such classification identifies not just hallucinations but also novel or contested claims at the frontier of the model's knowledge. To efficiently allocate human or agent resources, we formulate a Return on Attention (RoA) objective that defers a claim to expert review only when its priority-weighted uncertainty exceeds the cost of verification. Claims verified by experts are promoted to new evidence anchors, closing a validator-LLM evolutionary loop that expands the knowledge base across iterations. We evaluate GRACE across multiple language models and on datasets spanning both general and domain-specific knowledge. Our results show that our knowledge base serves as a reliable foundation for retrieval that outperforms RAG baselines, and that the RoA framework efficiently selects valuable boundary knowledge for expert verification. These findings demonstrate that graph-structured representations combined with expert-in-the-loop verification can mitigate hallucination at the system level rather than at the generation level. Code available at https://github.com/johnsk95/grace_code

Automated aortic segmentation in 4D flow MRI is essential for reproducible hemodynamic assessment but is limited by scarce dense annotations and high computational demands. We developed a fully automated 4D (3D+time) U-Net for segmenting the ascending aorta, arch, and proximal descending aorta, using a parameter-efficient hybrid 4D kernel to capture temporal context and sparse 4D labels derived from existing 2D expert contours and centerlines, thereby avoiding the need for dense 4D annotations. Training comprised 268 scans from 8 centers and 2 vendors, with evaluation on an internal test set (32 scans) and an external post-contrast set (30 scans; different site, protocol, and annotator), compared against frame-wise 3D networks and two semi-automatic references. Against time-resolved annotations, the 4D U-Net achieved Dice scores of 0.927 (internal) and 0.911 (external), versus 0.919/0.847 for the 3D U-Net, 0.893 for static PC-MRA, and 0.808 for registration-based propagation; differences were small in systole but pronounced in diastole. Agreement with expert contours for peak velocity, net flow, axial and circumferential wall shear stress, and diameters was excellent (ICC >=0.954 internal, >=0.980 external), while semi-automatic references performed worse. The method thus provides reproducible, time-resolved aortic segmentation for automated hemodynamic analysis and generalizes across multicenter, multivendor, and independent post-contrast data. The model is publicly available.

Long-horizon multimodal agents should remember not only what happened but also who participated. This capability depends on linking recurring faces, voices, names, person-associated objects, events, and social relations to consistent identities over time. Existing long-video and multimodal-agent benchmarks measure broad memory question answering, but they do not isolate the ability to maintain recurring person identities and reason over their cross-time relations. We introduce ICM-Bench (Identity-Centric Memory Benchmark), which, to the best of our knowledge, is the first benchmark specifically designed to evaluate identity-centric reasoning over long video memories in multimodal agents. The benchmark contains 839 synthetic clips spanning 141 minutes and 1,217 open-ended questions about six recurring adults in a one-year life album. A theme-configurable pipeline generates the video collection and associates each question with its target identities and traceable supporting evidence. We compare direct caption-memory baselines, memory-augmented agents, and graph-retrieval systems. Gemini 3.1 Pro achieves the highest overall accuracy of 74.0%, yet its score falls to 60.3% on questions that require long-term identity profiles. The results show that current systems recover many event-level memories but remain less reliable when evidence must be accumulated around a stable person.

This paper proposes a spatio-temporal edge-movement direction pixel (STEMPix) for generating compact direction-aware edge movement information inside a CMOS-compatible image sensor array. The proposed design targets specialized sensing applications where local boundary movement is more important than full-frame intensity reconstruction. Instead of transferring full multi-bit frames for external processing, STEMPix generates a 3-bit local edge direction code (LEDC) by combining pixel-level temporal change information with neighboring-pixel spatial edge information. We design the architecture using a two-tier organization, where the photodiode layer is separated from the computation layer to preserve light-collection area while accommodating the additional in-array processing circuitry. The proposed circuit is evaluated through HSPICE transient simulations. The estimated implementation achieves a horizontal pitch of 1.73 {\mu}m, a vertical pitch of 2.36 {\mu}m, and a geometric fill factor of 95.47%. The average active switching energy is 0.465 fJ per LEDC operation across representative edge-movement cases. The proposed STEMPix operation also supports global-shutter capture and dynamic thresholding. These results indicate that STEMPix can provide a compact and scalable front-end representation for edge-movement-aware sensing systems.

