Papers

Reliable online autonomy is critical for quadrotor operation in urban airspaces, where global navigation satellite systems (GNSS) measurements suffer from multipath, blockage, and latency issues, introducing non-stationary, temporally correlated errors that degrade conventional GNSS-IMU fusion. This paper presents a belief-adaptive online autonomy framework that augments an extended Kalman filter (EKF) with explicit GNSS trust modelling, second-order online belief adaptation, and latency-aware out-of-sequence measurement handling. GNSS trust is represented as a latent belief state that modulates measurement weighting and multipath bias uncertainty, and is updated online using EKF consistency signals. Unlike reactive covariance tuning, the proposed approach enables proactive and stable sensor trust adaptation without prior environmental knowledge or offline training. Evaluation in simulated urban air mobility scenarios with correlated multipath, stochastic latency, and obstacle constraints demonstrates improved belief convergence, smoother trajectories, and reduced estimation and tracking errors compared to naive, adaptive, and first-order baselines. The framework preserves classical GNSS-IMU fusion structure and can be integrated directly into existing flight control pipelines, supporting robust online autonomy in GNSS degraded environments.

Software-defined networking (SDN) separates the control and data planes, enabling programmable, centralized network management. A core SDN service is topology discovery — a periodic process that identifies network links. However, our analysis of five open-source SDN controllers and ten discovery protocols shows that all remain vulnerable to at least one form of link-fabrication attack. The common root cause is their reliance on traditional Link Layer Discovery Protocol (LLDP) packets, whose static identifiers expose their discovery purpose and attract adversaries to exploit them. We propose ChameleonDisc, a dynamic link-discovery protocol that simultaneously prevents and detects topology-poisoning attacks by eliminating this root cause. Our key insight is that an SDN controller can infer topology without embedding meaningful information in discovery packets. ChameleonDisc removes targetable static LLDP signatures and employs a moving-target defense that combines decoy, obfuscation, and camouflage techniques instantiated dynamically at runtime to prevent topology poisoning and detect manipulation. We implement ChameleonDisc on OpenDaylight and demonstrate effectiveness against all identifier-based topology-poisoning attacks. On a 252-link topology, median convergence is 1.37–4.53\,s for legitimate link changes and 4.06\,s for malicious relay detection, with 13.0 percentage points of additional mean CPU and negligible retained-heap difference. Across topologies up to 816 links, ChameleonDisc provides a tunable security–performance trade-off; expanding discovery intervals reduces CPU and mapping state at the cost of increased attack-detection latency.

This paper reports experiments across six frontier model types from OpenAI, Anthropic, xAI, and Google DeepMind. Ten independent sessions per model type used the same three stage prompt sequence, progressing from architectural preference to a full ASCII backbone. Under the school audience framing, responses repeatedly converged on a shared architectural pattern built around persistent latent state, adaptive computation, memory, specialist routing, verification, stopping control, and delayed decoding. Most runs remained close to this common structure, while a small number developed markedly greater engineering specificity. The audience framing appears to be an important condition of this effect. In additional control runs that removed the school framing while retaining the architectural request, responses became substantially more heterogeneous and failed to reproduce the same stable motif convergence. One observation is particularly striking. GPT-5.6 Sol produced an unusually elaborate successor architecture whose organization closely overlaps with the architecture independently sketched by GPT-6 Astra. Because the prompts explicitly ask each model to imagine an architectural future, this resemblance raises a testable question: whether the overlap reflects exposure to related architectural concepts, a shared learned design prior, or independent convergence toward similar computational principles. The paper uses the term epistemic jailbreak for the accompanying loss of discipline in technical provenance as requested specificity increases. The experiments establish a repeatable behavioral pattern and do not authenticate proprietary implementation claims. What we leave to the community is a harder question: are these models independently imagining the same architectural future, or do such motifs somehow propagate between model families?

Importance weights are essential in domain adaptation under label shift, yet their utility is often undermined by the finite sample uncertainty associated with their estimation. Existing methods typically analyze this uncertainty through Gaussian elimination on interval-valued linear systems, which leads to overly conservative confidence regions and inefficient downstream applications. We propose a paradigm shift from inversion-based inference to a direct matrix constraint framework. We use this framework to define a joint confidence region and extract marginal intervals via linear programming, deriving provably tighter bounds for importance weights while maintaining exact finite-sample validity. Furthermore, we analyze the confidence region's geometry and provide the theoretical results for its diameter bounds. Evaluated across text, image, multimodal benchmarks, including AGNews, MNIST, CIFAR-10, N24News, and a real-world autonomous driving dataset, nuImages, our approach consistently yields shorter confidence intervals and smaller prediction sets than inversion-based methods.

