Peer Revenue Returns subject to peer review

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Thursday 3 September

Quantum Physics

Quantum Complexity in Nuclear Scattering and Fission Dynamics

Quantum computers promise advantages for simulating strongly correlated quantum many-body systems, like atomic nuclei, that are beyond the reach of classical computers. Realizing this potential requires understanding the quantum complexity structure of the target problem. We investigate the time-evolution of two key indicators of quantum complexity, bipartite entanglement entropy and non-local magic (non-stabilizerness), in nuclear reaction dynamics. We analyze two representative dynamical processes: scattering in a one-dimensional model of strongly interacting fermions governed by the Negele potential, and a realistic simulation of $^{240}$Pu fission within time-dependent Hartree-Fock-Bogoliubov (TDHFB) theory. In the former case, we find that interactions dynamically generate both entanglement and non-local magic, leaving persistent signatures of quantum complexity in the outgoing states. In the latter, we observe that substantial quantum complexity survives in the spatial bipartition of daughter fragments well beyond scission. The presence of significant non-local magic and entanglement in both cases strongly indicate that quantum computers would provide substantial advantages for accurately simulating nuclear reaction dynamics.
Read the full abstract

All authors Saurabh V. Kadam, Antonio Bjelčić, Nicolas Schunck, Kyle Wendt

Machine Learning

B2B Customer Conversion Prediction: A Document Representation, Graph Theory, and CatBoost Driven Methodology

In the one-time selling B2B context, the buying cycle may last months or even years. During the long process, targeting customers that have a high potential to make purchases and recommending personalized campaigns accordingly are important for effective marketing. For this goal, we study the following problems, B2B customer data aggregation, customer feature generation, and prediction of whether a B2B customer would show interest in making a purchase (i.e., prediction of conversion into sales funnel). We propose an algorithm to aggregate individual contacts to the B2B customer level based on multiple keys. For non-standardized keys such as company names, we propose a novel architecture to cluster them in a domain encompassing irregularities such as spelling mistakes and spelling variants. We then define and generate a set of features and apply the CatBoost model for customer conversion prediction. Our framework achieves 91% prediction accuracy. Based on the prediction results and analysis of the model, we then discuss personalized campaign recommendations to foster conversion.
Read the full abstract

All authors Tianqi Wang, Sheikh Shams Azam, Wan Eih Huang, Anton Wiranata, Christopher G. Brinton, Jan P. Allebach

Number Theory

Representation Defects and Cassels Pairings for Congruent Number Curves: Rédei Symbols and Governing Fields

Using BSD results for CM elliptic curves, we relate Qin's quadratic form representation defects to the Cassels pairing on the pure $2$-Selmer group of a congruent number elliptic curve. When the pure $2$-Selmer dimension is even, we prove that the normalized representation defect modulo $2$ is the Pfaffian of the Cassels pairing matrix; the dimension of its radical yields sharper $2$-adic divisibility and information on the $2$-primary Shafarevich–Tate group. For products of primes congruent to $1$ modulo $8$ that are pairwise quadratic residues, we give explicit Cassels pairing matrices for both $E_n$ and $E_{2n}$ and express their entries in terms of quartic and Rédei symbols. For each fixed prime $p$, we construct a governing field of degree $256$ and determine the exact joint distribution of the two Pfaffians by Chebotarev's theorem. In particular, there is a set of primes $q$ of natural density $5/128$ for which both $E_{17q}$ and $E_{34q}$ have rank zero and $2$-primary Shafarevich–Tate group isomorphic to $(\Z/2\Z)^4$.
Read the full abstract

General Relativity and Quantum Cosmology

Non-extensive entropy signatures in compact star

The non-extensive entropy models applied to the black-hole horizon are connected to the generalized Einstein–Hilbert action, $f(R)$, via the Wald entropy formalism. We demonstrate that quark star configurations characterized by the MIT bag model under this modified gravity approach conform to the observational lower and upper bounds established for compact objects linked to HESS J1731$-$347 and GW190814. Furthermore, we assess the effective energy conditions and corresponding speed of sound to evaluate the physical plausibility and stability of the stellar configurations. All non-extensive entropy models studied satisfied these conditions. Our findings show that the exotic compact objects offer a compelling and credible framework for exploring aspects of non-extensive entropy models, such as the Barrow, Tsallis-Cirto, and Rényi formalisms, and vice versa.
Read the full abstract

