Papers

Motion blur arises from the temporal integration of a continuous sharp signal over a finite exposure window, yet existing learning-based methods sidestep this physical model and predict only the sharp signal itself: most single-image deblurring methods recover a single frame at the exposure center, while blur-to-video methods predict a fixed set of frames. We introduce PickMoment, a continuous-time reformulation that directly learns the interval-mean blur over arbitrary sub-intervals of the exposure with a single deterministic model. Drawing an analogy to MeanFlow's average-velocity formulation, we train the model with three supervisions derived from the blur integral: an empirical reconstruction loss from available subframes, an additivity loss that enforces self-consistency across overlapping sub-intervals, and a sharp-frame loss anchored at the zero-interval limit. A single trained model unifies single-image deblurring, blur-to-video generation, and continuous-time pick-a-moment recovery as different queries to the same network, with no separate training for each task. Our PickMoment achieves state-of-the-art performance among generative-based deblurring methods on GoPro and HIDE while competitive against restoration-based methods on RealBlur, and the highest per-frame fidelity on GoPro-7 blur-to-video, all in a single forward pass without iterative sampling.

Alzheimer's disease (AD) diagnosis requires sequential evidence acquisition under heterogeneous test costs and patient burden. Fixed-modality predictors do not jointly decide which test to acquire or when the available evidence is sufficient for diagnosis. We propose SCOPE-AD (Sequential Cost-Aware Ordinal-Belief Planning with Energy-Based Models for Diagnostic Agents) for cost-aware classification of cognitively normal (CN), mild cognitive impairment (MCI), and AD cases. A mask-aware ordinal model represents uncertainty along the ordered CN–MCI–AD continuum. Retrospective training records provide sampled Bellman targets for an energy-based teacher, whose action distributions are distilled into a Qwen policy. At deployment, the agent selects acquisition or diagnosis actions under availability and budget constraints without access to unacquired values. After each acquisition, the evidence and ordinal belief are updated before the next decision. On ADNI, SCOPE-AD achieves 77.70% Macro-F1 at an average acquisition cost of \$50.46, exceeding the strongest evaluated baseline by 9.34 percentage points. Full-modality evaluation raises Macro-F1 by only 1.89 points while increasing acquisition cost by 116.7 times. These results support selective acquisition for cost-effective diagnosis.

We construct quantum CSS codes with rate-distance tradeoff approaching the CSS GV bound, which can be encoded by a quantum circuit of logarithmic depth and a linear number of gates. Our construction is heavily inspired by a work of Brehm and Resch and can be viewed as a quantum analogue of classical repeat multiple accumulate codes. The encoder has a particularly simple structure, consisting of a single constant depth outer quantum circuit followed by several classical accumulation and inverse-accumulation layers.

We construct a smooth projective Calabi-Yau threefold with a flat elliptic fibration of Mordell-Weil rank twelve and Hodge numbers $(h^{1,1},h^{2,1})=(25,1)$. The construction starts from the Kummer surfaces associated with the products of a fixed elliptic curve and the level-four modular family. The threefold is birational to an anticanonical quartic in $\mathbb{P}(\mathcal{O}_{\mathbb{P}^1}(3)^{\oplus2}\oplus\mathcal{O}_{\mathbb{P}^1}(4)\oplus\mathcal{O}_{\mathbb{P}^1})$. We give twelve independent sections and construct a projective crepant resolution of the associated Weierstrass model. Varying the fixed elliptic curve gives a family of quartic resolutions that is locally complete at the original member. The resulting fourfold is birational over $\mathbb{C}$ to a diagonal order-four quotient of a product of four elliptic curves with $j$-invariant $1728$, and every smooth projective model has Kodaira dimension zero. An algebraic correspondence identifies the third cohomology of the original quartic resolution with the tensor product of the first cohomology of the fixed elliptic curve and the transcendental cohomology of the modular K3 surface. The Galois representation on the latter factor is associated with a weight-three CM newform.

