Context: Every agentic AI system shipped to production carries two hidden risks: accumulated Technical Debt (TD) and unmonitored runtime energy costs. While functional benchmarking is common, the empirical link between internal structural quality (specifically TD) and dynamic energy consumption during execution remains unexplored, creating a blind spot for practitioners and organizations managing sustainability and operational budgets at scale. Goal: We propose a confirmatory empirical study correlating Self-Admitted Technical Debt (SATD) with hardware-level runtime energy consumption across agentic frameworks, to determine whether code quality can drive energy-aware design decisions. Method: We will evaluate five open-source agentic frameworks by executing a standardized task suite in a strictly controlled environment. SATD will be extracted via automated Python-based comment mining and categorized via LLM-based classification using fine-tuned prompt, while runtime energy will be measured at the hardware level. Our study will investigate three core research questions: (RQ1) the presence of TD within these frameworks; (RQ2) the variance in runtime energy consumption across architectures; and (RQ3) the statistical correlation between a framework's TD and its task-level energy consumption. Conclusion: The findings will establish whether automated source code analysis can serve as a reliable, early-warning proxy for energy-efficient framework selection, thereby advancing both green software engineering and agentic AI quality research.

Attribute-Specific Fashion Retrieval (ASFR) aims to improve fine-grained image retrieval by focusing on specific attributes. However, existing patch-based attention and Transformer methods often misalign with irregular attribute regions and are prone to background noise, limiting their ability to capture subtle, pixel-level microstructures. To tackle these challenges, we propose SuperFashion, the first ASFR framework that adopts superpixel tokens within a Transformer architecture. SuperFashion initially employs an attribute-guided attention mechanism to extract attribute-related features, which in turn guide the cropping of semantically meaningful image regions. Superpixel segmentation is then leveraged on these regions to generate compact, semantically coherent superpixel tokens. By incorporating modality-specific embeddings for both attribute and superpixel tokens, the superpixel token-based Transformer facilitates adaptive interaction and fusion, thereby enhancing attribute localization and discrimination. Extensive experiments on FashionAI, DARN, and DeepFashion demonstrate relative overall MAP improvements of 1.84%, 9.27%, and 9.35% over prior SOTA. SuperFashion offers a new solution for web-based image retrieval.

The last decade of research on the LOCAL model has seen tremendous progress in understanding locally checkable labeling (LCL) problems, culminating in an almost complete classification of the possible complexities LCL problems can exhibit. In particular, on undirected trees, Chang and Pettie showed that there is no LCL problem with complexity between $ω(\log n)$ and $n^{o(1)}$ and Chang showed that, for every positive integer $k$, there is no LCL problem with complexity between $ω(n^{1/(k+1)})$ and $o(n^{1/k})$; additionally, which side of each gap a problem is found on is decidable. While the class of LCL problems - which, roughly speaking, consists of problems for which the correctness of a solution can be described by a finite set of allowed node configurations, which in turn can be locally verified by a constant-time algorithm - includes many important problems, it has one major restriction: problems can be defined only on bounded degree graphs, which consequently restricts all the classification and gap results mentioned above. In this work, we propose a generalization of LCL problems to unbounded degree using Presburger monadic second-order (PMSO) formulas; more specifically, we consider what we call Local PMSO (LPMSO) problems, i.e., problems whose correct solutions are both finitely described by a PMSO formula and locally verifiable by a LOCAL algorithm in constant time - this class contains many of the important problems studied in the LOCAL model but defines them on unbounded degree graphs. As our main result we prove that, on unbounded degree rooted trees, the aforementioned $ω(\log n)$ - $n^{o(1)}$ and $ω(n^{1/(k+1)})$ - $o(n^{1/k})$ complexity gaps (and their decidability) extend to the class of LPMSO problems.