All authors B. N. Jayawiguna, M. H. Al Ghifari, A. Rohim, A. Sulaksono

Artificial Intelligence

Speculative Macro Commit for Faster Tool-Using Agents

Tool-using LLM agents spend wall-clock time not only on model inference but also in serial action–observation turns, where each tool call, environment transition, and observation can delay subsequent decisions. We introduce Speculative Macro Commit (SMC), a runtime mechanism for a two-tier agent system: a large authoritative actor model produces the official trajectory, while a faster speculative drafter model continuously predicts and executes future action chains on an isolated environment snapshot. SMC mines recurring multi-action skeletons from training traces and stores them in a macro library used to match against action chains predicted by the drafter at runtime. When the actor's next tool call matches the first drafted action, SMC commits the remaining pre-executed draft steps, together with their observations, to the official trajectory. Using Qwen3.5-27B INT4 as the authoritative actor model and Qwen3.5-4B as the speculative drafter model, SMC matches the sequential agent's overall accuracy while reducing latency by 10.23% over the Speculative Actions (SA) baseline and 18.59% over sequential execution on the $\tau^2$-Bench Telecom subset. On AppWorld, SMC reduces wall time by 7.7% over SA baseline and 44.9% over sequential execution, with a small reduction in task completion. Overall, SMC provides a practical way to reuse multi-step speculative execution and reduce agent latency beyond single-step speculative actions. Our code is publicly available \href{https://github.com/zeyuliu1037/speculative-macro-commit}{\textcolor{magenta}{here}}.
Read the full abstract

Computation and Language

SGD-KV: Summarization Guided KV Cache Compression

Large language models (LLMs) face severe memory bottlenecks in long-context inference due to the linearly growing size of key-value (KV) caches. Existing KV cache compression techniques typically rely on simple heuristics, overlooking the distinct functional roles of different attention heads. We present SGD-KV (Summarization-Guided KV Cache Compression), a head-aware framework that leverages a novel chunk-summarization diagnostic task to systematically identify and prioritize attention heads specialized in hierarchical information aggregation. Experiments on Qwen2.5-7B-1M and Qwen3-32B across diverse long-context benchmarks demonstrate that SGD-KV achieves state-of-the-art performance with contexts up to 1M tokens, while reducing KV cache memory usage by up to 75%. Our findings show that strategically allocating the KV cache budget based on the summarization score distribution of attention heads yields a superior efficiency-accuracy trade-off for long-context inference.
Read the full abstract

All authors Zeyu Liu, Woomin Song, Xuandi Fu, Sai Muralidhar Jayanthi, Vivek Govindan, Aram Galstyan, Sravan Babu Bodapati, Srikanth Ronanki

Systems and Control

Trajectory-Based Co-Optimization of Arrival Scheduling and Descent Path Design in the Terminal Maneuvering Area

Terminal arrival scheduling and descent procedure design are studied in two largely separate literatures. Scheduling models reduce each aircraft to a travel time and deliver target landing times, and fuel-efficient descent procedures are designed one aircraft at a time with the schedule taken as given, although both decide where an arriving aircraft absorbs delay before final approach. Existing formulations therefore cannot trade a slower, earlier-configuring descent against level track miles, a schedule that is efficient in time can be expensive in fuel, and autonomous or reduced-crew operations will need a single trajectory plan that ground automation and the flight management system both accept. To close this gap, we propose a four-dimensional terminal arrival scheduler that selects each aircraft's lateral path extension, glideslope-capture distance, and flap-deployment trigger speeds in one decision. We evaluate every candidate idle-thrust descent offline with a wind-aware backward plan and a six-degree-of-freedom forward simulation that returns descent time, fuel burn, minimum track length, and stabilized-approach feasibility, and a rolling-horizon scheduler commits one verified descent and one extension per aircraft under wake-separation constraints and observed entry winds. Two Atlanta terminal airspace case studies quantify the benefit. We show that co-optimized continuous descents save about 15% of fleet fuel below saturation in a free-descent environment and that delayed deceleration saves about 23%, while on the six published Runway 8L arrival flows the charted altitude floors remove 31% of the design lattice and reduce the savings to 9–10% and 20–21%, respectively. We also find that wind moves single-aircraft descent fuel by 34–81% yet explains at most 4% of fleet fuel variance, because aircraft-specific wind effects average out across a scenario.
Read the full abstract

Computer Vision

Counting Animals in Camera-Traps Image Sequences without Count Labels: Winning Solution to the iWildCam 2021 Challenge

Camera traps have become an essential tool for wildlife monitoring, motivating the development of computer vision methods for the automated extraction of information from these data. While most prior work has focused on species identification, many ecological applications also require estimating the number of unique individuals appearing across short image sequences. This task is particularly challenging because camera traps typically acquire bursts of images at approximately one frame per second, creating large temporal discontinuities that may make conventional multi-object tracking methods unreliable, and because manually collecting individual count annotations is prohibitively expensive. In this work, we describe the winning solution to the iWildCam 2021 Challenge, which introduced a benchmark for counting animals at the sequence level under realistic annotation constraints where count annotations are unavailable for training. Our approach, MaxBoxCount, combines a strong species classification pipeline with a simple yet effective counting heuristic based on MegaDetector detections to estimate the number of unique individuals without requiring count annotations. Code is available at https://github.com/alcunha/iwildcam2021ufam.
Read the full abstract

Astrophysics of Galaxies

CLAS+: A large catalog of Changing-Look AGN candidates selected through S-PLUS narrow-band photometry