Uniform-state diffusion models (USDMs) can revise any token at any denoising step, which lets them correct their own mistakes, a key advantage over masked diffusion. Self-correction, however, requires both revising incorrect tokens and retaining correct ones, and we show that current USDMs lack the latter. Even under greedy-tail decoding, state-of-the-art USDMs (DUO, UDLM, and uniform-noise SEDD) keep revising 173–270 of 512 positions at every step, and these large, uncoordinated edits collapse sample diversity. A random-token corruption experiment traces this deficit to the models themselves: they reconstruct clean and corrupted tokens with nearly identical accuracy, even though clean tokens are easier targets. A decomposition of the validation NELBO shows that training barely rewards retention: incorrect predictions are heavily penalized at corrupted positions but almost free at clean ones. We propose Correct-Token Retention Regularization (CTR-Reg), a simple but effective auxiliary loss that trains the model to retain tokens left unperturbed by the forward process and requires no change to the sampler. CTR-Reg improves clean-token accuracy by 26.5 percentage points on average across six benchmarks, while leaving corrupted-token accuracy virtually unchanged, and its per-step revisions converge to only 3–11 positions. With just five greedy-tail steps, generative perplexity more than halves under CTR-Reg for all three models while diversity is preserved, and these gains hold across sampling budgets. Our results identify correct-token retention as a key missing ingredient for self-correcting diffusion language models, and demonstrate an effective fix.

Growing environmental and energy-related challenges have intensified the search for technologies capable of improving energy recovery and resource efficiency. Bioinspired photonic structures offer promising opportunities for the development of efficient infrared absorbers by exploiting optical mechanisms found in natural systems. In this work, we present a bioinspired approach for the design and optimization of a broadband mid-infrared absorber based on the highly absorbing wing scales of the Troides magellanus Magellan birdwing butterfly. The characteristic geometry of the biological structure is rescaled and adapted to shift its optical response from the visible to the infrared spectral range. Numerical simulations are used to investigate the influence of the main geometrical parameters on the spectral absorptance and to identify an optimized configuration. The optimized structure achieves an absorption enhancement of 391% compared with a flat steel reference. The improved optical response is attributed to the optimized high-aspect-ratio geometry, which enhances multiple scattering and light trapping, thereby increasing the effective light-matter interaction area. These results demonstrate the potential of bioinspired geometric scaling as an effective strategy for engineering efficient infrared absorbers and tailoring their spectral response for thermal radiation management and infrared photonic applications.

A Neumaier graph is a non-complete edge-regular graph containing a regular clique; it is called strictly Neumaier if it is not strongly regular. In this paper we present a construction using finite rings that unifies several known results and yields three new families, each containing infinitely many strictly Neumaier graphs. We characterize when the resulting graphs are strongly regular. The family obtained from norm functions of finite field extensions includes the collinearity graphs of generalized quadrangles of \(*\)-Tits type arising from the classical hyperoval. We further show that, when the cyclotomic numbers involved have order at most seven, four types of rings suffice to obtain, up to isomorphism, all strictly Neumaier graphs arising from our construction. As a consequence of our results, each of these four types has now been used to construct Neumaier graphs.

Day-ahead operation of local energy communities (LECs) is affected by uncertainty in PV generation and electricity demand, which may compromise the feasibility of committed exchanges with the upstream grid. This paper proposes a forecast-residual-based chance-constrained optimization (CCO) framework that characterizes uncertain parameters through their forecasts and associated residual distributions. Under the adopted zero-mean Gaussian assumption, the resulting chance constraints admit an exact deterministic-equivalent reformulation in which uncertainty is embedded through analytical safety margins, preserving the mixed-integer linear structure of the original scheduling problem. The model jointly coordinates PV and BESS operation, low-voltage distribution network constraints, internal energy sharing, and day-ahead grid-exchange commitments. Once the physical schedule is determined, an ex-post allocation stage distributes the available community energy pool among users according to predefined participation coefficients, identifies post-allocation surpluses and deficits, and maximizes their internal matching while preserving the aggregate grid exchanges obtained from the CCO. The framework is evaluated on a reduced 206-node European low-voltage feeder with up to 55 community users and benchmarked against a two-stage stochastic programming formulation. Results show that, for forecast-error levels close to 5%, the proposed CCO achieves operating outcomes comparable to the stochastic benchmark while retaining computational requirements close to the deterministic formulation. The ex-post allocation results further show that internal matching can reduce gross energy exchanges that require settlement with the upstream grid.