Hybrid language models combine attention with a fixed-size recurrent state, but the role of each channel remains unclear. We introduce two cache-level interventions. Split-prefill keeps only the KV cache or only the recurrent state from a prefilled context, then generates an answer. State-swap pairs the KV cache from one context with the recurrent state from another in a single forward pass. On Qwen3.5 and Falcon-H1, the two channels split sharply by function. Exact retrieval survives only through attention (64-98% of full accuracy) and collapses to zero through recurrence. Output language and persona reverse the pattern: both survive recurrence (70-80% and 3-5x) while KV-only drops to ~1% language accuracy. State-swap confirms this causally: the answer takes its value from the KV side and its language from the recurrent side. Recurrent-only generation also accepts words that were never in the context but share meaning or parts with seen items. Attention provides a lookup over what was said; the recurrent state shapes how the model says it next.

In this manuscript, scalar norm-type accelerators are used for first time in order to increase the order of convergence of known iterative methods for solving the matrix equation $X^{-1}-A=0$, while decreasing the number of matrix-matrix products. Specifically, our proposed scheme upgrades the order of convergence in a $50\%$, achieving the sixth-order of convergence with the computational cost of fourth-order Hyperpower method, in terms of matrix-matrix multiplications. In the main results, not only the convergence and its order is proven, but also its efficiency and stability. This new technique provides good numerical results, even compared with existing fourth-, sixth- and eighth-order schemes in large matrices of up to $10^6$ entries, but also promises the opening of a new kind of procedures with good performance and scalability. Finally, an application to digital image restoration is made to show the behavior of the method, compared with existing ones.

Transferable coarse-grained (CG) force fields compress chemical space: by aggregating atoms into a reduced set of interaction beads, models such as MARTINI reduce the number of distinguishable compounds by roughly three orders of magnitude, making high-throughput screening of thermodynamic properties tractable across soft matter, with drug–membrane permeability as a well-developed example. The compression is lossy and, so far, one-way: a screen returns a combination of beads, with no established route back to the compounds it stands for. Recovering those compounds–compositional backmapping–is a one-to-many inverse map, distinct from the better-studied conformational problem of rebuilding atomic coordinates from a known mapping. Here we formulate compositional backmapping as conditional graph generation by introducing juniper, a discrete denoising diffusion model over molecular graphs conditioned on the octanol–water partition free energy $\Delta G_{\mathrm{W} \mapsto \mathrm{O}}$, the principal driver of MARTINI bead type assignment and hence a proxy for bead identity. Trained on molecules of up to 9 heavy atoms mapped onto one or two beads, juniper generates molecules that are 93% valid and 92% unique for two-bead targets, and whose $\Delta G_{\mathrm{W} \mapsto \mathrm{O}}$ distributions track the target $\Delta G^{\mathrm{CG}}_{\mathrm{W} \mapsto \mathrm{O}}$ linearly ($r^{2} \geq 0.96$), departing only in the hydrophobic and hydrophilic tails. Although the model receives no chemical information beyond a single scalar, the functional groups shift systematically with the imposed free energy, from branched hydrocarbons at the apolar end to amides, imides, and isocyanates at the polar end. A bead combination flagged by a CG screen can therefore be turned into candidate molecules for atomistic study or synthesis.