Changing-look active galactic nuclei (CLAGN) exhibit rapid spectral transitions on timescales of months to years, challenging standard AGN unification models and providing unique insight into accretion-disk variability and obscuration processes. Despite their scientific importance, systematic searches for CLAGN remain limited by the scarcity of multi-epoch spectroscopy and the difficulty of constructing large samples efficiently. We introduce CLAS+, a catalog constructed by comparing synthetic photometry derived from DESI DR1 and SDSS DR17 spectra with contemporaneous narrow-band imaging from the S-PLUS survey. Applying our pipeline to a parent sample of 98,139 quasars, we select high-confidence CLAGN candidates above a threshold of reduced chi-squared $\chi_r^2 > 15$. We identify 3,554 strong CLAGN candidates, corresponding to 3.6% of the parent sample. This selected fraction is conditional on the adopted criterion and should not be interpreted as an intrinsic CLAGN occurrence rate. Cross-matching with existing CLAGN compilations shows that CLAS+ identifies a large population of previously unreported high-priority candidates (3,481). Among the spectroscopic–photometric comparisons with reliable temporal classification, 40% are classified as photometric turn-on, 41% as intermediate, and 19% as turn-off. We also find an apparent redshift dependence in the relative fractions, with photometric turn-on candidates becoming more common at higher redshifts within the CLAS+ selected sample. Therefore, narrow-band photometry provides a powerful and efficient alternative to purely spectroscopic searches for identifying CLAGN candidates at scale. CLAS+ substantially expands the known CLAGN candidate population and provides a valuable target list for future spectroscopic follow-up and time-domain studies with facilities such as Rubin/LSST.
Read the full abstract

All authors L. Nakazono, L. L. Oliveira, R. R. Valença, P. Sánchez-Sáez, G. Oliveira-Schwarz, N. M. Cardoso, A. L. O'Mill, S. Panda, E. Telles, R. Demarco, W. Schoenell, T. Ribeiro, A. Kanaan, C. Mendes de Oliveira

Machine Learning

The 2026 PNPL Competition: Word Classification and Efficient Cross-Subject Generalisation in LibriBrain100

The ambition of the 2025 PNPL competition (Landau et al., 2025) was to launch a multi-year curriculum for non-invasive speech decoding. Designed to progress from foundational tasks toward the linguistic complexity required for a practical brain-computer interface (BCI), it set the stage with speech detection and phoneme classification tasks. Winning submissions reached F1-macro scores of 95.6% and 73.6% on the respective tasks (Elvers et al., 2026), highly significant advances. This success was built on the LibriBrain dataset (Özdogan et al., 2025), the largest within-subject MEG dataset recorded at the time with ${\sim}50$ hours of data for one subject. However, while within-subject scale drives strong decoding performance, a practical BCI must generalise to new users from minutes of data, not hours. The 2026 PNPL competition responds to this challenge with LibriBrain100 (Mantegna et al., 2026), an extended LibriBrain dataset with 32 additional subjects (${\sim}40$ minutes each) plus even more within-subject data (${\sim}80$ hours). Advancing the curriculum of tasks to focus on word classification, two complementary tracks are presented in this competition: the Deep track targets within-subject word classification at scale, aiming at the best possible performance; the Broad track targets cross-subject generalisation, progressively reducing the amount of subject-specific fine-tuning data from ${\sim}40$ to ${\sim}20$ to ${\sim}10$ minutes, a duration that falls within a clinically feasible range and brings us a step closer to a non-invasive BCI capable of restoring communication to people living with profound paralysis.
Read the full abstract

All authors Francesco Mantegna, Gereon Elvers, Dulhan Jayalath, Gilad Landau, Tasha Kim, Miran Özdogan, Luisa Kurth, Teyun Kwon, SungJun Cho, Benjamin Ballyk, Alex Fung, Anna Greer, Pratik Somaiya, Christian Herff, Yorguin Mantilla Ramos, Hamza Abdelhedi, Karim Jerbi, Greg Farquhar, Brendan Shillingford, Mark Woolrich, Oiwi Parker Jones

Software Engineering

Two Truths and A Lie? Benchmarking Off-the-Shelf LLMs for Requirements Quality Assessment: Performance, False Alarms, and Misses

Requirements engineering (RE) governs the quality of everything downstream in systems engineering (SE); defective requirements that survive review cycles propagate into design rework, schedule delays, and cost overruns. Because requirements are often written in natural language, recent advances in generative AI have raised expectations that large language models (LLMs) can absorb requirement quality assessment, a task otherwise slow and human expertise-intensive. Yet empirical evidence on whether LLMs can be trusted to do so remains scarce. This study presents the first benchmarking analysis of off-the-shelf LLM performance for requirement quality evaluation. Against an expert-derived ground truth built on INCOSE quality criteria, we evaluate ten models spanning two families (OpenAI and Anthropic) and five generations each, across one hundred independent runs, two requirement sets, and five sampling temperatures. Four contributions follow. First, we quantify a strongly asymmetric error profile: across all models and runs, the best-performing Anthropic model detects a median of only 47% of expert-identified issues while false-flagging 11%. Second, performance degrades significantly where SE judgment is required, as necessity and correctness issues are almost always missed. Third, generational progress is non-monotonic, so newer models cannot be assumed better. Fourth, this error behavior shifts only modestly and non-monotonically across sampling temperatures, indicating characteristic model deficiencies rather than inherent stochasticity. Off-the-shelf LLMs are therefore not yet trustworthy autonomous evaluators. Findings also warrant caution for Agentic AI developers: orchestrating these LLM modules in specialized architectures risks compounding these deficiencies rather than correcting them. Their defensible near-term role is human-in-the-loop decision support.
Read the full abstract