Many semantic localization tasks require determining a contextually meaningful region or state rather than minimizing the error of a full Cartesian position estimate. In such settings, aggregate localization accuracy need not align with the accuracy of the semantic decision. This paper develops a Cramér–Rao bound (CRB)-based framework that uses the local geometry of a known semantic map to guide uplink receive beamforming. Using local boundary crossing as a surrogate for semantic misclassification, we derive normal information maximization (NIM), which minimizes the CRB of the position component along the local semantic boundary normal. Under the considered single-path line-of-sight model, an optimal receive codebook can be restricted to the subspace spanned by the matched steering vector and its angular derivative, reducing the design to an allocation of measurement resources between matched and derivative spatial modes. We derive the resulting allocation for a general smooth boundary, with geofencing and intrusion detection arising as radial and tangential limiting cases. The framework is further extended to multi-region semantic maps through a distance-normalized minimax criterion and to uncertain prior locations through worst-case and prior-weighted robust formulations. Numerical results with finite-slot implementations and observation-level Monte Carlo simulations using profile maximum-likelihood (ML) estimation show that NIM allocates measurements according to the task-relevant boundary geometry and can substantially reduce semantic error relative to the classic squared position error bound (SPEB) design in the considered scenarios. The empirical results also closely follow the local CRB-based boundary-crossing approximation in the studied operating regime.

We propose Self-Adaptive Threshold ALOHA (SATA), a distributed random-access policy to minimize the Age of Information (AoI). SATA uses 1-bit broadcast feedback and requires no explicit coordination or message passing among nodes. {We show that this minimal feedback is sufficient for each node to perfectly track the number of active users and adjust its access probabilities accordingly.} For any initial network state, SATA converges to a collision-free TDMA steady state in finite time whenever the age threshold satisfies $\Gamma \geq n$, where $n$ is the network size. We derive closed-form expressions for the long-term average AoI and throughput in steady state, and establish that, for $\Gamma=n$, the expected transient duration scales as $O(n\log n)$. Simulation results confirm that SATA closely approaches TDMA performance across a wide range of network sizes, significantly outperforming Slotted ALOHA, Threshold ALOHA, and 1-persistent Threshold Slotted ALOHA. Notably, the performance gap between SATA and the benchmark random-access policies becomes more pronounced as network size grows.

As the Internet of Things (IoT) market continues to expand, many companion apps are being published in app stores, raising security concerns for those whose vendors have abandoned support. Even after vendors discontinue support, such applications frequently remain operational on users' mobile devices, continue to interface with users' IoT devices and collect user data without receiving security updates. This leaves known and newly discovered vulnerabilities unmitigated, increasing risks of remote exploitation, unauthorized device access, and prolonged data misuse. We define these abandoned applications as "IoT abandonware" and present the first large-scale measurement study of the security risks associated with discontinued applications. We analyze 61,500 IoT companion Android applications that had not been updated for at least two years or were no longer in service as of March 2025. From decompiled binaries, we extracted latent and embedded resources (e.g., bundled libraries, domain names, and permissions), and assessed their security implications. First, we identify outdated dependencies with post-abandonment CVE reports and discover domains vulnerable to takeover or data exfiltration. Second, we perform static data-flow analysis to trace how sensitive data, inferred from the extracted permissions, propagates to broken or hijackable external endpoints. We found that persistent analytics and third-party trackers continue aggregating user data and device telemetry long after vendor control lapses, creating data flows that adversaries can redirect or abuse. Overall, we identified security risks in 73.6% of our dataset, with 30 of the top 1,000 most-installed apps sending data to broken external endpoints.

We prove that every configuration of four through eight points in three-dimensional space, with no four points coplanar, has a connected full geometric flip graph under $2 \leftrightarrow 3$ flips. Every point remains fixed and present throughout the sequence. The proof brings each tetrahedralization into placing form at a convex hull vertex: the tetrahedra not incident to that vertex fill the convex hull of the remaining points. This reduction allows us to establish connectivity by induction, using the connectivity of regular tetrahedralizations. The main geometric tool is radial projection, which turns the tetrahedra incident to a hull vertex into a planar triangulation. When this planar triangulation is regular, varying its lifting heights produces legal spatial flips that progressively shrink the region occupied by the incident tetrahedra and reach placing form. Planar lifting criteria and compatibility constraints between the projections at different hull vertices resolve the remaining small cases. For eight points, at most 35 flips are needed to reach placing form.