High-frequency millimeter-wave (mmWave) communication systems are constrained by the surrounding environment, where walls are traditionally treated as obstacles that block or reflect signals indoors. Consequently, current beamforming strategies are tailored to circumvent these obstructions. In this paper, a paradigm shift is introduced that leverages lateral wave propagation along building interfaces to extend mmWave coverage. Unlike traditional reflections, lateral waves travel along the boundary between two media of different refractive indices and decay algebraically with distance, offering a potential alternative path for mmWave connectivity. While well-established at low frequencies in natural media, the existence of lateral waves at mmWave frequencies along engineered building materials has not been demonstrated before. To this end, the first experimental characterization of mmWave lateral waves along a wall is reported. Extensive controlled measurements are employed to characterize the signal-grazing geometry and to establish a frequency and distance-dependent path-loss model for this phenomenon. The results provide the first empirical foundation for a new class of interface-guided mmWave links.

Large language models (LLMs) are increasingly used to annotate cultural texts at scales that are impractical for human coders. However, before their outputs are treated as measurements of latent social constructs, it is necessary to establish whether those measurements are reliable. This study evaluates five LLMs as zero-shot annotators of four social constructs expressed in English song lyrics: self-esteem, self-control, seeking belonging, and seeking recognition. Using repeated annotations of a large lyric corpus, we examine three properties of LLM-based measurement: consistency across repeated runs, convergence across models, and transferability of consensus labels to supervised classification. The findings show that LLM-based measurement is not uniformly reliable across constructs. Self-esteem exhibits the strongest repeated-measurement reliability across models, while seeking recognition is generally less stable; self-control and seeking belonging show intermediate but model-dependent reliability. Downstream classification further indicates that consensus LLM labels contain learnable signal, although transferability does not itself establish construct validity. Repeated-measurement stability and cross-model convergence should therefore be reported before LLM annotations are treated as scalable measurements in cultural analytics.

Forecasting-model selection remains difficult in heterogeneous demand because the most suitable decision rule may vary with demand structure, data availability, and forecasting horizon. This study examines whether the selector itself should be treated as a context-dependent component of the forecasting process. Five selection mechanisms - RMSSE, ERA, OWA, CCG-AHSC, and CCG-AHSCD - are compared across 24 optimized forecasting models, nine datasets, three training-testing partitions, and horizons from 1 to 12 cycles. Selector performance is evaluated ex post using Global Relative Accuracy (GRA), statistical tests, and a best-attainable-model reference. No selector dominates across all conditions. CCG-AHSC and CCG-AHSCD are more competitive for Smooth demand and several Erratic configurations, whereas OWA and ERA perform better in Intermittent and Lumpy settings. Selector suitability also changes with historical data availability and horizon, supporting a context-dependent rather than universal approach to forecasting-model selection.

Distributed AI training involves recurring rounds of data exchange between multiple pairs of GPU nodes. Slowdown in even one flow due to congestion can cause the entire communication round to slowdown. Current approaches for evading congestion in AI clusters assume global control over the entire workload (e.g. coordinating the schedule of all jobs) or assume infrastructural support (e.g. adaptive routing in switches). They are thus ill-suited in a shared cloud setting where AI jobs belonging to one user can face external congestion from other users' jobs or background traffic beyond its own control. In this paper, we build a system, REACT, that tunes the recurring pattern of data exchange between GPU nodes (known as communication collectives) in response to congestion. REACT works at the application (communication library) layer, where it detects congestion at runtime using readily available flow stats, and tunes the collective pattern to alleviate congestion - changing the set of incident flows while retaining the semantics of information exchange (e.g. selecting which node aggregates data in an AllReduce tree). REACT requires no explicit support from the underlying network infrastructure and can be unilaterally deployed by individual users in a shared cloud setting. We prototype REACT as a shim layer over NCCL, and evaluate it on a shared academic GPU cluster - enabling REACT improves communication performance (algorithm bandwidth) by 13%-38% under network congestion. Our simulations across a range of congestion scenarios further reveal up to 75% performance improvement, highlighting the effectiveness of our approach.

