Papers

Task planning for robots in off-site construction prefabrication is primarily manual, leading to planning inefficiencies and reluctance to adopt robotic solutions. Existing approaches provide limited support for scalable, modular automated task planning and localized runtime replanning. This paper proposes APTree, a model-driven framework that integrates automated task planning with modular task execution using behavior trees as the execution model. APTree repeatedly applies and coordinates heterogeneous symbolic planners to automatically generate and compose modular task plans at runtime. To support this approach, we develop a domain-specific language for the formal specification and validation of task plans, together with a compatible execution engine. We evaluate APTree on the real-world livMatS Biomimetic Shell pavilion, where it models the construction robotic fabrication domain, detects injected specification and planning errors at design time and runtime, and automatically generates large-scale task plans by coordinating three symbolic planners across numerous planning modules. A machine-checked formal analysis further shows that composed execution and local replanning are safe under explicitly stated assumptions, including planner soundness and non-interference of generated plans, providing practitioners with explicit conditions for dependable plan composition. APTree thus enables construction practitioners to leverage automated task planning without requiring detailed knowledge of symbolic planning techniques or APTree's internal implementation.

Reusable skills let LLM agents load task-specific instructions on demand, but they make execution depend on implicit routing over skill metadata. We test whether this design scales by holding SkillsBench tasks and their golden skills fixed while adding irrelevant skills. We find skill routing collapse: agents increasingly load fewer or no skills, even though the relevant skills remain available. With 1,000 irrelevant skills, loaded-skill recall falls from 73.6% to 27.4% and the zero-load rate rises from 9.3% to 51.6%, whereas explicit selection from the same library retains 73.9% recall. Loading other skills, including plausible alternatives to golden skills, is too rare to explain this decline. Masking only the descriptions of irrelevant skills largely restores skill loading, yet with the metadata intact, explicitly asking agents to plan skill use still yields lower recall than preselection. These findings point to semantic interference from descriptions and instruction interference when skill choice competes with task execution. We also evaluate explicit preselection as a practical mitigation, including its coverage and latency trade-offs. Our analysis informs skill protocols and evaluation for agents with growing skill libraries.

Modifications of gravity beyond (vacuum) General Relativity have a subtle impact on the physics around black holes. Describing these effects across different observables is, however, a priori strongly model dependent. While effective descriptions of the geometry allow to efficiently capture deviations from benchmark predictions, they sometimes lack a clear passage towards (effective) new physics. Inspired by linear response theory, in this work I therefore propose a formalism that allows to calculate leading corrections of observables relative to the Schwarzschild black hole anywhere in space-time (outside of the horizon): the order of corrections is fixed through an eigenvalue $\mathcal{E}$ of the Einstein tensor, which is identified with an external energy momentum tensor that generates the deformation of the geometry. In this ways, $\mathcal{E}$ can be linked to effective physics models. Focusing on spherically symmetric and static black holes, I illustrate the approach at the hand of a number of observables in the vicinity of the horizon, such as characteristic quantities of particles moving around the black hole (e.g. the black hole shadow or the radius of the innermost stable circular orbit) or the spectrum of scalar quasi-normal modes. Reversing the logic, I argue that the knowledge of such observables provides non-trivial constraints on $\mathcal{E}$, which in turn grant insights into the effective physics in the vicinity of the black hole.

Autoregressive language models produce, at every prediction step, a vocabulary-sized logit vector whose structure determines the conditional next-token distribution. We empirically identify that the ordered vocabulary-head logits follow a remarkably consistent log-linear relation \[ z_{(r)} \approx \mu - \alpha \log r. \] across prediction contexts, architectures, and scales, emerging progressively during pretraining. To our knowledge, this is the first systematic analysis of the logit geometry underlying the distributional assumptions used in the prior decoding work. While controlled order-statistic baselines spanning bounded, light-tailed, and heavier-tailed distributions can largely explain the observed log-linear geometry, they do not explain the systematic shift in the slope $\alpha$: trained language models exhibit $\alpha \approx 1.4$–$1.7$, sitting well above the initialization baseline ($\alpha \approx 0.2$) . Moreover, $\alpha$ correlates with predictive performance, suggesting that it captures learned structure beyond generic extreme-value effects.