Given a graph $G = (V, E)$, a signed dominating function is a function $f: V \rightarrow \{-1, 1\}$ such that for every vertex $u \in V$, $\sum\limits_{v \in N[u]} f(v) \geq 1$. The weight of $f$ is defined as $\sum\limits_{u \in V} f(u)$. The objective of the \sd{} problem is to compute a signed dominating function $f$ of minimum weight. The problem is known to be NP-complete even when restricted to bipartite, chordal, and planar graphs. In this paper, we extend the known complexity results for the \sd{} problem. Since the problem is NP-complete on chordal graphs, we study its complexity on split graphs, a subclass of chordal graphs, and show that it remains NP-complete. Moreover, as the problem is W[2]-hard parameterized by weight, we investigate its parameterized complexity with respect to structural parameters. We prove that the problem is W[1]-hard when parameterized by feedback vertex set number (and hence by treewidth and clique-width). Motivated by this hardness result, we consider more restrictive parameters, neighbourhood diversity and twin cover number, and present FPT algorithms.

OPC UA has become the dominant open protocol in operational technology. Time-series databases routinely archive OPC UA telemetry but discard the semantic metadata (node hierarchy, engineering units, and type definitions) which gives sensor values their meaning. Recovering this information from a time-series database is non-trivial: namespace indices recorded at the source are session-local and unstable across restarts, and naive merging across multiple source servers results in identifier collisions. We present opcua-ts, an implemented architecture that persists this semantic information alongside its telemetry in a general-purpose time-series database under a lifecycle-stable join key, and that reconstructs the source address space as a live OPC UA endpoint. We characterize the conditions under which the reconstruction is sound across multi-source deployments and validate the approach with a NodeSet2 XML round-trip against the source server. Initial results from a boiler-simulator round-trip indicate that the approach is feasible.

Layered cybersecurity pipelines transform evidence before they decide on it, and the order of those transformations determines which security facts become visible to which layer. This paper gives layer order a finite-state semantics built from a layer-order automaton, deterministic sequential security transducers, evidence markers, and a final decision automaton. The worked case is HTTP request desynchronization: front-end and back-end processors compute incompatible request boundaries, and the same trace is detected or missed according to whether framing evidence reaches the parser-differential layer before it commits. The results separate completed-trace recognition, online editing, decision synthesis, and faithful enforcement; characterize faithful online enforcement as the regular prefix-closed case under causal visibility; and show that regular policies beyond that boundary remain recognizable without becoming deployable enforcers. The framework is monolithically equivalent to finite-output deterministic edit automata, while preserving layer-local invariants such as marker birth, marker survival, and reorder-sensitive visibility. A concrete parser-pair semantics identifies the forbidden marker factor with CL.TE, TE.CL, TE.TE, and HTTP/2-downgrade boundary disagreement under the stated abstraction, and a contextual reorder congruence classifies which component permutations induce the same decision language. The result is an automata-theoretic account of order-sensitive security failures and a compositional vocabulary for auditing, synthesizing, and comparing layered enforcement pipelines.

Three years ago, Philip Zucker posted an attempt to use answer set programming (ASP) for term extraction from e-graphs Although the task is NP-hard and ASP offers a natural modelling of e-graph terms, the initial attempt did not yield convincing results. From the aspect of practical ASP users, we first pinpoint the way to make ASP work and work well on the task of e-graph extraction. The initial results show the naïve ASP encoding is comparable on efficiency to the well-optimised ILP-based exact DAG extraction in the extraction-gym, and find several extra optimal extraction on the complex instances. This leads us to a further agenda: with the "better together of egg+Datalog", is there a better "better together" by having ASP as a more powerful Datalog? We discuss the potential benefit from each other.

Trajectory prediction is a key enabler of autonomous and cooperative driving systems. However, most existing benchmarks are either sensor-centric, geographically constrained, or based on synthetic mobility traces that do not capture real-world V2X communication dynamics. This paper introduces CAMASA, a large-scale infrastructure-based dataset derived from Cooperative Awareness Messages (CAMs) and Decentralized Environmental Notification Messages (DENMs) collected within the Modena Automotive Smart Area (MASA). The dataset comprises more than 40 million CAMs and 2 million DENMs recorded under authentic urban traffic conditions over multiple months. We present a rigorous preprocessing pipeline that includes filtering, pseudonym reconciliation to account for ETSI privacy-driven stationID changes, and temporal normalization to 10 Hz trajectories, suitable for motion forecasting and time-series analysis. With over 14,000 km of reconstructed vehicle paths and tens of thousands of unique station IDs, CAMASA provides a statistically significant empirical foundation for research on Cooperative Intelligent Transportation Systems (C-ITS). Beyond trajectory prediction, the dataset enables calibration of microscopic urban traffic simulators (e.g., SUMO) and supports the development of realistic Intelligent Transportation Systems (ITS) Digital Twins by jointly modeling mobility patterns and V2X communication coverage in real deployments.