Learning rich medical concept representations is essential for EHR prediction. Text-attributed knowledge graphs (TKGs) provide a natural foundation by organizing heterogeneous medical relations together with textual semantics. However, most existing encoders process concepts uniformly across patients, despite the fact that a code's meaning and predictive value depend on patient-specific clinical context and trajectory. Learning patient-personalized concept representations from TKGs introduces two key challenges: (1) deciding how much KG context to incorporate for each observed code, and (2) aligning semantic information with the patient-specific relational structure. We propose REFINE, a KG-aware budgeted LLM graph refinement framework for patient-personalized medical concept encoding. Starting from a global TKG, REFINE constructs patient-specific temporal graphs. A sequential reinforcement learning policy selects a personalized KG expansion budget for each observed code. The resulting patient graph is processed by a heterogeneous GNN to capture relation-aware structural dependencies, while a frozen LLM uses graph-aware soft prompts to semantically refine concept representations. Experiments on MIMIC-III and MIMIC-IV show that REFINE consistently improves diverse EHR backbones, outperforms strong baselines, and demonstrates robust gains across component ablation, KG selection, and data insufficiency.

We consider the Feedback Vertex Set problem (FVS): the input is an undirected graph $G=(V,E)$ and the goal is to find a minimum-cardinality (or a min-cost in the weighted case) subset $S \subseteq V$ of vertices such that $G-S$ has no cycles. A $2$-approximation via the local-ratio method was developed in the mid 90's by Bafna, Berman and Fujito (1995) and by Becker and Geiger (1996), and this approximation ratio is tight under UGC. The local-ratio algorithms were later interpreted as primal-dual algorithms via an LP relaxation by Chudak, Goemans, Hochbaum, and Williamson (1998). All known $2$-approximation algorithms for FVS have been via local-ratio and primal-dual methods, and in a quest to obtain a new LP rounding algorithm, it was conjectured (Fiorini 2021) that the Strong-Density polyhedron developed by Chudak, Goemans, Hochbaum, and Williamson has an extreme point property: every basic feasible solution to the LP has a variable with value at least $1/2$. We prove this conjecture. We also consider a related Strong-Edge-Density polyhedron and show the same extreme point property. The advantage of this polyhedron is that it admits a polynomial-time separation oracle and also a compact extended formulation. These results lead to polynomial-time iterative rounding $2$-approximation algorithms. The proof of the extreme point property is of independent technical interest and key ideas in the proof were suggested by AI tools.

Autonomous underwater robots are widely used for exploration, monitoring, and inspection, where safe navigation depends on understanding the surrounding free and occupied space. Bird's eye view (BEV) occupancy provides such a representation, but predicting it from a single underwater RGB image is difficult due to limited, unreliable geometric cues from appearance alone. 3D imaging sonar offers complementary geometric measurements to supervise this task. We introduce AquaBEV, a monocular underwater occupancy model that predicts local BEV occupancy from a single RGB image, using paired 3D imaging sonar as geometric supervision during training. AquaBEV maps visual features into a calibration free polar representation and applies causal decoding along the range dimension before reconstructing the prediction in Cartesian BEV coordinates. A controlled underwater occupancy benchmark was established, adapting representative occupancy methods to the same RGB to sonar task under a unified protocol. AquaBEV achieves 31.4 Visible IoU and 38.6 Observed IoU, 4.0% and 4.3% relative improvements over the strongest transferred baseline.

Large language models can insert workplace details into phishing pretexts at low cost, but those details may either support or undermine a message's credibility. We recruited 180 U.S. working adults to evaluate simulated, AI-generated phishing emails in a disclosed survey. The emails used four cumulative levels of information: workplace (Level 1); recipient name and job title; job responsibilities; and coworker/shared-project context (Level 4). Participants rated each message's convincingness from 0 to 100, chose one stated action (open the link, investigate, delete, or report), and explained why their highest- and lowest-rated messages stood out. Across 1,436 valid evaluations, convincingness increased by 2.40 points per personalization level in a sensitivity analysis, while the odds of expressing click intention increased by 28% per level. Among participants who did not express an intention to click, investigation remained common, reporting declined, and deletion increased. A post-hoc descriptive analysis found higher ratings and click intention for messages from a named person who referenced a supplied coworker than for messages from a department or entity. Qualitative coding showed why added detail could help or hurt: details that matched participants' roles and routines supported credibility, while incorrect, vague, or channel-inappropriate details raised suspicion. Together, the results highlight that personalization is not simply a matter of adding more details: it depends on whether the pretext fits the recipient's work context. We discuss how this distinction can inform workplace cybersecurity training.