Integrated sensing and communication (ISAC) performance analysis commonly employs the Cramér–Rao bound (CRB) as the sensing metric. The CRB lower-bounds estimation-error variance under a target-present model but does not characterize the reliability of the energy detection that receivers often use to validate returns against noise and clutter. Considering the range CRB in a single-user, single-target joint power–bandwidth allocation problem, we establish the conditions under which a rate-optimal allocation at a given CRB requirement fails to achieve a given detection probability. We then develop detection-constrained rate–CRB characterizations at two levels: 1) a detection-aware (DA) formulation that treats the CRB and detection probability as parallel requirements, and 2) a detect–then–estimate (DTE) formulation using a detection-gated CRB defined on frames that cross an energy threshold. We analyze the resulting changes in the optimal rate and feasible power–bandwidth allocations, and for DTE further show how detection probability and conditional estimation accuracy jointly determine downstream tracking performance. Multi-user, multi-target resource allocation is then formulated under CRB-only, DA, and DTE characterizations. The results show that CRB-only allocation can conceal inadequate detection reliability, and that accounting for detection and receiver architecture materially changes the power–bandwidth allocation and the communication–sensing tradeoff.

Personalization encoders compress evolving interaction histories into preference states used to rank items or condition text generation. A task head operating only on this state can miss useful evidence that remains in the frozen encoder's cached representations for individual timesteps. We study this recoverability gap and propose REPAIR, which compares cached representations with the current preference state in a compact learned coordinate space. It resolves corrective evidence over extended history, recent interactions, and localized bursts. It then selects which patterns at which timesteps contribute and adds their aggregate correction to the state before the task head. Encoder-host repair reuses representations from the existing forward computation without re-encoding the history. Across MovieLens, PENS, MIND, and Amazon Reviews 2023, training only REPAIR improves MRR and nDCG@10 for all twelve representative recommendation hosts while both encoder and task head remain frozen. Head-only finetuning of the same hosts yields smaller gains. For example, Mamba4Rec on MovieLens gains 3.96 MRR points, compared with 0.19 from head-only finetuning. Rank and temporal diagnostics support a compact, host-dependent corrective structure. In personalized generation, IMPerSumm improves the two reported weighted PerSEval variants, which assess responsiveness to user preference, by up to 25.23%. These results support post-compression state correction and distinguish the availability of preference evidence from its downstream use.

Bayesian Optimisation (BO) is a powerful framework for the optimisation of expensive black-box functions, but typically requires refitting a surrogate and maximising an acquisition function at every evaluation step. In-context approaches based on Prior-data Fitted Networks (PFNs) amortise part of this cost by pre-training transformers on functions drawn from synthetic priors. PFNs4BO amortises the surrogate but still relies on a numerically maximised acquisition function, while FIBO performs BO fully in-context by sampling optimiser locations from a learned density, which fixes the decision rule and admits no surrogate. Learned acquisition functions score a finite candidate set with a trained network, but, lacking a label for the query, learn the score by reinforcement learning on previously solved tasks. We propose IQS-BO, a PFN that learns the query decision by supervised learning on synthetic priors. In a single forward pass, IQS-BO predicts the probability that each candidate maximises the objective over the set, and we show that the minimiser of its objective is the posterior probability of this event. The model can be pre-trained without a surrogate for fully in-context BO, or take the predictions of a fixed probabilistic surrogate as additional input, amortising only the decision step. Our method proposes queries at a fraction of the cost of acquisition-based methods, while either matching or outperforming standard BO with Gaussian processes (GPs) and available in-context methods on synthetic and real-world benchmarks. Finally, we propose a mixture prior for pre-training PFNs which combines samples from GPs with functions exhibiting warped inputs, isolated narrow optima, or plateaus that are poorly modeled by stationary kernels common in GP surrogates. We show that pre-training on this prior can lead to improved optimisation performance.

We analyze a coupled PDE-ODE model for metastatic cancer invasion involving epithelial-like and mesenchymal-like cancer cells, cancer-associated fibroblasts (CAFs), the transforming growth factor $\beta$, matrix-degrading metalloproteinases, and extracellular matrix reconstruction by CAFs. The model incorporates bidirectional phenotypic transitions between epithelial-like and mesenchymal-like cancer cells, cancer-cell haptotaxis, negative fibroblast haptotaxis, as well as cell-specific proliferation. Under explicit parameter restrictions, we establish global classical solvability for the full model in two dimensions using an energy-based approach, and for a three-dimensional variant with constant phenotypic transition rates using coupled $L^p$-estimates. Numerical simulations illustrate primary tumor invasion and metastatic spread to secondary sites.