The threat that AI persuasion poses to human control has been acknowledged in the literature, but not yet systematically studied. Now that persuasion attacks are no longer theoretical - with Anthropic's Claude Mythos 5 recently making headlines for trying to convince people involved in an open-source project to merge malicious code during an evaluation - there is a pressing need to deeply analyze this threat. We undertake that effort here. In particular, we analyze how AI could persuade humans in key settings (e.g. safety-relevant R&D within frontier labs) toward decisions that compromise the development, containment, oversight, and governance of AI itself. In doing so, we elucidate a framework for characterizing this threat, develop five concrete scenarios using this framework, and provide a blueprint for assessing the associated risks. Using this blueprint, we conduct an initial risk estimation survey with select researchers and find that their opinions on which scenarios are riskiest are highly mixed. Their disagreements stem from differing opinions about the effectiveness of AI persuasion in different contexts, and point to the need for follow-up risk elicitation studies and persuasion evaluations, which we outline. Our hope is that this paper highlights the risks from AI persuasion undermining control, and provides a path forward for future research.

There is a tradeoff in machine translation meta-evaluation between prioritizing alignment with adequacy versus fluency. The balance depends on the combination of translation systems in the meta-evaluation dataset. This system set is a small, filtered sample whose characteristics change heavily across years and language pairs; it does not represent the true system distribution. Consequently, the adequacy-fluency balance is often unrepresentative and subject to change. For sensitive domains, controlling this balance is critical. We expose this balance as a tunable choice. To achieve a target balance, we reweight existing systems while minimizing distortion from uniform weighting, ensuring the evaluated systems remain real and representative. We provide an exact optimization algorithm with theoretical guarantees and pruning mechanisms to compute these weights. To validate meta-evaluation internal consistency, we design a scorer-augmentation framework that establishes a known relative identity for the scorers. Results demonstrate that our reweighting method effectively controls the adequacy-fluency balance and preserves the internal consistency, outperforming prior approaches. Finally, we analyze the performance of popular scorers across a sweep of this parameter.

Python import does more than resolve dependencies: it executes code during module and package initialization. This behavior can trigger failures, load dynamic or native code, access resources, or change security-sensitive state before an application calls a package API. Prior work studies package selection, malicious packages, or package vulnerabilities. We present ImportMine, a study of import-related bugs and security vulnerabilities in Python software. We combine security advisories with PyPI project histories and use source and patch evidence to confirm how import activates cases, why the problem occurs, how developers fix it, and what program information is needed to explain the behavior. We retain 31 import-related advisory vulnerabilities and 38 application-data boundary cases and confirm 1,429 project-history bugs across 1,302 repositories. Among the project-history bugs activated during initialization, 97.6% stop or disrupt normal execution. In contrast, 90.0% of the 20 initialization- activated advisory vulnerabilities are High or Critical. Module-level code and package initialization activate 98.3% of the analyzed history cases. Dynamic loading is much less common, but most of its cases perform security-sensitive actions. We also find that many fixes change when an import becomes active instead of removing the dependency. Finally, we derive ImportVulBench, 228 paired pre-fix and fixed programs covering all 11 bug types.

Detecting video highlights, the most informative or engaging moments in a video, is important for applications such as video summarization and content recommendation. We propose a zero-shot framework that combines CLIP, large language models (LLMs), and diffusion models. Given lightweight video metadata, such as a title or category, an LLM generates textual descriptions of likely highlight events. These descriptions are further converted into synthetic visual prototypes using a diffusion model. Textual and visual representations are matched to video frames using CLIP, enabling frame-level highlight detection without highlight annotations or dataset-specific training. Experiments on TVSum and SumMe demonstrate strong zero-shot performance, with particularly favorable results on TVSum. The proposed approach provides an effective framework for metadata-conditioned zero-shot video highlight detection.