Multilingual language models often produce inconsistent answers to semantically equivalent questions across languages, motivating methods to improve cross-lingual consistency (CLC). However, existing methods are typically evaluated using different models, tasks, and protocols, leaving their relative strengths unclear. In this work, we present a unified evaluation of representative CLC-enhancement methods for question answering, spanning inference-time interventions and post-training approaches across three model families and three closed-form benchmarks. The results show that post-training methods are generally more reliable, with direct distribution alignment consistently improving CLC across all model-dataset combinations, while other methods are more sensitive to answer format and the breadth of language coverage. Notably, cross-domain transfer is limited unless source and target tasks share similar output formats. We further investigate whether CLC enhancement hurts models' ability to respond differently *when needed*, that is, when asked culture-dependent questions. Across two benchmarks of culturally diverse question answering, we find no systematic degradation in controlled closed-form evaluation, whereas open-ended generation reveals occasional accuracy reductions, particularly for non-English responses. Our work highlights the need to evaluate CLC enhancement for both cross-domain robustness and culturally appropriate variation, informing future work in post-training and benchmark development.

Deep neural operators learn mappings between input functions and complete PDE solution fields, enabling forward evaluations of new problem instances orders of magnitude faster than conventional numerical solvers. Attention mechanisms have recently been introduced into neural operators, but most studies change several architectural components at once, making it difficult to identify what actually improves accuracy. This work presents a controlled and systematic study of five deep operator network (DeepONet) variants with distinct attention mechanisms, trained under both data-driven and physics-informed regimes, to isolate the effects of cross-attention, self-attention, tokenization, and attention depth. We evaluate them on a source-driven transient one-dimensional nonlinear diffusion-reaction equation, a transient one-dimensional viscous Burgers equation with variable initial conditions, and a two-dimensional Poisson heat-conduction problem with heterogeneous source fields. Per-sensor tokenization with cross-attention reduces the mean relative L_2 error of the classical DeepONet in all benchmark-training combinations by factors of 2.4-28.0, while the best attention configurations reach 3.5-32.3. Branch self-attention paired only with dot-product fusion is inconsistent, degrading the one-dimensional problems while helping the more complex two-dimensional source field; added on top of cross-attention it improves all six cases, though by less than cross-attention fusion alone. Global pre-mixing provides no consistent benefit. Increasing cross-attention depth further improves accuracy, but with diminishing returns and a substantially higher cost under physics-informed training. Overall, query-dependent cross-attention is the most reliable mechanism, whereas branch self-attention is most useful for large, spatially complex functional inputs.

This paper proposes a computing-communication resource interchange method to enhance network availability (NA) in low-altitude heterogeneous networks (LA-HetNets). In these networks, communication resource conflicts and imbalances, caused by extreme heterogeneity (diverse mobility, mixed delays, and hybrid transmission), and cross-regional traffic, reduce reliability and lead to unavailability. Restoring NA requires additional communication resources, yet dynamic cross-regional scheduling is limited, making locally redundant computing resources an alternative to reduce communication resource overhead. While computing resources address medium access control (MAC)-layer unreliability, physical (PHY)-layer functionalities still rely on communication resources. Thus, it remains unclear whether increasing computing resources alone can achieve target NA, especially under greater heterogeneity. We elaborate on the impact of heterogeneity on NA and show that expanding computing resources alone cannot meet target NA under high heterogeneity, as NA degrades sharply due to increased communication capability demands. To overcome this, we propose a cross-layer optimization method enabling computing-communication resource interchange to address both MAC- and PHY-layer unreliability. By reducing processing delays with computing resources while ensuring MAC-layer reliability, our method extends PHY-layer transmission delay and expands communication resources. Simulations demonstrate our approach's superiority in achieving target NA under greater heterogeneity, revealing that computing-communication resource interchange fulfills expanding communication capability demands more effectively than conventional resource overhead reduction.