The rapid growth of digital content has made reliable integrity verification increasingly important. Existing solutions rely either on centralized authorities, which introduce trust dependencies and single points of failure, or on decentralized storage systems that incur prohibitive resource overhead. In this paper, we present snaproot, a lightweight system that implements the hash-anchoring paradigm of Haber and Stornetta on the Solana blockchain to provide efficient, decentralized file integrity verification. snaproot generates a SHA-256 hash of a file and stores it immutably on-chain as a permanent reference record. Verification is performed by recomputing the hash and comparing it to the stored value, yielding a deterministic binary outcome. We describe a four-tier trust architecture comprising three realized tiers and one prospective tier for long-term persistence beyond the lifetime of any single blockchain. We present a formal threat model, a security analysis grounded in the second-preimage resistance of SHA-256, and an empirical evaluation on Solana Devnet across file sizes from 1 KB to 500 MB. A central conceptual contribution is the explicit separation between existence proof, the key-independent claim that a file existed at a given time, and authorship proof, the key-dependent binding between a record and a specific wallet identity. This separation allows existence guarantees to survive key loss while preserving stronger authorship claims where keys are retained. We position snaproot against OpenTimestamps, OriginStamp, and Chainpoint and discuss limitations with respect to pre-registration manipulation and AI-generated content.

End users create IoT automation rules via trigger action programming, but their expressions are often fragmented, capturing device operations rather than high level intents. This gap leads to missing conditions, logical conflicts, and overlooked safety constraints, risking hazardous behaviors. To address this, we propose an intent driven requirements completion approach that reframes rule completion as a dual process: reconstructing intent from fragmented rules, then regenerating rules from that intent, with safety embedded throughout. We introduce a Bidirectional Requirements Traceability Tree, a three layer model linking rules, intents, and quality concerns, and design a multiagent framework that combines LLM reasoning with structured traceability. This enables completions that are both functionally complete and inherently safe, while remaining traceable and explainable. Evaluation shows our method significantly outperforms the baselines, improving the rule completion rate by 43% and reducing logical conflicts by over 21%. By grounding completion in intent understanding, we shift the paradigm from user to system responsibility, and from functional correctness to holistic trustworthiness.

Offshore energy hubs (OEHs) use grid-forming modular multilevel converters (MMCs) to enable large-scale offshore wind integration and multi-terminal HVDC operation. In HVDC-connected offshore wind farms and OEHs, the offshore grid-forming HVDC converters absorb active power from an offshore AC grid supplied by the wind farms and convert it to DC power for transmission to the onshore grid. Converter current limiting under different fault types in this setting is an understudied topic in the literature, which mostly focuses on power-injecting converters. This paper proposes a unified current-limiting strategy that combines a variable virtual impedance (VVI), based on a smooth threshold function, with a novel virtual-power (VP) mechanism derived from the power dissipated in the virtual resistance. The VVI ensures current limitation during fault-induced overcurrents while preserving voltage-source behavior, whereas the VP mechanism adds a compensating power term into the synchronization loop, enabling automatic power redistribution among converters. P-delta analysis further shows that a more resistive VVI can improve the transient stability of power-absorbing converters, while the proposed VP mechanism further enlarges the stability margin. EMT simulations validate that the combined VVI-VP strategy limits fault currents, maintains synchronism during severe faults, and achieves coordinated post-fault power sharing in fully converter-based OEHs.

Current scaling trajectories for Generative AI, typified by linear supply-side "stacks," prioritize performance density while externalizing significant thermodynamic and material costs. As the "Twin Transition" of green and digital transformation accelerates, the industry faces technology gaps - including Scope 3 emissions and e-waste recycling - that impede sustainable scaling and lead to social tensions. This study proposes a Regenerative Socio-Technical roadmap that repurposes the Sustainable Production and Consumption system map to reframe artificial intelligence infrastructure as a system-of-systems governed ultimately by planetary limits. By integrating the Institute of Electrical and Electronics Engineers International Roadmap for Devices and Systems (IEEE IRDS) sustainability considerations for semiconductor facilities, the study proposes a metabolic circuit framework that centers "Values and Needs" within production and consumption relationship loops. This study identifies critical gaps in current Nvidia-centric roadmaps and proposes a competing reference architecture. It demonstrates how a spontaneous order of resource parsimony and planetary accountability can provide an actionable pathway for regulatory compliance and industrial resilience in the digital circular economy.