All authors Jannatul Shefa, Alejandro Salado, Paul Wach, Taylan G. Topcu

Machine Learning

Language-encoded network topology enables large language models to reason about complex networks

Networks describe systems in biology and beyond, from protein interactions and social relationships to power grids and citation records. Reasoning about such systems requires understanding their structure: which elements are central, which connections bridge separate communities, and how it changes when elements are removed. Although large language models (LLMs) excel at natural language, they struggle with such questions when networks are given as edge lists, sentences or measurement tables, because their structural meaning must be inferred. Here we introduce BioGlyph, which compiles network topology into an interpretable and transferable language of structural roles. BioGlyph combines graph partitioning and structural measurements to identify roles such as hubs, community cores and cross-community connectors, and fixed rules to translate them into a universal vocabulary. The representation describes each element through its structural role, supporting evidence and semantic consequences, leaving both the network and the LLM unchanged. Across twenty networks spanning five domains, BioGlyph substantially improves open LLMs' ability to answer structural reasoning questions, outperforming edge-based, numerical and learned representations by up to 26 percentage points in system accuracy. Ablations show that the gain comes from explicitly encoding structural roles in semantically interpretable terms. The gain is more prominent in dense, community-structured networks and diminishes in sparse networks whose topology is more readily inferred from text. In a budding-yeast protein-interaction network, BioGlyph exposes biological organization: cross-community connectors are enriched for essential genes, whereas peripheral proteins are depleted. BioGlyph thus provides an interpretable representation for both language models and scientists to reason about network structure.
Read the full abstract

All authors Ucchwas Talukder Utsha, Sakib Mostafa, James Zou, Md Tauhidul Islam

Systems and Control

Locating Power System Oscillation Sources by Extracting Interharmonics from Synchrophasor Data

Recent studies have shown that oscillating phasors arise from beating waves caused by interharmonics. This finding has led to an interharmonic-based oscillation source-location method. But waveform data needed for the method are less available than PMU synchrophasor data. This paper investigates whether interharmonics can be extracted directly from phasor data for similar applications. The results show that, for certain oscillation phenomena, phasor data can support interharmonic extraction and oscillation source location. Sensitivity studies have identified the requirements, and a cloud-based software tool is developed for PMU-based source location.
Read the full abstract

Statistical Methodology

Randomization Inference for Matched Pairs with Binary Outcomes

We give an exact randomization-based confidence set for the average treatment effect (ATE) in matched-pair studies with a binary outcome, requiring neither monotonicity nor any distributional assumption beyond the within-pair coin flip. At its core is an analytic solution to the worst-case allocation of attributable effects: two binomial-symmetry lemmas identify the pattern hardest to reject as a single boundary corner, so testing null hypotheses needs no integer program and no numerical search. Inverting the test via binary search yields a prediction set for the attributable effect in O(log S) Binomial tail calculations; the Bonferroni proposition of Rigdon and Hudgens (2015) produces the ATE confidence set at the same computational cost. The same corner extends without further machinery to a sensitivity analysis for matched observational studies under Rosenbaum's $\Gamma$-model. A simple formula for the design sensitivity illuminates when an observational study can hope to provide evidence for an effect.
Read the full abstract

Quantum Physics

Revealing Many-Body Phases at finite temperatures through Quantum Coherence Distribution

We introduce a diagnostic framework based on the full probability distribution of quantum coherence to probe single-particle and many-body phases at finite temperature. We first apply this approach to the localization transition in the non-interacting Aubry-André model and then extend the analysis to its interacting counterpart, whose phase diagram is substantially richer. We show that the moments of the coherence distribution provide clear signatures of quantum phase transitions and capture the emergence of many-body phases. More importantly, we find that the entropy of the coherence distribution emerges as an even more powerful indicator, retaining clear signatures of the underlying quantum phases while remaining remarkably robust against increasing thermal fluctuations.
Read the full abstract

All authors Antonio Palamara, Nicola Lo Gullo, Antonello Sindona, Francesco Plastina

Robotics

Long-Horizon Consistent and Interaction-Aware World Models for Multi-Style End-to-End Driving