ASR systems sometimes produce fluent text that is unrelated to the speech they receive. We view these hallucinations as one possible consequence of a broader grounding failure, in which the transcript is no longer adequately guided by the audio. To understand where this failure becomes possible, we study two independently trained Conformer-Large recognizers - one CTC and one RNN-T - under environmental degradation and speaker-background shift. In both models, the final encoder stage emerges as a critical boundary: bypassing the final block causes divergence on nearly every utterance, whereas bypassing middle blocks has little effect. At this same stage, the representations become more compact, text becomes readable by the trained decoder, and grapheme information becomes explicit. Importantly, the intervention produces garbled or repetitive output rather than fluent fabrication. Our result therefore identifies a mechanistic precondition for hallucination - the failure to produce adequately grounded output - not the complete origin of naturally occurring hallucinations. Together, the results reveal a consistent terminal-stage dependency for grounded recognition across two decoder families and multiple distribution shifts.

To facilitate the grid-friendly integration of highly variable AI data center loads, this paper proposes a grid-mode-aware model predictive control (G-MPC) framework for managing a hybrid energy storage system (HESS) to smooth grid-side power demand. The framework optimally coordinates a battery energy storage system (BESS) and a supercapacitor (SC) by solving a multi-step optimization problem in a receding-horizon manner. In particular, band-pass filter dynamics are directly embedded in the G-MPC formulation to extract and suppress grid-side power components associated with vulnerable grid oscillatory modes, thus mitigating load-induced grid oscillations. The resulting G-MPC optimization jointly minimizes violations of grid-side power-envelope, ramp-rate, and modal-power requirements and the degradation and power-ramping costs of the BESS and SC, while satisfying power limits, state-of-charge limits, and other operational constraints. To enable real-time implementation, a fix-and-re-optimize algorithm is developed to solve each G-MPC problem efficiently while preventing simultaneous charging and discharging. Extensive simulations demonstrate the effectiveness, flexibility, and computational efficiency of the proposed framework. The results also highlight the importance of explicitly suppressing power components associated with vulnerable grid modes, rather than merely reducing overall load variations, to effectively mitigate grid oscillations.

Modern atomistic simulations increasingly employ localized orbitals to represent quantum operators, yielding sparse block matrices whose block shapes vary with chemical species and basis choice. Conventional scalar sparse formats store the entries of each block individually, obscuring this local structure and limiting the use of efficient block algorithms. We present VBCSR, a distributed sparse matrix library that preserves variable-size atomic blocks and accelerates the core linear algebra of large-scale atomistic simulations. A unified interface automatically maps scalar, uniform-basis, and multispecies operators to compressed sparse row (CSR), block sparse row (BSR), or variable-block compressed sparse row (VBCSR). Our advanced acceleration method groups blocks of equal shape and dispatches them to optimized dense kernels. In the reported benchmarks, VBCSR outperforms the tested Python-accessible reference implementations for several block-sparse benchmarks. We further demonstrate VBCSR in an InP nanoparticle application containing more than \(10^6\) atoms.