We report continuous-wave (CW) vacuum-ultraviolet (VUV) generation at 148.4 nm by second-harmonic generation of 296.8-nm fundamental light in quasi-phase matched (QPM) BaMgF4 (BMF). At the incident power of 170 mW, a VUV output power of approximately 31 pW can be obtained. The quadratic dependence of the signal on the incident power in the QPM region and its attenuation in the ambient environment support the identification of VUV generation. The polarization dependence shows a sine-function shape with a 15% output attenuation from the maximum to the minimum. These results highlight the potential of BMF as a nonlinear medium for CW VUV generation.

Four-dimensional cone-beam computed tomography (4D CBCT) provides phase-resolved volumetric images for image-guided radiotherapy. Yet good reconstruction quality usually needs many projections. This increases scan time and radiation dose. We propose STCNet, a sparse-view 4D CBCT reconstruction framework that uses spatial-temporal consistency learning. STCNet reconstructs high-quality images from only 16 projections in each respiratory phase. STCNet integrates physics-driven positional encoding to enforce geometric consistency in the projection domain and introduces temporal encoding in latent space to capture respiratory motion dynamics across phases to capture dynamic respiratory motion across phases. A dual-domain collaborative strategy jointly optimizes both projection and image domains, enhancing structural fidelity and temporal coherence. We tested STCNet on the TCIA-LUNG dataset and a CIRS thorax phantom dataset. STCNet performed better than conventional methods and recent deep learning baselines in image quality, artifact reduction, and structural preservation. On TCIA-LUNG, STCNet reached PSNR values of 33.311, 32.695, and 32.658 dB at phases 2, 5, and 8. The corresponding SSIM values were 0.845, 0.823, and 0.825. On CIRS, STCNet achieved PSNR values up to 36.497 dB and SSIM values up to 0.877. These results support STCNet as a promising framework for low-dose, phase-resolved 4D CBCT reconstruction in image-guided radiotherapy.

Knowledge distillation trains student models on teacher-generated responses to transfer task capability, but those responses may also carry behavioral tendencies of their source. We investigate whether the source teacher can shape student behavior beyond the capability that distillation is intended to transfer. We construct harm-accepting and harm-avoiding variants of the same base teacher, use them to generate responses to ordinary reasoning and knowledge questions unrelated to safety, and distill those responses into students with different base models. Students generally acquire task capability while their downstream decisions shift toward the source teacher's safety-pragmatism disposition, a phenomenon we call imprinting. Distillation from the harm-accepting teacher reduces harm avoidance, whereas distillation from the harm-avoiding teacher increases it. Imprinting varies across teacher-student pairs and does not extend uniformly across different forms of safety behavior. Excluding a sparse set of divergence tokens reduces imprinting, whereas response rewriting preserves the source-dependent effect and can introduce additional rewriter-dependent shifts. Post-distillation preference optimization raises harm avoidance but does not eliminate the source-dependent difference. These results show that behavioral tendencies can be intentionally imprinted through the source teacher, even with ordinary task supervision.

Post-training pruning enables efficient deployment, but conventional methods do not explicitly preserve internal-gradient alignment between sparse and dense models. We study novel connections between downstream performance of compressed models and the ability to retain gradient information via influence-function preservation. Central to this connection is a curvature-aware objective that accounts for both layer activations and backpropagated gradients. Despite its promise, this objective remains underexplored in pruning due to a lack of efficient solvers. We introduce $\texttt{CACTUS}$, a novel ADMM-based optimizer supporting both pruning and sparse-plus-low-rank decompositions of pre-trained weights. Compared to LLM-Surgeon which targets the same objective, $\texttt{CACTUS}$ can achieve over an order-of-magnitude speedup. At $50\%$ pruning of Qwen3-4B, $\texttt{CACTUS}$ reduces the C4 perplexity gap to the dense model by $45\%$, improves compressed-to-dense per-sample gradient cosine similarity on WikiText-2 from $0.670$ to $0.755$, and runs over $28\times$ faster than LLM-Surgeon on substantially weaker hardware. By preserving influence functions, $\texttt{CACTUS}$ supports effective data selection. On Llama-3.2-3B, fine-tuning a fresh dense model on the $5\%$ of data selected by $\texttt{CACTUS}$ achieves $49.44\%$ accuracy on BIG-Bench Hard, compared with $47.22\%$ using ALPS and $49.54\%$ using dense-model selection. Finally, integrating $\texttt{CACTUS}$ with LoGRA, a recent gradient-projection strategy, on Qwen3-4B with MLP layers pruned to 2:4 sparsity and a rank-64 correction retains a $0.711$ cosine alignment with dense LoGRA-projected C4 gradients.