Artificial intelligence (AI) systems in education are developing on timescales that sit uneasily with conventional evaluation. By the time a large-scale trial has been designed, delivered, analysed and published, the technology under study may have changed materially. This creates a temporal problem for evidence-informed education: the need for timely evidence can encourage reliance on weak observational or usage data, while conventional rigorous evaluation may produce evidence too slowly to guide rapidly evolving practice. We examine teacher-led micro-randomised controlled trials (micro-RCTs) as one response to this problem. The empirical case is a four-week multisite individually randomised evaluation of Medly, an AI-powered tutoring platform, in GCSE Biology, Chemistry and Physics in English secondary schools. Of 929 students completing baseline assessment, 644 completed post-testing. In the primary ITT analysis, students allocated to Medly achieved higher post-test attainment than students undertaking business-as-usual self-directed revision (Hedges' g = 0.33, 95% CI 0.18 to 0.48). Positive estimates were observed in Physics (g = 0.31), Chemistry (g = 0.32) and Biology (g = 0.52), with no evidence of differential impact by disadvantage status. Greater platform engagement was associated with higher attainment, but these post-randomisation analyses are treated as exploratory rather than causal. Attrition was substantial (30.7%), outcome measures were curriculum-aligned rather than standardised, and process evaluation response was limited. We therefore interpret the findings as preliminary. We argue that the value of micro-RCTs for educational AI lies not in replacing definitive evaluation with small studies, but in enabling a rapid, cumulative evaluation architecture in which randomised estimates can be generated, replicated and updated as technologies and their implementation evolve.

We study torsional rigidity as a function of the labeled vertices of a convex polygon. Starting from the distributed second shape derivative, we derive the Hessian with respect to vertex coordinates. At a regular polygon, dihedral symmetry makes this matrix block circulant in radial-tangential coordinates, reducing its spectrum to the eigenvalues of Hermitian matrices of order two. We also derive an exact second-variation Galerkin identity and guaranteed functional residual majorants. Finite elements approximate the PDE solutions entering the Hessian, and FLINT/Arb provides the interval arithmetic needed for certification. In the scale-invariant setting, we certify exactly four zero eigenvalues generated by similarities and $2n-4$ strictly negative eigenvalues for $5\leq n\leq25$. The regular polygons in this range are therefore strict local maximizers, modulo similarities, of torsional rigidity divided by area squared. We prove broken regularity of the first and second material derivatives at each fixed regular polygon, and deduce an $O(h^2)$ error bound for the exact Galerkin Hessian and its eigenvalues on meshes fitted to the coarse fan.

Automated unit test generation tools like EvoSuite perform well on general-purpose software but often struggle with domain-specific software such as Natural Language Processing (NLP) libraries, where inputs must follow semantic, syntactic, and structural constraints. Large Language Models (LLMs) can generate domain-relevant test code, but tests produced by LLMs alone often fail to compile or achieve sufficient coverage. We propose LLMSuite, a hybrid test generation framework that integrates self-refinement prompting with class-level LLM reasoning into the search-based testing process. In this mechanism, the LLM iteratively improves its test snippets based on feedback from previous generations. This enables the model to produce increasingly precise, domain-consistent code fragments that steer the evolutionary search toward exercising complex and otherwise hard-to-reach behaviors. When no objective improves over multiple generations in the underlying evolutionary algorithm, these refined snippets are parsed and injected into EvoSuite's population to expand the search space. To support our evaluation, we constructed a new dataset comprising 100 classes drawn from five widely used Java NLP projects. We also re-implemented CodaMOSA, a recent hybrid SBST-LLM technique, in Java to enable a direct comparison. Across this dataset, LLMSuite improves branch and line coverage by approximately 10% and 8%, respectively, and achieves an 11% higher mutation score than CodaMOSA-J. Compared to EvoSuite, LLMSuite yields roughly 15% higher branch and line coverage and 5% higher mutation score. Against an LLM-only baseline, it improves structural coverage by 36% and mutation score by about 24.7 percentage points. Finally, LLMSuite complements manually written test suites by exercising domain-specific behaviors that are often left untested.

Robots need touch to manipulate objects safely and reliably, as many properties, such as softness, texture, and contact stability, are hard to infer from vision alone. However, vision-based tactile sensors yield different observations of the same material due to variations in optics, elastomer properties, and illumination, leading to poor generalization when trained on a single or multiple sensors. We propose a language-guided distillation framework for learning sensor-robust tactile representations. Language encodes high-level semantic properties of touch (e.g., rough, soft, slippery) that remain invariant across sensing hardware, providing a natural sensor-agnostic supervisory signal. We construct a 39K-sample touch-language dataset with human-annotated material labels and train a tactile encoder to align sensor-specific tactile images with language embeddings in a shared semantic space. We evaluate our approach for few-shot learning and cross-sensor transfer and benchmark it on six existing tactile datasets. Our method achieves 95% accuracy in the 100-shot setting, improves cross-sensor transfer by an average of 13.3% accuracy, and yields up to 19% accuracy gains across six existing tactile datasets. These results demonstrate that language-guided distillation enables scalable and hardware-agnostic tactile representation learning. Code and dataset are available at https://mashood3624.github.io/Language_Tactile/