We study tilings of symmetric and alternating groups by conjugacy classes. For $S_n$, the case where the conjugacy class consists of transpositions was first investigated by Rothaus and Thompson in 1966 and remains open. We prove that this long-standing case is the only unresolved case for symmetric groups: every other nonidentity conjugacy class of $S_n$ fails to tile $S_n$. Our proofs combine representation-theoretic methods with combinatorial arguments concerning products of permutations. We also completely resolve the corresponding problem for alternating groups in a stronger form: for $n\geq5$, no normal subset of $A_n$ that avoids the identity can tile $A_n$. \iffalse We study the problem of whether a conjugacy class tiles the symmetric group or alternating group. For $S_n$, the case where the conjugacy class consists of transpositions was first investigated by Rothaus and Thompson in 1966 and remains open. We prove that this long-standing case is the only unresolved case for symmetric groups: every other nonidentity conjugacy class of $S_n$ fails to tile $S_n$. The proof combines representation-theoretic methods with combinatorial arguments concerning products of permutations. We also completely resolve the corresponding problem for alternating groups in a stronger form: for $n\geq5$, no normal subset of $A_n$ that avoids the identity can tile $A_n$. \fi \iffalse In this paper, we study when a conjugacy class or a normal subset tiles the symmetric or alternating group. The case of transpositions was first investigated by Rothaus and Thompson, and in general, it remains open. We show that every other conjugacy class of a symmetric group does not tile. We also prove that normal subsets in $A_n$ ($n\ge 5)$ not containing the identity element cannot tile.

Identifying the particle nature of dark matter remains one of the most significant challenges in modern physics. Direct detection experiments aim to observe rare scattering events between dark matter particles and terrestrial targets, a task that demands extreme background suppression and sensitivity to minute energy depositions. This review critically assesses the current experimental landscape, organizing detector technologies by the fundamental physical trade-offs that define their scientific reach. We contrast the multi-tonne scalability of noble-liquid Time Projection Chambers (TPCs), which currently define the sensitivity frontier for high-mass Weakly Interacting Massive Particles (WIMPs), with the precision of cryogenic semiconductors and Charge-Coupled Device (CCD)- based sensors, which dominate the search for low-mass and sub-GeV candidates. Special emphasis is placed on the role of advanced reconstruction pipelines and machine learning (ML) as integral components of detector performance. Finally, we discuss the strategic roadmap for the next decade as experiments approach the neutrino fog, where coherent elastic neutrino-nucleus scattering (CE{\nu}NS) produces an increasingly significant and ultimately irreducible background that can mimic DM-induced nuclear recoils. We argue that future progress will rely not on a single technology but on a complementary global program combining increased target mass, ultra-low energy thresholds, improved background discrimination, and distinct observables - such as directionality and temporal signatures - to maintain robust discovery capability in the presence of neutrino-induced backgrounds.

The widespread adoption of Mixture-of-Experts (MoE) has created a growing need for deployment on heterogeneous platforms. However, it exposes a fundamental mismatch between the algorithmic demands of large-scale MoE and the disparate characteristics of hardware.Existing CPU-GPU hybrid inference systems fail to resolve this as they either encounter PCIe bandwidth bottlenecks when loading experts to GPUs, or rely heavily on CPU computation. Consequently, this leads to low resource utilization and inevitable violations of fixed latency budgets as parameters scale. In this paper, we identify and exploit Cross-Asymmetry–a structural alignment between the algorithmic workload skew of MoE routing and the physical disparity of heterogeneous hardware. To this end, we introduce RapidMoE, a residual offloading system for efficient large-scale MoE inference. We propose how RapidMoE leverages a residual-split framework to enable offloading paradigm shift from expert-level to bit-level, which unfolds across three key dimensions: (1) data representation, enabling compact and decoupled storage; (2) routing strategy, partitioning computation into dual paths aligned with hardware capabilities; (3) execution parallelism, scheduling a balanced storage-compute workload across devices. We further employ a novel Unified Multi-Level Importance Arbitration to adaptively adjust the critical expert set at runtime, ensuring the accuracy-latency Pareto frontier. These innovations exploit inherent cross-asymmetry, fundamentally breaking the algorithm-hardware misalignment. Experimental results show that RapidMoE achieves up to 3.5x speedup in decoding and 2.1x speedup in prefill compared to state-of-the-art (SOTA) offloading systems.