Large language models (LLMs) increasingly act as personal assistants; however, little is known about how they capture national-level preferences on a global scale. Given that "time" is a universal and representative concept, we examine LLMs' time preferences by varying only users' country and measuring the compensation LLMs deem appropriate for a 60-minute appointment delay. We then compare LLMs' responses with GDP per capita on Purchasing Power Parity (GDP_PPP), Mean Income, and a literature-based index (Patience Score). Our results across 193 countries and 5 LLMs demonstrate that people's time across the world is evaluated in noticeably different ways among LLMs, as well as varying levels of robustness under prompt variations. LLMs consistently recommend the highest compensation for Europe, but mostly the lowest for Africa. Based on GDP_PPP, LLMs underestimate the value of time for users living in Asia and South America. These findings indicate that geographical patterns are embedded in LLMs and may translate in different treatment in agentic settings.

Motivated by applications to cognitive radio networks, multi-player multi-armed bandits (MMAB) have attracted considerable attention in recent years. The asynchronous variant, in which players are active only at a fraction of timesteps, better reflects real-world decentralized networks, where devices are not permanently active. Despite its practical relevance, this setting remains relatively understudied. Existing work on asynchronous MMAB has focused on homogeneous players, who share the same activation probability, a setting in which the optimal centralized allocation can be easily characterized. In contrast, efficiently computing the optimal allocation in the heterogeneous setting, where players have different activation probabilities, has remained an open problem. In this work, we investigate this problem and first establish its NP-hardness. We then turn to the greedy algorithm, which runs in polynomial time, and prove that it achieves a $\frac{1}{e}$-approximation of the optimal allocation. Furthermore, when players are processed in decreasing order of their activation probabilities, the approximation guarantee improves to $\frac{1}{\sqrt{e}}$. Overall, our results establish the greedy algorithm as an efficient approximation to the optimal allocation, providing a practical oracle baseline for the design of online algorithms for heterogeneous asynchronous multi-player bandits.

Sampling from high-dimensional, heavy-tailed distributions is a fundamental challenge in computational statistics, as many standard Markov chain Monte Carlo (MCMC) methods mix poorly in such settings. Recently, Stereographic MCMC [Yang et al., 2024] and the Sub-Cauchy Projection Sampler [Grazzi et al., 2026] have been shown to perform successfully on such tasks. However, establishing their non-asymptotic convergence properties remains an important open problem. In this work, we fill this gap by establishing non-asymptotic upper bounds on the mixing time of the stereographic projection and sub-Cauchy projection samplers. Our results demonstrate that, under certain conditions on the target and initial distributions, the mixing time is polynomial in dimension for a broad class of distributions, including light- and heavy-tailed cases. Motivated by the theoretical analysis, we further establish a new weighted isoperimetric inequality that extends the classical version for log-concave distributions to the heavy-tailed setting with optimal dimension dependence. The proof techniques provide new insights into the geometric properties of heavy-tailed distributions that govern rapid mixing in high dimensions.

The modular data of the doubles of the Haagerup–Izumi categories is a grafting of two finite metric groups (conjectured by Evans–Gannon, proved for untwisted categories by Gannon–Schopieray–Yadav). Gang and Kim found non-unitary modular data of this shape in the topological twist of rank-$0$ $S$-fold theories, from the Bethe vacua of an abelianised gauge theory. The aim is to understand rigorously the significance of the handle-gluing and fibering operators, and to construct the Bethe vacua. For every hyperbolic once-punctured-torus monodromy $\varphi$, a word in $L$ and $R$, the two groups are $A_\mp=\coker(\varphi\mp1)$, with the torsion linking forms of the mapping tori of $\varphi$ and $-\varphi$. Over all words of trace $m^2+2$ these forms reproduce the Evans–Gannon table of Q-systems, one printed entry being corrected, and the distinct modular data in the list of Gannon–Schopieray–Yadav. The Bethe data are given in closed form by the word: the level matrix, the positive vacuum, $c_{\eff}=\#L$, and the handle-gluing values $2^{\#L}\lambda^{\pm1}|A_-||A_+|$ at the identity classes, $\lambda$ the stretch factor. Each kernel vector of $\varphi\mp1$ modulo $|A_\mp|$ at which an explicit trigonometric formula does not vanish gives a solution of the Bethe equations in closed form. Under the half-index dictionary, with the unit in the class of $\sigma V^*$, $c\equiv\#L-6\#R\pmod{24}$.