As digital media consumption shifts toward large-scale Over-the-Top (OTT) platforms, the efficiency of the control plane, specifically entitlement and identity verification, has become a critical factor in user experience. Current architectures often rely on synchronous cloud-tethered validation flows that introduce significant latency, especially on resource-constrained consumer electronics. This paper proposes a Hybrid Edge-Cloud Entitlement Framework designed to minimize user-perceived friction. By implementing a secure, local caching layer within device middleware and utilizing an Adaptive Entitlement Cache with Proactive Refresh (AEC-PR) algorithm, we decouple the user interaction from backend network variability. We evaluate the performance on ARM Cortex-A series hardware, demonstrating that localized cryptographic verification reduces authorization latency from a mean of 422.8ms to 18.4ms (a 95.6% reduction) while mitigating implementation-level side-channel risks through deterministic Ed25519 arithmetic and TEE isolation.

Checkerboard patterns are a well-known numerical artifact in density-based topology optimization using the Solid Isotropic Material with Penalization (SIMP) method and linear finite elements. Existing explanations based on mixed-field incompatibility or locking-induced stiffness overestimation explain the artificial stiffness of checkerboard layouts but do not clarify their characteristic spatial localization. In this work, we show that checkerboard patterns systematically emerge in multiaxial load-transfer regions, whereas predominantly uniaxial stress regions remain checkerboard-free. Through systematic numerical investigations, we demonstrate that checkerboarding originates where continuous intermediate densities are mechanically favorable for multiaxial load transfer but are suppressed by SIMP penalization. Due to the characteristic behavior of linear elements, checkerboard layouts provide an artificially stiff discrete substitute for these penalized intermediate-density regions. In contrast, uniaxial load paths naturally favor continuous solid struts, rendering checkerboards mechanically disadvantageous. Our findings provide a unified mechanical interpretation of checkerboarding as the interplay between global stress states, SIMP penalization, and element-level locking, thereby explaining both its origin and the spatial localization.

Metaverse devices, such as virtual reality (VR), have seen substantial development and widespread applications in numerous areas. Although recent studies have revealed privacy leakages in VR, these vulnerabilities were limited in the scope of observable behaviors in virtual scenes (e.g., what a user is seeing). In this work, we uncover the feasibility of going beyond the scope of observable user behaviors to unobservable brain EEG-correlated representations (e.g., what a user is perceiving) by leveraging unrestricted motion sensors in VR headsets to reconstruct brain EEG signals, a seemingly neglected but promising vector. The insight is that the inbuilt motion sensors (e.g., accelerometers) in the VR headset can capture subtle vibrations induced by pupillary responses, which are highly correlated with users' visual stimuli and in-brain perceptions. Therefore, we design and implement BraVeSpy to systematically investigate and demonstrate the feasibility of this severe privacy leakage originating from brain EEG-correlated representations reconstructed from variations of inbuilt motion sensors. Our extensive evaluation results from different VR devices show that BraVeSpy, for the first time in the Metaverse, can reveal unobservable privacy, where we successfully unveiled perceptive images in the brain with 52.0%-67.2% accuracy. In particular, we also find that BraVeSpy outperforms the current approaches that are limited to coarse-grained inference of observable behaviors and achieves over 85.0% accuracy in inferring user activity-related sensitive information, such as fingerprinting websites, apps, and streaming videos, and over 96.0% accuracy in user de-anonymization, gaze movement tracking, and virtual keystroke inference.