End-to-end autonomous driving has increasingly adopted world model-based reinforcement learning frameworks to improve learning efficiency through imagined rollouts. However, existing world models suffer from three key limitations: temporal inconsistency in long-horizon imagined rollouts, inadequate modeling of ego-environment interactions, and limited adaptability to diverse driving styles. To address these challenges, we propose StyleDrive, a world-model-based learning framework that jointly enforces long-horizon consistency, explicitly disentangles interactive traffic states, and supports multi-style policy optimization within a unified learning paradigm. First, we introduce a temporal consistency regularization that integrates historical latent states through gated cross-attention, stabilizing long-horizon imagined rollouts and mitigating error accumulation. Second, we design an explicit state disentanglement module that separates ego-relevant from ego-irrelevant interactive states, enabling more interpretable and efficient decision-making in complex traffic scenarios. Third, we enable multi-style driving behaviors through Group Relative Policy Optimization, which replaces per-step reward optimization with trajectory-wise relative advantages, reducing reward variance and supporting diverse driving styles without retraining. We evaluate StyleDrive on the Bench2Drive closed-loop driving benchmark, achieving a driving score of 88.44 (+17.08 over the previous best world model-based method) and a success rate of 66.82 (+16.58). Furthermore, we deploy StyleDrive on a real automated guided vehicle platform and demonstrate promising sim-to-real transfer capability in dynamic driving scenarios.
Read the full abstract

All authors Yuxuan Han, Kunyuan Wu, Liyunong Yang, Zilu Wang, Cansen Jiang, Yi Xiao, Liang Hu

Wednesday 2 September

Software Engineering

Stochastic Modeling of the Mean Time Between Software Failures: A review

A review of several models for the software failure detection stochastic process is per- formed. We focus on how the Mean Time Between Failures is obtained from those models and present a comparative study. The models considered involve some of the well known based on non- homogeneous Poisson processes and a recently proposed model based on contagion. Our main target is to evaluate the models performance when they are applied to modern software projects developed under Agile methodologies. Recently developed Software Engineering practices (TDD, Scrum, etc.), propose an iterative procedure be- tween development, testing and debugging, introducing this way, particular challenges to Software Reliability modeling. As opposed to the non-homogeneous Poisson processes, the failure rate of the contagion model depends not just on time but also on the number of previously detected failures. All the considered models are particular cases of non- homogeneous Birth Processes, a general fundamental theory on how to obtain time to failure and time between failures expectations is also shown.
Read the full abstract

Software Engineering

The Defense Rests: A Courtroom-Inspired Multi-Agent LLM Framework for Automated Safety Case Generation

Safety cases are widely used to demonstrate that mission-critical systems, such as autonomous vehicles and unmanned aerial systems, can operate safely after deployment. However, manually constructing safety cases is complex, time-consuming, and labor-intensive, which calls for the proposal of automated approaches to support their creation. Inspired by the legal nature of safety case construction, we formalize the safety case creation problem through a courtroom analogy, where stakeholders collaboratively propose, examine, criticize, review, improve and validate safety claims and their supporting evidence. In this analogy, the safety analyst acts as a prosecutor by identifying hazards and safety concerns, while the assurance case developer serves as a defense attorney by constructing and substantiating safety arguments. An orchestrator acts as a judge by coordinating stakeholder interactions, and a review board serves as a jury by assessing argument completeness and persuasiveness. Based on this formalization, we propose eTrustWare, a Large Language Model (LLM)-based multi-agent framework for automated safety case generation and review. Our experiments on several safety cases of ML-enabled autonomous driving systems demonstrate the feasibility of coordinated LLM-based agents for generating safety cases.
Read the full abstract

All authors Gerhard Yu, Amanuel Aknaw, Amir Ali Mohaghegh, Yong Song Cao, Alvine B. Belle

Computation and Language

One Question, Four Voices: How Advice for Alzheimer's Caregiving Differs Between Caregivers, Clinicians, and Large Language Models

Online caregiver forums are an important source of emotional and informational support for Alzheimer's disease and related dementias (ADRD). We compare responses to 85 real caregiver questions from ALZConnected across four responder groups: peer caregivers, physicians, ChatGPT, and a retrieval-augmented CareGPT grounded in peer discussions. Using 28 measures spanning linguistic form, emotional support, informational support, and topic scope, we examine how these voices differ when answering the same questions. Peers provided time-anchored narratives and community grounding; clinicians offered concise, advice-forward, clinically bounded guidance; and LLMs produced a formal, verbose, highly structured consultation style. CareGPT showed minimal measurable separation from ChatGPT, suggesting that retrieval plus prompt injection alone was insufficient to shift communicative stance. LLM outputs scored higher than peers on politeness and emotion-support markers and were comparable to clinicians on informativeness and physician-alignment measures, but were less narrative and less contextually situated. Topic modeling showed that LLM responses concentrated on advice-oriented themes, whereas peers covered a broader range, including emotional processing, legal and financial concerns, and daily care. These findings suggest that AI can complement, but not replace, human contributions to ADRD caregiver support and motivate hybrid systems that combine structured AI delivery with the distinct strengths of human voices.
Read the full abstract

All authors Congning Ni, Jinkyung Katie Park, Yang Li, Sarvech Qadir, Sean S. Huang, James S. Powers, Julia A. Hiner, Vania Leung, Patricia Commiskey, Lijun Song, Qingxia Chen, Pamela J. Wisniewski, Bradley A. Malin, Zhijun Yin

Computation and Language

Retrieval-Augmented Generation for Clinical Decision Support in Gait Analysis: A Proof-of-Concept