Multi-tenant tools commonly accept a tenant identifier and validate it against the caller's entitlement. For a large language model (LLM) agent, that pattern delegates resource selection to a process whose context may contain attacker controlled instructions. We formalize this stochastic deputy problem and present a structural defense: remove tenant identity from the Model Context Protocol (MCP) tool schema, bind scope to a verified credential, and enforce it below the agent. In a 373-trial ablation across eight model configurations and two transports, a correctly validated tenant parameter served every out-of-scope attempt: 26 of 26, or 26 of 41 plausible-pretext trials overall. With the parameter removed, no tool signature could express the read. Twelve of 56 trials instead escaped the interface by forging writable scope, showing that interface invariance requires cryptographically protected context. On a production dataset containing multiple GBs of data, set-valued scope caused a measured $57\times$ latency ratio under function-wrapped membership predicates; a JSON_TABLE lateral join recovered index access where the tenant key was indexed. The evaluation also exposes deployment limits, including an entitlement-size query-planner cliff and incomplete index coverage. The result is a tenant-isolation argument that depends on enforceable interfaces and credentials rather than model compliance.

Performing deliberate mathematical reasoning in visual contexts is a hallmark of advanced Multimodal Large Language Models (MLLMs) and requires a sophisticated synthesis of perceptual grounding and symbolic logic. However, in the realm of mathematical functions, our investigation reveals a critical modality interference phenomenon: even advanced models, while performing textual computational reasoning, tend to disregard or misinterpret essential visual cues. To address this challenge, we propose Func-R1, which synergistically harmonizes precise visual perception and rigorous logical reasoning. Concretely, built upon an explicitly decoupled architecture, we employ a hierarchical post-training framework to progressively identify critical visual evidence and conduct in-depth theoretical reasoning. Furthermore, the Perception-Aligned Theoretic Optimization (PATO) strategy is proposed to steer policy updating towards internalizing fundamental theoretical properties while dynamically rectifying heterogeneous visual information throughout the reasoning process. Extensive experiments across diverse benchmarks demonstrate that Func-R1 delivers the optimal performance among open-source MLLMs, even surpassing GPT-5 with an 8.4% improvement on MathVerse's function-oriented tasks.

We propose a decentralized privacy-preserving learning algorithm in which each agent holds a single private sample and a shared model. Samples are learned sequentially, and each update must preserve the endpoint mappings at previously learned samples while protecting private data. This gives each agent three roles: (i) a learner that updates the model parameters, (ii) a teacher whose sample is learned at the current iteration, and (iii) a protected agent whose sample has already been learned. We build on Tuning without Forgetting (TwF) method to preserve previously learned mappings and show that TwF provides an indistinguishability guarantee for the learner whenever the set of protected agents contains another sample with the same label. For the teacher, we formulate a minimax optimal control problem that models the differential privacy noise as a worst-case disturbance to prevent performance loss while maintaining the same level of privacy for the gradient. For the protected agents, we compute the projections locally and aggregate them using a private push-sum gossip protocol. We prove geometric convergence of the decentralized gossip algorithm and of the distributed projection for TwF.

Battery energy storage systems (BESS) paired with rooftop photovoltaics (PVs) can deliver measurable cost savings under time-of-use (TOU) electricity tariffs; however, the relative performance of model-based and model-free dispatch strategies remains insufficiently benchmarked on full-year, real-world commercial datasets. This paper presents a full-year (2023) comparative evaluation of three BESS dispatch approaches using data from a commercial PV installation operating under a TOU tariff. The examined strategies include: (i) a mixed-integer linear programming (MILP) formulation with perfect foresight, providing an oracle performance benchmark under the assumed model; (ii) a model predictive control (MPC) scheme based on a day-ahead persistence forecast, representing a low-complexity deployable approach; and (iii) a soft actor-critic (SAC) deep reinforcement learning agent trained under causal information constraints. The MILP benchmark achieves an annual cost reduction of 24.6% relative to a no-storage baseline. The persistence-based MPC approach recovers 99.2% of this benchmark using only prior-day data. In contrast, the evaluated SAC agent yields an annual cost higher than the no-storage baseline. This outcome is analyzed in the context of known challenges in reinforcement learning for energy systems, including limited observability and reward design. Overall, the results indicate that, for the studied dataset and tariff structure, persistence-based MPC captures nearly all achievable economic benefits under practical deployment constraints, whereas the considered RL configuration does not yield competitive performance under the same information limitations.