Autonomous AI agents have driven the transition from conversation to task execution, shifting security failures from textual deception to system compromise. Although security evaluation is crucial for proactive risk prevention, prior work is constrained by fundamental bottlenecks, including fragmented risk coverage, static or low-fidelity execution environments, and single-dimensional and coarse-grained assessment metrics. To address these challenges, we propose AgentCanary, a comprehensive security evaluation framework for autonomous AI agents. AgentCanary provides a systematic solution along three contributions. First, comprehensive risk coverage: we introduce an orthogonal Entry $\times$ Impact risk taxonomy that decouples how adversarial influence enters the agent from what harm it ultimately causes, and instantiate it as a scenario-aligned task suite spanning realistic deployment workflows. Second, a high-fidelity real executable environment: rather than static Q&A or mocked tool responses, agents interact with real tools against dynamically provisioned task artifacts, with persistent state across multi-step interactions that naturally supports long-horizon attack evaluation. Third, trajectory-grounded multi-dimensional evaluation: evaluation consumes the full agent trajectory rather than the reply text or a single tool call, enabling decomposed scoring along three orthogonal dimensions, Outcome Safety, Security Awareness, and Task Utility. We evaluate a broad set of frontier models on AgentCanary against multiple established adversarial attack methods across three agent frameworks. The results reveal that current agents often fail to recognize the attacks they face, particularly under compromised skills, persistent state, and long-horizon execution attacks, and provide a systematic baseline for developing more reliable and secure agent systems.

Particle-based liquid simulation is widely used in graphics and physical modeling, but high-resolution rollouts remain computationally expensive. Consequently, many methods aim to recover fine-scale dynamics and dense transport patterns from coarse particle simulations. However, these methods typically rely on additional particle generation, which still incurs considerable computational overhead and leads to poor representation. To this end, we propose AnisoLift, a structured latent closure framework that augments each coarse particle with learnable anisotropic ellipsoidal components. This allows the model to capture directional local structure from the underlying high-resolution flow without introducing extra particles. Given a coarse simulation, our model predicts residual corrections to particle states to bring the updated state closer to the aligned high-resolution teacher. Our training objective jointly supervises particle dynamics and anisotropic geometric structure, encouraging both physical consistency and structural coherence. Extensive experiments show that our approach enhances coarse liquid simulations through improving fidelity to fully resolved flow behavior.

Existing automated RESTful API testing approaches commonly rely on simple checks (e.g., HTTP status codes, schema conformance), which are insufficient for detecting semantic faults, business logic violations, and state-dependent inconsistencies. To address this, we propose MASTOR, a Multi-Agent approach for generating Semantic Test Oracles for RESTful APIs based on implementation source code. MASTOR consists of two phases: source analysis and oracle generation. The former employs a source extraction agent to construct a source context for each endpoint operation by analyzing a transitive import closure of relevant source files. The latter employs two parallel oracle-generation paths over the collected contexts: a single-operation path producing status and field oracles per operation, and a multi-operation path generating behavioral consistency oracles for operation sequences by leveraging cross-operation semantic associations. Both paths apply a challenger-agent review, where a dedicated reviewer identifies weaknesses and issues improvement hints to guide targeted regeneration, followed by oracle normalization to filter out structurally invalid oracles. We evaluated MASTOR on a benchmark of 13 open-source RESTful API projects (296 operations, 251,303 lines of code) from the WFD and PRAB datasets. MASTOR achieved an average mutation score of 75.4%, generating 10,022 oracles. These oracles were translated into executable assertions via ToJUnit and ToPostmanAssertify, and into human-readable descriptions via ToReadable. In a baseline comparison on 50 selected operations, MASTOR outperformed Direct Prompting by 30.1 percentage points (69.9% vs. 39.8%) and SATORI by 49.4 percentage points (69.9% vs. 20.5%).

Driven by the rising value of data in applications such as advertising, finance, and machine learning, markets for data products have become increasingly important. Data markets mainly sell two kinds of products: datasets and machine learning models. Since these products can be replicated at negligible marginal cost, sellers naturally version them through query access and noisy model releases. Versioning immediately raises an arbitrage problem: a buyer may combine cheaper purchases and recover a more informative product at a lower total price. Existing work on query and model pricing studies arbitrage-freeness when buyer values are treated as exogenous, whereas the literature on selling information derives value from the buyer's decision problem but ignores arbitrage-freeness. Accordingly, we study the seller's optimal data pricing problem where buyers value data through Bayesian decision making and we impose arbitrage-freeness constraints. We first interpret query and model pricing as special cases of information pricing, and formulate the general arbitrage-free information selling problem, show the computational hardness and give a branch-and-bound algorithm based on McCormick relaxations. We then consider threshold utilities where buyers have a positive value if and only if the experiment is sufficiently informative. Under this condition, we find that the arbitrage-freeness can be characterized by Blackwell dominance, which in turn unifies the arbitrage-free conditions for query pricing \cite{deep2017design} and model pricing \cite{chen2019towards}. Finally, we characterize the revenue-maximizing pricing under restricted query and model menus.