Large language models (LLMs) demonstrate significant potential for clinical decision support but lack specialized biomechanical knowledge required for accurate gait analysis interpretation. We developed and evaluated a Retrieval-Augmented Generation (RAG) framework to enhance LLM-generated clinical explanations for ankle fracture rehabilitation using ground reaction force data. Using the GaitRec dataset (503 subjects: 211 healthy, 292 ankle fracture patients), we implemented a five-stage pipeline integrating: (1) machine learning classification (91.3% accuracy); (2) Layer-wise Relevance Propagation for explainable feature identification; (3) RAG knowledge base from peer-reviewed biomechanics literature; (4) GPT-5 explanation generation with and without retrieval augmentation; and (5) double-blind evaluation by experienced physical therapists across five clinical criteria. RAG augmentation produced statistically significant improvements in Clini-cal Benefit (mean difference=0.60, p=0.047, {\delta}=0.49) and Completeness (mean difference=0.70, p=0.049, {\delta}=0.49), both demonstrating large effect sizes. Baseline GPT-5 achieved strong performance across all criteria (\textgreater4.0/5.0). This proof-of-concept demonstrates that augmenting general-purpose LLMs with domain-specific knowledge meaningfully enhances actionable, evidence-based clinical guidance—dimensions most critical for rehabilitation decision-making. The framework provides a scalable approach for integrating AI into specialized clinical workflows without costly model retraining.
Read the full abstract

All authors Georgios Giarmatzis, Fabian Horst, Djordje Slijepcevic, Nikolaos Aggeloussis, George Papakostas

Computation and Language

Internal Verdicts Track Evidence: A Template-Controlled Neutrosophic Reading of Epistemic States in Large Language Models via the Jacobian Lens

For twenty-five years, neutrosophic logic has assigned independent degrees of truth (T), indeterminacy (I), and falsity (F) to epistemic states that classical and fuzzy semantics cannot express. Previous applications measured these components on outputs: answers, judgments, expert evaluations. This paper reports, to our knowledge, the first neutrosophic reading of components inside the internal representations of a large language model. Using the recently released Jacobian lens on Qwen3.5-4B, we project intermediate-layer readouts onto lexicons of support, refutation, and hedging, obtaining layer-wise (T, I, F) profiles under conflicting evidence, first-person false belief, and factual control (n = 20 each). A first battery yielded three striking signatures: a surge of refutation mass under conflict, sustained indeterminacy under false belief, and an apparent "verdict collapse" before the output layer. A second, template-controlled battery showed that all three, as initially stated, were largely artifacts of prompt grammar: matched templates with agreeing evidence inflate the same lexicon masses, true beliefs elicit the same hedging, and the model verbalizes its verdict when allowed to generate. What survives the controls is stronger than what died: the neutrosophic verdict balance B = log10(F) - log10(T) tracks evidence polarity under identical templates (about 3 orders of magnitude of separation, Cohen's d = 2.3-2.5, p < 10^-6 in 9 of 9 layers; paired Wilcoxon: median difference +2.4 orders, 19/20 items, p = 3.8x10^-6) and residually discriminates false from true first-person beliefs. We argue that the independence of the neutrosophic components is what keeps the common-mode channel observable, making both the measurement and its self-correction possible, and we distill the two-battery design into a reusable protocol for auditing language model internals with lens-based readouts.
Read the full abstract

Statistical Methodology

Null-Corrected Conditional Coverage in Materials Property Prediction: A Five-Dataset Replication

Uncertainty quantification for materials property prediction increasingly reports conditional coverage, coverage within strata of the predicted value, alongside marginal coverage. The worst-bin statistics used for this are downward-biased in finite samples, a known defect that applied materials work does not generally correct for. We replicate a single-dataset conditional coverage study across five regression datasets (n = 437 to 21,263), seven interval methods and 50 random splits each, correcting the worst-bin statistic against its expected value under exact conditional validity. The null is simulated from realized per-bin test counts and referenced to each method's own marginal coverage, separating marginal miscalibration from uneven coverage. On the smallest dataset the correction runs from 0.061 to 0.096, between a third and two thirds of the shortfalls actually reported, against 0.009 to 0.011 on the two largest. Split conformal prediction's failure survives the correction on four of five datasets; two conclusions of the antecedent study do not. Conformalized quantile regression, characterized there as a negative case, fails on one dataset only. Cross-fitted normalized conformal prediction, grouped there with the successful methods, fails on three of five. Group-conditional (Mondrian) conformal prediction stays below our significance criterion on every dataset, yet exceeds the null on three of five under an octile refinement of the variable it calibrates on. A level-matched ablation locates that uniformity in group-conditional calibration rather than in the elevated quantile level, and removes Mondrian's apparent width cost entirely. Conditional coverage evaluations should report the null distribution of the worst-bin statistic used, the method's realized marginal coverage, and the resolution of the evaluation partition relative to any partition the method calibrates on.
Read the full abstract