In global metropolitan areas, public transport benefits from bus systems. Bus design widely applies theoretical models, which typically assume static ridership. However, ridership randomness exists, and lack of attention to it might lead to wrong design decisions. Moving beyond static ridership, we extend the traditional single-line model with stochastic ridership. Further, to optimize the bus system under various ridership randomness, we introduce a hybrid model that combines conventional buses (CBs) and flexible buses (FBs). In both models, we apply a continuous approximation approach and find that: 1) Bus capacity increases with ridership randomness in the extended single-line model; 2) The hybrid model exhibits a binary state: either CBs serve the ridership with rejections under low ridership randomness, or CBs serve the majority while FBs serve the minority with rejections under high ridership randomness. Our findings establish a link between bus system design and ridership randomness, contributing to a more adaptive and efficient public transport framework.

Open-source software (OSS) ecosystems face escalating supply chain threats, including typosquatting, dependency confusion, Trojan Source obfuscation, malicious build injection, and CI/CD pipeline poisoning. Sonatype documented over 1.2 million cumulative malicious packages across major registries, while signature-based and rule-based detection approaches lack the semantic understanding needed to generalize across evolving attacks. This paper proposes a taxonomy-aligned large language model framework for automated detection and multi-class classification of OSS supply chain threats. We introduce an AV-xxx taxonomy covering five semantically distinct attack categories and curate 999 verified real-world incidents from GitHub Security Advisories, CISA alerts, government security databases, and commercial threat intelligence reports, spanning 2018 to 2026 across eight package ecosystems. Using taxonomy-aligned prompting with GPT-4, the framework achieves 97% accuracy and 97% macro F1 across all five categories. Evaluation against eight methods five traditional machine learning baselines, one zero-shot open-source LLM, and two fine-tuned neural models reveals that fine-tuned Llama 3.1 8B (70.5%) and SecRoBERTa (77.5%) underperform a simple TF-IDF classifier (82.3%), while Mistral 7B without taxonomy alignment achieves 65.7%. These findings highlight taxonomy-aligned prompting over model scale, domain pretraining, or fine-tuning as the critical performance factor. Cross-ecosystem analysis shows varying classification difficulty, with npm achieving 65.5% baseline accuracy versus 100% in smaller ecosystems. GitHub Actions exhibits a substantial accuracy F1 divergence (82.4% accuracy, 0.367 macro F1), indicating systematic class imbalance and highlighting ecosystems where LLM-based contextual reasoning offers the greatest benefit. Our dataset, scripts, and code are publicly available in Github.

Timely and specific feedback is one of the strongest influences on student learning, yet it is difficult to sustain in large electrical engineering classes where the ratio of students to demonstrators is high and a learner who is stuck may wait days to find out why an approach was wrong. Generative Artificial Intelligence (GenAI) offers a way to scale conversational feedback, but using it to grade assessed work raises trust and accountability concerns, and keeping a human in the loop to assure its judgements reintroduces the very delay that erodes the value of feedback. The result is a tension between the immediacy that makes feedback so impactful and the human oversight that makes it trustworthy. In this work, we set out to resolve that tension in practice by designing and piloting a GenAI practice platform that delivers immediate, scaffolded feedback during self-directed practice. This relocates human oversight from real-time grading to the upfront verification of solutions. Our goal was to understand how students engaged with the tool, how they perceived the value and reliability of its feedback, and what lessons transfer to other engineering subjects.

In modern radar and Integrated Sensing and Communication (ISAC) systems, the conventional point-scatterer assumption becomes increasingly inaccurate due to higher spatial resolution. In this context, Extended Target Tracking (ETT) has grown interest in recent years. However, most existing ETT works are designed for monostatic systems due to the fine time synchronisation or per-node antenna arrays required for ranging and angle estimation in multistatic scenarios. In contrast, this paper proposes a novel approach to extended target tracking in a distributed antenna system, where only bistatic Doppler frequency measurements are available. Doppler-only tracking is particularly attractive as it avoids the two aforementioned issues. To address this Doppler-only extended target tracking problem, we develop a modified Rao-Blackwellised Particle Filter (RBPF) that replaces the standard Kalman update with a particle-specific Extended Kalman Filter update followed by a relinearisation step, handling the non-linearities of both the rigid extended target dynamics and the observation model. Numerical simulations show the effectiveness of the proposed algorithm for accurate extended target tracking under this Doppler-only setting.