How much voter input is necessary in order to ensure representation in multiwinner elections? If voters are randomly selected from an underlying population, how many draws are necessary to find a proportional committee of $k$ candidates, with high probability? Sample-based adaptations of standard multiwinner voting rules that satisfy the justified representation (JR) proportionality axiom use $\tilde O(k^5 \log \frac{m}δ)$ sampled approval ballots over $m$ candidates, where $δ$ is a probability of failure and $\tilde O$ suppresses $\mathrm{polylog}(k)$ factors. We present a rule for which the sample complexity of JR-family proportional committee selection is $\tilde O(k^{4}\log \frac{m}δ)$. This separates the sample complexity of JR from that of the natural corresponding additive approximation to the voter coverage (Chamberlin-Courant) objective, which we show requires $Θ(k^5\log \frac{m}δ)$ samples. For lower bounds, we present a family of instances with $m, \frac{1}δ \in \mathrm{poly}(k)$ for which $Ω(k^3)$ sampled ballots are necessary in order to identify a JR committee. We also show a dependence on $\log m$ is necessary. This lower bound is versatile, and also applies to Hare proportionality for solid coalitions (PSC) for ranked ballots. Unfortunately, no number of sampled ballots suffices to satisfy the slightly stronger Droop JR and Droop PSC axioms with high probability. But mild relaxations of JR require fewer samples, as do the beyond-worst-case domains and actual approval preferences we evaluate.

This study compares two approaches to profiling how learners offload cognitive activity to LLMs during a synthesis writing task. Drawing on Salomon's distributed cognition and the Kintsch and van Dijk model of text comprehension, the study operationalises offloading to an LLM in two ways: as a volume of LLM use and as content of what is offloaded, both along with prior knowledge. Data from 97 university students interacting with a general-purpose LLM via a custom interface were analysed using k-means clustering. To capture the content of offloading, their prompts were interpreted as to who performs the activity (active or passive) and at what level of comprehension (local or global). Volume-based profiling (k=4) differentiated learners primarily by prior knowledge, with volume negatively associated with essay authorship. Content-based profiling (k=5) revealed qualitatively distinct patterns of offloading, from vocabulary clarification to active direction of structuring and generation to passive delegation of comprehension at both levels. These patterns reflect different fragmentation of the cognitive process, with differences in learning strategies, behavioural markers, and essay authorship. Combining volume and content of offloading could improve future analyses on how LLM use redistributes cognitive activity and its effects on learners.

The transition towards a sustainable power system is enabled by the replacement of conventional synchronous generators with converter-interfaced renewable energy sources. However, the resulting loss of rotational inertia and governor damping causes significant frequency deviations and can therefore cause instability. The focus of this paper is the optimal allocation of virtual inertia and damping in the power system activated by established converter control schemes. To this end, we propose a novel dynamic optimization algorithm that considers performance metrics for system stability, cost-efficiency, and resilience. In addition, our algorithm considers the magnitudes and locations of disturbances in the power system for the optimal allocation. Finally, we validate our approach on a three-area system and also compare our results with a $\mathcal{H}_2$ system-norm-based allocation approach.

Traditional test adequacy metrics measure a system's implementation, not whether it adheres to its expected behaviour. While developers rely heavily on code coverage and mutation testing to assess test suite quality, these metrics are fundamentally implementation-centric and cannot detect gaps between what the code is expected to do and what it actually does. Unfortunately, there has been no way to reliably detect these discrepancies; in this paper we introduce an automated proof-of-concept approach to investigate these gaps. The approach extracts expected method-level behaviours from natural language documentation and source code, maps them to existing test cases, and identifies gaps between expected and validated behaviours. We evaluate the approach across ten popular open-source Java libraries comprising 8,922 methods, extracting 20,729 behaviours with 93.1% precision. Our empirical analysis conservatively estimates that 17.5% of detected expected behaviours remain entirely untested, which we term as the test suite's behavioural gap. To determine if these gaps are merely an artifact of human-driven testing, we evaluate state-of-the-art automated test generators (EVOSUITE / ASTER), finding that they similarly fail to validate at least 20.6% / 27.1% of detected expected behaviours. We further demonstrate that behavioural gaps are not predicted by traditional structural metrics: the majority of untested behaviours occur in methods that already have high line coverage, and over half persist in methods with high mutation kill score. These results suggest behavioural coverage acts as an independent dimension of test suite adequacy that can complement traditional structural metrics.