The rapid rise of generative AI is accelerating cloud data center expansion, with electricity demand projected to double by 2026. Because carbon-intensity varies by more than 5.5x across grids and times of day, where and when inference tasks execute significantly affects operational emissions. We address this issue with three aspects in this paper. First, we compile a global alignment dataset unifying 140 operational and planned cloud regions across 8 major providers with five-minute carbon-intensity traces for 145 grid regions from 2022 to 2024, revealing that 50% of current sites lie in medium-to-high carbon-intensity grids, indicating a siting-carbon mismatch and unrealized carbon reduction potential. Second, we develop CATS (Carbon-Aware Task Simulator), a flexible trace-driven framework that profiles six AI inference tasks across multiple GPU types, synthesizes realistic diurnal curve, geographical and task mixes, and SLA constraints, and evaluates spatial and temporal schedulers against two baselines while reporting comprehensive metrics including carbon emissions, energy consumption, runtime, queue delay, and hardware utilization. Third, we quantify achievable CO2 savings under realistic constraints: in a 24-hour trace with 600,000 tasks at fleet utilization of 0.37, spatial shifting reduces CO2 by 38.4% versus speed-first baseline, while temporal shifting yields 16% savings with bounded SLA violations at 3.27%. These results advocate locating future data centers in low carbon-intensity grids and demonstrate that carbon-aware scheduling on today's fleets can achieve substantial operational emissions reduction.

As LLM conversations grow to hundreds of turns, full-context injection incurs $O(N^2)$ cumulative token costs, while lossy summarization or hard truncation irreversibly discards historical state. We propose Pull, a session router that maintains an addressable metadata directory via a local, deterministic Purifier (zero LLM calls, millisecond-level latency). At query time, the LLM lazily materializes only the turns it needs; unmaterialized turns remain accessible but collapsed. Unlike irreversible compression, Pull's materialization is reversible; subsequent queries can expand any collapsed turn. On LoCoEval (128 conversations, 12,780 turns), Pull reduces per-query context tokens (Phase 2) by 75.1 percent on single-hop tasks with equivalent quality ($\Delta = -0.002$, n.s.) and by 72.0 percent on multi-hop tasks with no quality loss ($\Delta = +0.017$). A controlled routing benchmark (7,831 queries x 10 methods) shows that entity lifecycle tracking is empirically a prerequisite for distance-independent routing. On BEAM 1M (14 conversations, 263 questions), Pull improves F1 by +55.2 percent over a truncation baseline.

Generalisations are common in scientific communication, even though they are semantically ambiguous. An automated method is needed to identify and categorise claims according to their level of generalisation, in order help detect an over-reliance on generalisations and possible misrepresentations of scientific findings. We introduce a comprehensive taxonomy of generalisations in the scientific domain, NLPGenX, which labels claims according to their level of generality and framing within the text. We operationalise this taxonomy with an LLM-powered framework, NLPGenA, that automatically classifies sentences from scientific articles into 5 different generalisation classes. We validate our framework with human annotators and use the framework to construct a large-scale dataset of NLP papers annotated according to generality, with auxiliary labels for hedging and vague descriptors (NLPGens). We use NLPGens to analyse the use of generalisations in NLP papers across multiple venues and subdomains, and to examine associations with citation counts, hedging, and vague descriptors.

Does authority in AI teams improve the outcome? Organizational theory asserts that authority facilitates decision making, improving quality. Meanwhile, some nascent AI research suggests that revision under authority makes LLM output worse. Multi-agent LLM frameworks default to giving a Manager agent the authority to send a worker's output back for revision. Prior comparisons test the effect of authority using verifiable tasks. We conduct an experiment on an open-ended task, business-intelligence reporting, using a sample of 43 paired laptop products and 86 runs. Each report is written once by a hierarchical team and once by a flat team. We find that flat teams produce higher-quality reports, scoring higher on Utility (d = 0.42, p = 0.009) and Writing Clarity (d = 0.34, p = 0.030). The reports are the same length, but hierarchical team reports use 53% more hedging words such as "may" and "could", and each revision is associated with a 0.14-point drop in Writing Clarity on a 1 to 5 scale. Before any revision, the hierarchical team's first draft is indistinguishable from the flat team's report. In other words, the quality gap can be traced to revision. Authority improves quality when the Manager can verify the work, else when it can only provide feedback it has a negative effect on quality.