Human-Computer Interaction

VIVID: Bringing Tool-Augmented Agents into Enterprises

Tool-augmented large language model agents connect natural-language requests to enterprise services, yet their operational viability depends on reliable execution, organizational control, and system responsiveness. We present VIVID, a deployed tourism-management agent system designed to study these requirements in a real enterprise environment. On an audited dataset of 283 queries, its Qwen3-8B configuration achieves 59.06% strict end-to-end success across the evaluated decision stages, indicating useful but bounded autonomous capability. Stage-wise KV-cache eviction reduces mean latency by 34.4%, while targeted optimization reduces the mean latency of the five slowest measured business tools by 81.1–92.3%. In a three-month deployment serving more than 1,500 personnel, VIVID reduced average per-order processing time from approximately ten minutes to under two minutes, with a final-task error rate below 1% after human-in-the-loop control. These results show that enterprise agent viability requires system-level consideration beyond isolated tool-calling accuracy.
Read the full abstract

All authors Yichu Fang, Ziqiang Zhang, Qiyang Shi, Zilong Wang, Lin Shen, Zhikang Ou, Jing Ma, Jiayuan Chen, Lei Sun, Dai Cheng, Yi Qiao, Yu He, Wei Zhang, Xiaoyu Shen

Human-Computer Interaction

Cellery: Gradual, Composable Structure for Spreadsheets

Structure in spreadsheets is typically restricted to ranges of cells and flat tables. However, replacing the open-ended two-dimensional grid with a more structured alternative risks limiting users' freedom and impeding exploratory workflows. How can we enable expressive structure in spreadsheets while preserving their flexibility? We propose a model that allows users to gradually add structure on top of the traditional grid. To identify common structural patterns, we conducted a content analysis of 30 real-world spreadsheets. Based on our findings, we developed Cellery, a novel spreadsheet system that makes these structures explicit. In Cellery, users gradually define the structure of their spreadsheets using named groupings of cells, composing together four structural primitives: Atoms, Lists, Tables, and Blocks. Once defined, users can directly manipulate structures, dynamically reference them in formulas, and incorporate them in reusable abstractions. We evaluate Cellery with an expressiveness study reproducing spreadsheets from our sample, demonstrating that it is capable of representing types of structure that traditional spreadsheets cannot. In a preliminary user study of five participants, we find that they successfully use Cellery to design spreadsheets with explicit structure, and we uncover opportunities for further work on the formula language.
Read the full abstract

All authors Jacob Yim, Sathwika Peechara, Brian Hempel, Michael Coblenz

Computational Physics

pcNN-SN: a physics-enhanced predictor-corrector neural network algorithm coupled with the discrete ordinates method for neutron transport equations

Super-resolution machine learning methods hold promise for addressing the high computational costs associated with high-dimensional features in traditional transport simulations. This paper proposes a physics-enhanced neural network coupling algorithm for discrete ordinates (SN) transport calculations. Based on high- and low-accuracy data, we develop a predictor-corrector neural network (pcNN) to predict correction factors for coarse-grid solutions. Based on transport equations, local physical features containing subgrid-scale information are selected as network inputs. Dimensionless transformation of physical quantities reduces the input space dimensionality and narrows the range of parameter distributions, thereby facilitating network training. Physical constraints are incorporated into the network model, and the corrected solution retains its original conservation properties. pcNN and SN transport sweeps are internally coupled, and numerical solutions are corrected online. Numerical results demonstrate that this algorithm can effectively improve coarse-grid solutions to achieve a level of accuracy comparable to that of fine-grid results. The physics-based network design enables the model to demonstrate a certain degree of reusability under different computational conditions. Compared to mesh refinement, the pcNN-SN algorithm reduces the computation time for the three-dimensional test problems by more than 60%.
Read the full abstract

All authors Cong Liu, Junxia Wei, Shuai Wang, Heng Yong

Probability

Approximation of Probability Geometries via Convex Families of Coordinate Charts

Within probability-coordinate geometry, a continuous distribution function restricted to an interval can serve as a probability coordinate chart. Averaging in these coordinates defines probability barycenters and associated Kolmogorov moments. Because the class of admissible charts is infinite dimensional, this paper develops a finite-dimensional approximation theory. Relative to a fixed base chart, every chart on the same interval is represented uniquely by a strictly increasing deformation of the unit interval. Monotone piecewise-linear deformations are shown to be exact convex combinations of elementary ramp functions, thereby parametrising probability geometries by allocation vectors in a simplex. Such geometries are dense in the admissible class. We study two constructive schemes: ramp interpolation and Bernstein approximation. Bernstein approximants also admit a ramp-dictionary representation with smooth beta atoms derived from uniform order statistics, while automatically preserving admissibility. Uniform approximation of deformation maps yields explicit stability bounds for probability barycenters and initial Kolmogorov moments. For deformed intrinsic charts, the barycenter is a quantile whose level depends only on the deformation, connecting the construction with chart-based quantile estimation. Deterministic numerical illustrations demonstrate both approximation schemes and their induced barycenter errors.
Read the full abstract