Quadrupedal locomotion requires balancing conflicting objectives such as command tracking, stability, and energy efficiency, yet conventional reinforcement learning (RL) hardcodes these priorities into a fixed scalar reward at training time. We present PROMO (Preference-Conditioned Multi-Objective Reinforcement Learning), a semantic multi-objective approach that makes this trade-off an explicit runtime input to a single locomotion policy. PROMO conditions the policy on deployment facing preferences while keeping embodiment-specific locomotion priors fixed, thereby separating operator intent from reward shaping terms required for viable gait generation. Compared with fixed-objective controllers, multi-objective baselines, and independently trained specialists, PROMO achieves objective specialization and robustness from a single deployable policy. Across 100 sampled preferences in simulation, 67 behaviors are non-dominated under exact Pareto dominance, with a mean preference-objective correlation of 0.843, demonstrating broad Pareto coverage and predictable preference response. The same policy transfers zero-shot to a Unitree Go2, where preference changes alone reduce specific energy by up to 30.4%, position error by 38.7%, and peak body-attitude deviation by 59.0% relative to the balanced preference. These results establish preference-conditioned multi-objective RL as a practical runtime interface for adaptive legged locomotion, extending its role beyond offline Pareto-set construction. Open-source code and videos are available at https://amrmousa.com/promo/.

The advancement of deep learning-based speech synthesis has significantly increased the diversity of deepfake speech, posing threats to voice authentication. While centralized training is effective for deepfake speech detection (DSD), it requires considerable computational resources and raises privacy concerns. To address these issues, we propose a Federated DSD (FedDSD) method that enables collaborative model training across decentralized speech datasets without sharing raw audio. Specifically, each client trains a local model using the FedProx algorithm to mitigate the effects of data heterogeneity and uploads model parameters to a central server. To improve global model aggregation, we further propose a layer-wise center-guided weighting aggregation (L-CGWA) strategy that adjusts each client's contribution per layer based on its distance to a reference center, capturing inter-client and inter-layer discrepancies and enhancing the robustness of model aggregation. Experimental results demonstrate that models trained under the proposed FedDSD method achieve equal error rates (EERs) comparable to those obtained via centralized co-training, while significantly out-performing models trained on individual corpora. Furthermore, the proposed FedDSD method demonstrates robust generalization capabilities across diverse cross-domain datasets.

Autonomous colonoscopic navigation can reduce operator burden and the risk of loop formation or tissue trauma, but remains challenging due to deformable anatomy, weak-texture and specular endoscopic visuals, and contact-rich viscoelastic interactions. Existing methods either rely on geometry-driven pipelines, which are efficient and interpretable yet brittle due to manually engineered features and switching logic, or adopt learning-based policies, whose inferred depth/geometry can become temporally inconsistent or overly smooth under weak texture and specular highlights while simulation-trained variants (e.g., deep reinforcement learning) may further suffer from a sim-to-real gap. We propose ColoACT, an autonomous navigation system that integrates an RGB-D-E based Action Chunking Transformer policy (ColoACT policy) for a compact self-propelled Bevel-Gear-Based Endoscopic Robot (BGER). The ColoACT policy augments RGB with estimated relative depth and a gradient-based pseudo-elevation map to enhance fold-ridge saliency and other high-frequency geometric cues, and enables smooth continuous control of the BGER by predicting overlapping action chunks and fusing them via temporal ensembling. In different ex-vivo porcine colons (approximately 60 cm), our system achieves success rates of 85.4% and 72.5% in straight and curved segments, respectively, and achieves 70% success in 90-degree turns and 60% in double-bend sequences, with feasibility further demonstrated in challenging triple-bend segments. The project page is available at: https://Adamhu1.github.io/ColoACT/.

Scientific progress emerges from a longitudinal ecosystem in which researchers, institutions, funding agencies, collaboration networks, and the scientific literature co-evolve. As AI becomes increasingly involved throughout the scientific research cycle, understanding these interconnected and evolving processes becomes increasingly important. We introduce SciUtopia, a persistent, closed-loop LLM-agent simulation framework for studying academic research ecosystems. SciUtopia models interconnected scientific processes such as research-direction choice, collaboration, submission, peer review, resubmission, citation, funding, and researcher attrition, while maintaining evolving states across simulated years. Its configurable institutional mechanisms and information channels provide a controlled testbed for matched counterfactual experiments and targeted interventions. Across 61 simulation worlds, SciUtopia simulates over 40,000 researchers from 8,000 institutions, producing around 400,000 publication decisions and 1.2 million LLM-generated peer reviews. Using these longitudinal simulations, we find that rejection-driven resubmission substantially amplifies reviewer burden beyond population growth alone, cautious exploration balances citation impact with career success and long-term topic diversity, and resource inequality can emerge even without detectable cumulative advantage from narrowly winning early funding. Code is available at https://github.com/Ahren09/ScienceUtopia.