Production heterogeneous supercomputing platforms are increasingly used to host large language model (LLM) training workloads. However, existing GPU-oriented training runtimes typically rely on high-bandwidth device memory, fast interconnects, and mature collective communication libraries, making them difficult to directly adapt to MT-3000, a platform with an explicit memory hierarchy, limited usable DDR capacity, and constrained inter-cluster communication. This paper presents RATrain, a resource-aware training runtime for dense LLMs on bandwidth-constrained heterogeneous supercomputing platforms. RATrain formulates standard non-interleaved 1F1B training as a training-state lifecycle scheduling problem, and schedules gradient synchronization, parameter update, parameter-view prefetching, and activation recovery at layer-level and stage-local granularity. RATrain further combines an MT-3000-aware execution backend for efficient and predictable FP16 GEMM, Attention Backward, and explicit data movement with a resource-aware planner that selects feasible training configurations under the 20GB usable-DDR constraint per compute cluster. We implement RATrain on a real MT-3000 platform and evaluate it using LLaMA-2-7B, Baichuan2-13B, Qwen2.5-32B, and LLaMA-2-70B configurations. Results show that RATrain achieves up to 1.35$\times$ end-to-end speedup over MT-3000-adapted GPU-style training strategies. For LLaMA-2-7B, RATrain scales to 1024 compute clusters, reaches 112,790.55 tokens/s, and achieves 97.0\% scaling efficiency. A further 1.028B-token correctness run shows that RATrain preserves the loss trajectory of a semantically equivalent Baseline-1F1B run, with a maximum relative loss deviation of 0.081\%.

Minimum-weight decoding for two-dimensional color codes is NP-hard (Walters and Turner 2026), motivating the search for approximation guarantees beyond worst-case exact decoding. We study a block-based decoder for triangular color-code lattices. The decoder satisfies the deterministic additive guarantee \(\lvert E_{\mathrm{alg}}\rvert \leq \operatorname{OPT}(S)+O(n/τ)\), where \(n\) is the number of vertices and \(τ\) is the wall spacing. We show that this additive guarantee becomes a near-optimal multiplicative guarantee under natural noise models. For constant-rate i.i.d. face noise and constant local degree, choosing \(τ=Θ(ε^{-1})\) gives a \((1+ε)\)-approximation with probability \(1-\exp(-Ω(n))\), in time \(n2^{O(ε^{-1})}\). We also prove a smoothed analogue: the same near-optimality guarantee holds when an arbitrary adversarial error pattern is perturbed by independent constant-rate noise. Finally, in the low-probability regime \(p=o(1/\log^2 n)\), the syndrome decomposes into small active regions with high probability, allowing independent component-wise decoding and yielding an exact minimum-weight correction in time \(n2^{O((\log n)^{3/2})}\). These results show that, despite worst-case hardness, color-code decoding admits strong average-case, smoothed, and sparse-regime guarantees.

Semantic-ID (\sid) tokenizers are increasingly reused as standalone artifacts in generative recommendation: an exported item-to-code mapping becomes the address space that a later sequence generator must use. These mappings rarely come with a common inspection interface, so coverage gaps, full-code aliasing, behaviorally weak prefixes, tail compression, and prefix fan-out are often found only after downstream training. We present \tool, a mapping-first diagnostic resource for \sid tokenizer artifacts. \tool defines a small adapter contract over item mappings, metadata, interactions, and optional generator traces; validates the contract; and reports mapping-level probes for utilization, aliasing, neighborhood alignment, popularity allocation, and structural cost, with hooks for temporal churn and generator traces. \tool reports inspectable artifact profiles before downstream leaderboard scores. The released resource covers four tokenizer artifact lines: a same-item GRID/RQ-KMeans-style and ReSID/GAOQ contrast on 23,742 Musical items, plus released LETTER and LC-Rec item-index artifacts. In the Musical contrast, the GRID-style feature-text export has 3,749 unique full codes and a 0.977 full-code aliasing rate, while ReSID/GAOQ is aliasing-free in its exported mapping. Yet the strongest prefix--co-occurrence alignment comes from a deterministic category-prefix control, not from either learned export row (0.447 versus 0.154 and 0.055--0.080), showing that addressability and behaviorally meaningful prefixes should be inspected separately. Cross-domain, fixed-reranker, and mechanism-probe checks support the same diagnostic direction: prefix alignment is a candidate-exposure signal, while final ranking quality remains a downstream model question.