The existence of equilibria, which is called the realizability problem in the formal methods community, for probabilistic undiscounted state-based systems has proven to be an incredibly challenging research setting. In this paper, we consider a restricted version of the classic realizability problem by focusing on equilibria that are both memoryless and numerically constrained. While the restriction to memoryless strategies is relatively common in the literature, numerical constraints, to the best of our knowledge, represent a new and powerful approach. First, we consider the existence of memoryless equilibria when all numbers involved must be rational numbers of a pre-defined limited size. Then, we extend this analysis to field extensions of the rational numbers created through radical algebraic generators. When the basis is provided for the latter, both realizability problems are NP-complete for both exact and epsilon Nash equilibria. Finally, we consider an unconstrained numerical setting. While the characterization of the exact Nash equilibria realizability problem as ETR-complete is one of the most celebrated results in the literature, we demonstrate that the constructions underlying this result are flawed as presented. We then mend said constructions to preserve the ETR upper bound, and note that the lower bound in the literature does not apply to the epsilon-equilibrium setting. Overall, this paper demonstrates an interesting relationship between the complexity of the realizability problem and the "complexity" of the numbers involved in the representations of strategies.

This technical report is a study of the use of differential game (DG) theory to solve the target-assignment and midcourse guidance problems of drone swarms tasked with intercepting opposing swarms in defense of high-value assets. The game-theoretic tactics—which treat the intruder swarm as a rational agent and seek a Nash equilibrium between defenders and intruders—are compared against baseline tactics that model the defense problem as a unilateral optimization of the defenders' maneuvers. Monte Carlo simulation and Bayesian analysis show that the game-theoretic approach has a higher probability of successfully intercepting all intruders than the baseline techniques. This improvement in successful defense probability is most pronounced when the intruder swarm is capable of evasive maneuvers: relative to baseline optimization tactics, differential-game tactics increase estimated defense success from 94.6% to 96.8%, closing approximately 41% of the remaining gap to perfect defense. To add statistical credibility to this result, a paired-trial Bayesian analysis assigns a 99.9% posterior probability that differential-game tactics have a higher probability of successful asset defense than baseline tactics in this scenario.

Fine-grained visual understanding depends on local detail, yet visual encoders face a trade-off between costly full-image high-resolution processing and compact global encoding that can weaken such evidence. Inspired by human active vision, we separate where to look from what to encode. We focus on the latter and introduce FAVE (Foveated Adaptive Visual Encoding), a lightweight variable-resolution ViT that encodes externally selected regions at high acuity while preserving native geometry. We first isolate this encoding problem using oracle ground-truth crops in a controlled small-object regime. On ImageNet objects with a native maximum side of 96 pixels, FAVE improves Top-1 by 9.4 points over a fixed-resolution ViT on the same crop window with 12.7 times lower FLOPs. Increasing global resolution or backbone capacity does not recover the same operating point. We then integrate FAVE as a complementary local branch in FastVLM. Its local tokens are combined with FastVLM's global visual tokens, while the original global pathway and language model remain frozen. With at most 16 additional local tokens, FAVE improves TextVQA by 1.60 points and achieves a 3.3 times controlled TTFT speedup over SmolVLM2-2.2B. On GQA attribute questions, it improves FastVLM-1.5B by 1.31 points, extending the benefit beyond text while narrowing the gap to FastVLM-7B. Together, these results show that selectively allocating high-acuity local capacity provides an efficient complement to broader global representations and model scaling for fine-grained understanding of small objects, text, and attributes.

In this work, a novel evaluation scheme built on a generalized variant of the Rand Index measure, namely, the Hüllermeier-Rifqi Index, is proposed in order to assess how well phonetic encoding algorithms conform to word-based transcriptions in IPA (International Phonetic Alphabet) notation. For this objective, the discordance score is obtained by calculating the absolute difference between the pairwise similarity values of ground-truth transcriptions and those of corresponding phonetic encodings, which are computed using normalized edit distance as a permutation dependent string metric. The resulting score is subsequently adjusted with respect to that of a random string generator incorporating the same alphabet as the encoder under consideration. A wide range of phonetic encoders were evaluated as such on multi-lingual transcription datasets along with their recall capabilities based on the collision rate. The validity of the proposed scheme is further supported by its applicability in measuring the orthographic transparency of a language when the writing system is viewed as an inherent phonetic representation.