Databases

BuildGraph: A Synthetic Multi-Archetype Building Knowledge Graph Dataset

Semantic querying of building knowledge graphs (KGs) underpins the integration of artificial intelligence into building operations, from natural-language access to cross-building analytics, but such KGs are rarely public owing to proprietary, security, and cost barriers. BuildGraph is a synthetic building KG dataset of 120 buildings in the Brick Schema ontology, grounded in U.S. Department of Energy prototype models and sensor-placement patterns from real buildings. It spans eight commercial building types across three ASHRAE energy-code vintages, with five realizations per archetype varying URI naming, sensor-attachment predicates, and topology. A 75-query SPARQL benchmark confirms structural completeness: BuildGraph reaches 91.5% Query Answerability Rate versus 35.8% for 59 real-world Brick files. Independently, it reproduces real buildings' sensor-type proportions on their shared vocabulary (cosine 0.86-0.94), evidence of realistic instrumentation where measurable. A downstream text-to-SPARQL experiment with Gemma 4 (26B) reaches 27.3% Row-Matching F1 (+17.5 pp over zero-shot) on 12 held-out buildings. BuildGraph gives facility managers and digital-twin developers a testbed for portable analytics and natural-language interfaces, and its dataset, generator, and benchmark are openly available at https://github.com/humanbuildingsynergy/BuildGraph.
Read the full abstract

Databases

Guarded Commits: Transactional Human Approvals for LLM Workflows

LLM workflows often require human approval before an irreversible external action. Most systems keep that approval outside the workflow, as an interface click or an audit entry. The workflow therefore lacks a commit-time check that every risky path reached an approval gate. Its logs may not preserve the reviewed evidence or the conditions for reusing an earlier decision. We present a guarded-commit design that makes human approval part of workflow state. Before the external action runs, a decision source appends a resolution record and its referenced evidence to a ledger. A credential-confined commit adapter then checks that record against the artifact, policy version, and executed path. Our evidence is limited to trace reconstruction. A validator test on synthetic acyclic workflow plans accepts unfaulted plans and rejects plans with each injected fault: a missing gate, the wrong gate type, or incomplete path coverage. Across three public workloads totaling 271,035 traces, replay reproduces recorded artifact hashes when present. Avoided reviews and disagreement with recorded decisions vary by workload. The resulting record supports audit and replay under the policy in force when the resolution was recorded.
Read the full abstract

All authors Laurent Bindschaedler, Ferdi Kossmann, Chunwei Liu, Jason Mohoney

Mathematical Physics

Closed-form evaluation of the cyclic four-electron atomic correlation integral

For positive exponents $\alpha_1,\ldots,\alpha_4$ we evaluate the cyclic four-electron Slater integral, the topology recorded as case V in Bonham's 1965 classification, in a finite real expression of weight at most three. The general answer is an explicit sum of eighteen blocks involving only logarithms, dilogarithms and trilogarithms of rational functions of the exponents; all logarithm arguments are positive and all nonconstant polylogarithm arguments may be placed in $(-1,1)$, so no radical, generalized polylogarithm, contour prescription or unevaluated limit is required. At equal exponents the value collapses to $I(\lambda,\lambda,\lambda,\lambda)=(4\pi)^4\lambda^{-4}[\mathrm{Li}_3(1/4)/2-7\zeta(3)/24+\pi^2\log 2/12-2\log^3 2/3]$, a weight-three period over $\mathbb{Q}$ ramified only at 2 and 3. Three independently implemented numerical routes agree at four exponent tuples to at least 32 decimal places, an exact-dyadic evaluation of the classical formula gives eight 100-place certificates, and three published four-electron calibration values are reproduced to every printed digit. We also prove the small- and large-exponent limit laws with closed-form coefficients, dihedral symmetry, and a strict rearrangement law for the three inequivalent cyclic placements of four exponents. The formula applies to the four-cycle and does not evaluate the two chorded topologies.
Read the full abstract

Machine Learning

Integrating Fairness and Explainability in a Multiple Instance Reinforcement Learning System

Predicting student performance from educational interaction data requires models that are both accurate and sufficiently transparent to support meaningful intervention, while demographic information introduces an additional risk of unfair predictions. This study investigates a multi-objective framework that combines reinforcement learning-based multiple instance learning (RL-MIL), adversarial debiasing, and preference-conditioned hypernetworks for student-at-risk prediction. MIL represents each student as a bag of weakly labeled interactions, while an RL agent selects informative instances for downstream classification. Two hypernetwork variants are evaluated to determine whether a user-defined preference scalar can continuously control the trade-off between predictive performance and Equalized Odds. The underlying RL-MIL baseline achieves strong classification performance, but both hypernetwork extensions exhibit mode collapse: changing the preference weight produces little systematic movement along the intended fairness-performance frontier. The failure is associated with objective dominance, weak gradient propagation through the conditioning mechanism, and interactions between dynamically generated parameters. The results show that fairness objectives can be incorporated into an interpretable RL-MIL pipeline, but preference conditioning alone does not guarantee controllable multi-objective behavior. Robust fair RL-MIL therefore requires explicit mechanisms for gradient balancing, objective separation, and stability analysis.
Read the full abstract