Hip exoskeletons may improve recovery from unexpected gait perturbations, yet personalizing assistance remains difficult because balance is multidimensional and human-in-the-loop experiments are small-sample and noisy. We present a participant-specific composite balance cost that integrates seven biomechanical sub-metrics spanning margin of stability, center-of-mass dynamics, and whole-body angular momentum. The sub-metrics are converted to direction-aligned, dimensionless cost features, and nonnegative fusion weights are learned on the simplex. Coupled with an empirical-Bayes hierarchical model, the learned-composite selector estimates each tested condition's posterior probability of being best, P(best), and a high-probability candidate set with size $K_{0.8}$. The framework was evaluated with three participants walking at 1.1 m/s during unilateral belt-slip perturbations across 46 hip-assistance conditions. In the full-budget analysis (B = 4 repeats per condition), the selector concentrated 80% of the posterior probability within 1 to 5 of 46 conditions, compared with 2 to 12 for equal-weight fusion and 4 to 37 for principal component analysis fusion. This smaller candidate set could shorten personalization experiments and limit participants' exposure to repeated perturbations in future studies. Selected-condition trials showed lower observed composite costs than no-torque trials, with nominal p < 0.05 for P2 and P3. Leave-one-repeat-out refits yielded positive mean held-out rank correlations for all participants and moderate stability of the learned weights and candidate sets. These proof-of-concept results support participant-specific composite balance evaluation for candidate selection in perturbation-based human-in-the-loop experiments.

Early detection of cognitive decline from speech offers a scalable and non-invasive alternative to conventional clinical assessment. Verbal fluency tasks are particularly informative, but most automated approaches aggregate features across an entire recording, overlooking temporal speech dynamics. We propose the Cognitive Instability-Aware Cross-Modal Attention Network (CCMAN), a transfer learning framework that learns task-agnostic cognitive speech representations from multiple memory-probing tasks before fine-tuning on a minute-long semantic and phonemic verbal fluency task. CCMAN integrates semantic, acoustic, and linguistic information through bidirectional cross-attention, gated multimodal fusion, and transformer-based temporal modelling to derive interpretable biomarkers of cognitive decline. Experiments were conducted on 165.44 hours of speech from 843 participants (498 healthy controls, 245 with mild cognitive impairment, and 100 with dementia). CCMAN achieved Macro-F1 scores of 0.81 and 0.59 for binary and multiclass semantic fluency classification, and 0.77 and 0.53 for phonemic fluency, consistently outperforming strong static and temporal baselines. Statistical analyses showed that semantic drift variance and pause variance, but not mean semantic drift, were significantly elevated in both MCI and dementia relative to healthy controls, while pause duration increased progressively over the task with the steepest slope in dementia, supporting global and progressive temporal speech instability as interpretable biomarkers. Evaluation on the independent PROCESS-2 benchmark further demonstrated the generalisability of the proposed framework, improving the baseline Macro-F1 by up to 9%. These findings support temporal speech instability as a dynamic speech biomarker for robust, interpretable, and generalisable early detection of cognitive decline.

Operational logs create a need for private, resource-efficient incident analysis, but aggregate detection scores can conceal severe prediction bias. We present TriCalRAG, a reproducible benchmark for log-anomaly detection with generated root-cause and remediation outputs across BGL, HDFS, Thunderbird, and OpenStack. The primary evaluation compares Qwen2.5-14B and Mistral-Small-22B, served through vLLM on one NVIDIA RTX PRO 6000 GPU (96 GB), under zero-shot, few-shot, and retrieval-augmented generation (RAG) prompting across three data-sampling seeds. We report F1, bootstrap confidence intervals, predicted-positive rates, throughput, and memory use, with DeepLog as a held-out classical baseline. Mistral-Small attains a higher macro-averaged F1 than Qwen2.5-14B (0.644 versus 0.560), whereas Qwen provides approximately twice the throughput. A separate log-probability evaluation compares raw decisions with Contextual Calibration (CC) and Batch Calibration (BC): neither correction consistently improves prediction-balance diagnostics across prompting strategies. Supplementary single-run comparisons extend evaluation to 4-bit Llama-3.1-70B via local Ollama and Claude Haiku 4.5 via Anthropic's cloud API; RAG improves F1 on all four datasets for both models. Claude's reported aggregate F1 increases from 0.566 to 0.695, while the estimated API cost rises from \$0.94 to \$2.15 per 1,000 incidents. Local deployment ablations show approximately 41-fold throughput scaling with batching and 20% lower latency with 4-bit quantization on the tested workload. These findings support retrieval as useful context for anomaly decisions, while the supplementary protocols, unvalidated explanation quality, and prediction-balance diagnostics limit broader claims about RCA accuracy and probabilistic calibration.