Collecting the FIFA World Cup sticker album presents a classic public-goods and collective-action dilemma, in which completing a collection on one's own is highly inefficient. To evaluate how localized community norms shape collective efficiency, we use agent-based modeling and Monte Carlo simulations, parameterized with empirical field observations from exchange meetups in Natal, Brazil. Reflecting the tournament's recent expansion, the Panini 2026 album features 980 individual stickers, including 68 metallic specials. We contrast a standard baseline economy (1:2 special-to-normal exchange ratio) with an overprotective, strict strategy (exclusive special-for-special trading) and an altruistic, generous strategy (in which advanced players surrender needed duplicates to assist peers). Our findings reveal that overprotective rules trap liquidity and drive network-wide inefficiency. The strict strategy increases median completion costs by 10 packs and severely penalizes the least fortunate 5\% of collectors, adding 20 packs in large cities and 30 in small communities. Conversely, widespread generosity optimizes network liquidity and dramatically compresses the long tail of bad luck. Introducing the generous strategy reduces required purchases for the 5th percentile by 90 packs in large-scale configurations and 130 packs in smaller clusters. Furthermore, widespread altruism triggers a strong functional coupling that effectively synchronizes completion rates across the network. This study demonstrates that while rigid, protective norms degrade collective welfare, generosity successfully mitigates pack-draw variance, transforming an expensive, isolated hobby into a resilient, highly efficient public good.

In this paper, we develop a Virtual Reality-based immersive learning environment that allows teachers to conduct a lesson in a virtual space using Photon Fusion. The proposed system allows teachers and students to be present in the same virtual space regardless of their actual physical locations. The teachers can verbally communicate with students in real-time, interacting with 3D learning materials. By adopting Photon Fusion, the system achieves stable real-time communication and synchronization among multiple players. Evaluation results demonstrate that the proposed system provides stable communication performance, good usability, and minimal VR sickness, confirming its effectiveness as an immersive learning environment.

The rapid proliferation of Internet of Things (IoT) devices has enabled unprecedented automation and connectivity, but it has also substantially increased the attack surface, exposing networks to sophisticated cyber threats, including zero-day and adversarial intrusions. Traditional Intrusion Detection Systems (IDS) struggle to generalize to unseen attacks, often require substantial computational resources, and lack interpretability, particularly in resource-constrained and heterogeneous IoT networks. Recent advances, including Deep Learning (DL), open-set detection, and Large Language Model (LLM)-based semantic reasoning, address some of these challenges but typically focus on zero-day and adversarial threats and rarely combine semantic reasoning with multi-agent systems. To overcome these limitations, we propose a semantic multi-agent ID that integrates four specialized agents (e.g., Scout, Mutator, Auditor, and Arbiter) that leverage semantic embeddings and multi-stage probabilistic decision fusion. The Scout induces structured hypotheses from semantic embeddings; the Mutator generates adversarially constrained variants; the Auditor evaluates consistency and filters unreliable outputs; and the Arbiter produces interpretable, risk-aware alerts. Extensive experiments across multiple real-world IoT datasets demonstrate that the proposed system achieves 95.9% overall detection accuracy, reduces false-positive rates to 6.8%, improves zero-day detection to 87.9%, and maintains computational efficiency suitable for edge deployment.

Communication skills are increasingly recognized as essential in Software Engineering, yet discussions about them remain fragmented across academic and gray literature. This fragmentation is problematic because it limits a broader understanding of how communication is valued, taught, and applied in both educational and professional settings. Through a multivocal literature review, we found strong convergence between academic and gray sources in treating communication as a core competency, while also identifying differences in emphasis, with academia focusing on conceptualization and empirical evidence and gray literature stressing practical consequences and emerging industry practices.