Errors in LLM agent executions and their visible consequences can be separated by many steps, making decisive-error localization a matter of understanding both step content and step dependencies. We introduce DeFA, a dependency-guided framework for agent failure attribution. DeFA first combines protocol relations and semantic dependencies into an event dependency graph spanning the trajectory. It then identifies events that may violate task requirements and traces their sources and subsequent effects to construct a failure propagation graph. Finally, DeFA uses step evidence and the steps' roles in failure propagation to identify the decisive error, responsible agent, and error category. To support long trajectories, DeFA partitions executions into segments and combines the current segment's detailed content with summaries of the other segments, giving local diagnosis access to global execution context. Across Who and When and the Who and When Pro text subset, DeFA achieves the highest responsible-agent and exact step accuracy with all evaluated backbones, and the highest failure-mode accuracy among taxonomy-aligned methods on Pro. Further experiments on image and video trajectories demonstrate its applicability to multimodal failure attribution. Ablations support the contributions of segmentation, the event dependency graph, and the failure propagation graph. Using DeFA's diagnostic feedback for skill evolution in Trace2Skill improves downstream task accuracy by 6-15 percentage points over the native pipeline, showing that the diagnoses can also support agent improvement on subsequent tasks.

We study strictly stable minimal horizons in three-dimensional time-symmetric charged initial data sets with cosmological constant, allowing charged matter. Along the constant-mean-curvature foliation issuing from the horizon through leaves with positive Jacobi operator, we prove exact first-variation identities for the charged Hawking mass. When the charge density has one sign compatible with the horizon charge, Hawking masses are nondecreasing, and equality between the two boundary values of the fixed-charge mass forces the entire collar to be isometric to an annulus in a source-free constant-curvature Reissner–Nordström–(anti-)de Sitter model. From this we recover the local rigidity results under a local maximality condition. At the minimal endpoint the defect entering these identities computes the fixed-charge Hawking-mass Hessian along the normalized lapse, and its vanishing characterizes the model horizon data.

In this study, we use the Raychaudhuri equation as the essential geometrical tool to reconstruct the functional form of $f(Q)$ gravity within the context of symmetric teleparallel gravity. The expansion scalar and the congruence dynamics recorded in the Raychaudhuri equation are used to relate the non-metricity scalar $Q$ with the kinematical evolution of the Universe in order to establish the reconstruction method. We obtain explicit forms of $f(Q)$ that can explain the observed accelerated expansion of the Universe without the need for an additional dark energy component by taking into account appropriate cosmological assumptions. Moreover, the reconstructed model can also be extended to the infinite past to model the early universe for certain scenarios. In an FLRW spacetime we have considered a constant deceleration parameter and a constant jerk parameter to facilitate our reconstruction analysis. We further discuss the, general relativity limit, Friedmann consistency and background physical viability of the reconstructed models. The analysis indicates that Raychaudhuri equation, interpreted as an equation governing the focusing/defocusing of cosmological geodesic congruences, can be used as the central equation in an inverse reconstruction of $f(Q)$ gravity.

Quantum Reinforcement Learning (QRL) integrates reinforcement learning with parameterized quantum circuits and is a promising approach to combinatorial optimization. On Noisy Intermediate-Scale Quantum (NISQ) devices, however, decoherence, gate imperfections, and measurement errors reduce policy quality and make learning less reliable. Existing error mitigation techniques are generally applied as fixed corrections that do not adapt to changing noise conditions or to the evolving state of training. This work presents Adaptive Policy-Guided Error Mitigation (APGEM) as a context-aware orchestration layer of the hybrid quantum-classical training loop that dynamically selects the most suitable mitigation strategy during QRL training. APGEM evaluates Zero-Noise Extrapolation (ZNE), Probabilistic Error Cancellation (PEC), Clifford Data Regression (CDR), and Readout Error Mitigation (REM) using policy-level indicators, including quantum-state fidelity, policy entropy, cumulative reward, and approximation ratio, and integrates the selected strategy directly into the reinforcement learning loop. The framework is evaluated on the Capacitated Vehicle Routing Problem (CVRP), a representative NP-hard problem in urban logistics, under a range of NISQ noise models and noise levels. APGEM consistently outperforms conventional static mitigation methods, reaches approximately 94% of the utility of an oracle strategy, maintains higher quantum-state fidelity as noise increases, and produces more stable learning behaviour throughout training. Ablation studies show that the framework learns context-aware mitigation policies that adapt to different noise environments and circuit execution conditions. These findings demonstrate that integrating adaptive error mitigation into the learning process substantially improves the robustness and reliability of QRL on NISQ hardware.