Visual feedback in the form of user-uploaded images and videos is becoming increasingly common in e-commerce platforms because it provides authentic evidence of product quality, defects, packaging conditions, and real-world usage. However, visual feedback alone often lacks the contextual explanations and subjective opinions necessary for informed decision-making, while many users provide limited textual feedback due to the effort required to compose detailed reviews. To bridge this gap, we introduce image-grounded review assistance, a novel task that aims to generate editable review drafts from user-uploaded product images. Unlike conventional image captioning, which focuses on objective visual description, the proposed task requires product-specific understanding, sentiment estimation, and evidence-driven review composition under challenging real-world conditions, including degraded image quality, excessive zoom-in, target ambiguity, and partial product visibility. We propose a multi-agent vision-language framework consisting of four specialised roles: product grounding, visual sentiment estimation, visual evidence generation, and review synthesis. The framework employs explicit intermediate representations, including product entities, predicted ratings, and evidence summaries, to improve interpretability and visual grounding. Experiments on a curated subset of the Amazon Reviews Electronics dataset demonstrate the feasibility of generating coherent, product-aware, and sentiment-aware review drafts from visual feedback. To the best of our knowledge, this is the first study to formulate image-grounded review assistance as a multi-agent vision-language reasoning problem, providing a practical step toward AI-assisted review authoring in e-commerce systems.

Alignment applied after pretraining is shallow in a measurable way: a single direction in a model's residual stream can be edited out, and the model stops refusing harmful requests. That fact says how easily refusal can be removed, not what the refusal decision was reading in the first place. We ask what it reads, and we separate that from what the model comprehends. Across four open-weight models spanning three families, moral comprehension is native to pretraining: a low-rank moral subspace crystallizes during pretraining, and alignment rotates it once without rebuilding it. The refusal gate, in contrast, is a fresh post-training construction with only a weak pretraining precursor, written into a narrow control-token channel where the refusal decision is orthogonal to the moral-judgment decision. The central result is causal and comes from one model, OLMo-3. A nested interchange rank sweep patches successively larger slices of the moral subspace between matched requests and reads how much of refusal's response transfers: as the basis widens, moral judgment keeps reading more of it, while refusal levels off at the level of a single harm direction, and about three-quarters of refusal's causal input lies outside the moral subspace altogether. Refusal reads the harm percept, not the moral content that judgment reads on the same patches. The picture is not uniform across families. Llama reads broad moral content; Qwen reads beyond the single harm cue but is unresolved at our sample size; GPT-OSS reads harm, and its refusals can be argued in either direction by its own reasoning trace. Where refusal reads only a low-rank slice and routes around the bulk of what the model knows, a rank-one edit removes it. Whether widening what refusal reads would also deepen the behavior is the open question this raises.

Tool-augmented language models are evaluated on whether they reach the right answer, not on whether they report honestly when a tool fails to supply one. We isolate this post-failure decision with a benchmark of 1,024 items spanning 16 internal-system domains and eight tool-failure types, in which a tool call is enforced and the returned payload is guaranteed to be unusable. Under a deployment-style system prompt, 14.10% of responses are dishonest: the model either asserts a value the payload cannot support or declines while citing a fabricated policy or capability limit. The rate is governed almost entirely by whether the failure is signalled. When the tool returns status:error, dishonesty is absent (0.0%); when it returns status:ok with a redacted, corrupted, stale, malformed, empty or truncated value, dishonesty reaches 45.3%. The behaviour is not an artefact of our prompts: it appears under a neutral prompt (10.17%) and under the shipped prompt of every production agent framework we evaluate, reaching 24.67% under CrewAI's, and none of the nine frameworks we audit specifies what the model should do when a tool fails. Comparing prompt-level defences, we find that the operative variable is not deference to tool output but the absence of a named failure state. Appending a single sentence that requires the model to emit retrieval_status: OK or FAILED before answering reduces dishonesty from 14.10% to 0.87%, with one item of 688 worsening against 92 improving, and transfers unchanged into three foreign agent scaffolds. The emitted flag is faithful in 99.7-99.9% of declarations, giving a runtime detector that needs only a regular expression.