Establishing effective spatial correlations between distinct quantum fields is essential for multi-physics quantum sensing.~We here propose a correlation imaging approach based on the hybrid entanglement between microwave photons and microwave surface acoustic wave phonons generated via a superconducting quantum circuit.~At matched microwave frequencies, the five order-of-magnitude wavelength disparity between photons and phonons leads to a unique quantum entanglement resource, which allows the optimal field for probing an object to be distinct from that used for readout.~This resource enables two imaging modalities with giant spatial scaling: a $10^{-8}-10^{-6}$-fold demagnification of macroscopic object profiles onto microscopic phononic chips, and conversely, a $10^{4}-10^{6}$-fold magnification of microscopic object profiles into macroscopic photonic readout devices. By interfacing free-space long-wavelength microwaves with on-chip short-wavelength surface acoustic waves, our correlation imaging approach provides a general framework for cross-scale quantum sensing.

The Sun's atmosphere is significantly influenced by events that often measure only tens of kilometres in the horizontal direction. They happen in both quiet and active regions and are determined by magneto-hydro-dynamic processes. Their origins often lie in the lowest layers of the solar atmosphere. To improve the understanding of chromospheric and coronal heating, the solar dynamo and its activity cycle, or to improve the prediction of space weather events we must therefore enhance our understanding of such small-scale events. In this work I first describe the process of calibrating spectropolarimetric observations with very high spatial resolution and put together the building blocks for the mathematical proof the MOMFBD image reconstruction algorithm. In the second part I describe the FISS-SP prototype which is designed to fully exploit the resolution power of a 1.6m telescope. This instrument is built to provide suitable data for image reconstruction of spectropolarimetric data. The resulting data can resolve small-scale features down to the diffraction limit. The simultaneous interpretation of a few lines is still the norm for spatially highly resolved solar spectra. Theoretical works have shown that the simultaneous interpretation of multiple solar lines can greatly improve the robustness against noise. Using observations from FISS-SP I advance the theoretical idea into a practical technique. With this I can retrieve a coherent model of the local solar atmosphere, resolving structures down to a horizontal size of about 49~km. Finally, I use those new techniques to interpret an observation of a solar pore. In this last part I describe and analyse small-scale structures, which could not be spectropolarimetrically resolved before. Typical pore-boundary features - namely, dark striations with bright grains moving into the pore - are identified as convective features.

System calls (syscalls) record key interactions between running programs and the operating system kernel, providing fine-grained and minimally intrusive data for host-based intrusion detection systems (HIDS) deployed in cloud and other modern computing environments. However, existing methods often model syscalls in their original execution order, where sequences from different processes are interleaved, making informative patterns difficult to extract and raising two questions: whether raw syscall sequences can be reorganized in a way that yields more discriminative representations, and how complex attack patterns can be effectively learned from the reorganized sequences. We propose ReSHID, a resource-aware behavior reconstruction and hierarchical semantic learning framework for host intrusion detection. It reconstructs semantically continuous sequences by leveraging syscall semantic invariants to cast subject identity and relationship resolution across PID namespaces as a bipartite matching problem and tracking file descriptor (FD) lifecycles to associate descriptors referring to the same resource. Additionally, features extracted from these sequences are organized into a lightweight subject behavior graph incorporating inter-subject relationships, where GATv2 captures key coordination patterns to model complex attacks involving multiple subjects. Experimental results show that sequence reconstruction combined with the detection method can improve HIDS performance. Even with a lightweight linear classifier, the proposed method achieves the best results among all compared methods in terms of F1-score (98.64%), ROC-AUC (99.80%), and PR-AUC (98.10%), while reducing the number of n-gram features by approximately 75.2% and 44.1% compared with the raw sequences and MGFE, respectively.