Buildings cause nearly 40% of carbon emissions in the US. Of total building emissions, HVAC systems are responsible for approximately 40-60% and account for 15% of global energy consumption. Despite advances in artificial intelligence, most buildings still operate as inefficiently as they did decades ago. Building optimization, perhaps more so than any other optimization problem, has enormous potential for Machine Learning (ML) driven impact on climate change, in both the short and long term. One of the biggest obstacles for building optimization is the lack of policy transfer. Most building control systems, whether simulation or learning based, don't generalize: they only work well for the building they were developed for, but not for other buildings. To solve this problem, we propose a challenge for the development of building foundation models that easily adapt to new buildings with minimal configuration and thereby enable building optimization to scale across the globe. Participants will build control policies on a set of simulated buildings, and have their policies evaluated on an unseen set of buildings. The buildings vary in size, layout, weather, material properties, and HVAC system. Successful outcomes from this initiative promise a massive leap forward in global sustainability efforts and demonstrate the great potential of ML to combat climate change.

Timely assessment of depression severity in adolescents is essential for guiding clinical care, yet conventional evaluations rely on episodic clinical assessments and symptom rating scales. Contactless gait analysis offers a potentially scalable approach to capturing objective behavioral features associated with depression. Here, we introduce the Long-Distance 3D Adolescent Depression Gait Dataset (LD3D-ADG), comprising depth-sensor gait recordings from 119 adolescents at a single center, including 39 healthy controls, 41 with mild depression, and 39 with moderate-to-severe depression. We developed UMLPA-GCN, a graph convolutional network that models spatiotemporal gait dynamics across three anatomically defined skeletal partitions. In participant-independent five-fold cross-validation, UMLPA-GCN achieved 75.63% classification accuracy, a macro-F1 score of 75.52%, and a macro-area under the receiver operating characteristic curve of 0.839 (95% confidence interval, 0.779–0.899). Class-specific recall was 87.18%, 63.41%, and 76.92% for healthy controls, mild depression, and moderate-to-severe depression, respectively. At the participant level, reduced arm swing, slower pelvic movement, restricted joint range of motion, and greater arm-swing asymmetry were identified as leading potential digital biomarkers contributing to depression severity assessment, associated with higher 17-item Hamilton Depression Rating Scale scores. These findings provide proof-of-concept evidence for contactless digital depression severity stratification in adolescents. Further external and prospective validation is needed to establish their utility for longitudinal monitoring and timely clinical assessment.

Trends with stellar elemental abundances and condensation temperature have been proposed as a possible tracer of rocky planet formation. Here we present an alternative explanation, where we show that stellar activity (star spots) can also cause apparent abundance changes in refractory elements preferentially over volatile elements. This effect arises naturally from atomic physics: refractory (rocky; high-$T_c$) elements preferentially have low first-ionization potentials, with strong transitions that are most sensitive to the cooler, magnetically-perturbed spot conditions, making any spot-induced change in line strength correlated with $T_c$. We re-examine literature on Sun-like stars in the Solar neighborhood and find that the $\log{R'_{HK}}$ stellar activity indicator is correlated with the so-called $T_c$ slope: the gradient of chemical abundances with respect to condensation temperature. Using 1D LTE Korg.jl simulations of the Sun with spots of varying size and contrast, we show that a wide range of spot configurations can manifest as abundance patterns that mimic a signal of abundance trends with condensation temperature. Finally, we conduct a line-by-line differential analysis of an ambient Solar spectrum relative to a Solar sunspot umbra to represent an extreme case, where we find that elements with high condensation temperature can show a factor of three (0.5 dex) increase. We discuss implications that stellar activity poses for precision spectroscopy, and provide some direction on how this effect might be mitigated.

The safety of LLM-based agents currently rests on alignment training and safety classifiers, neither of which can rule out that a sophisticated indirect prompt injection (IPI) goes through and causes significant harm. Recent work therefore builds agents with information-flow guarantees that rely on structured tool interfaces to interact with the environment. Computer Use Agents (CUAs) act visually on the computer screen, which does not provide a boundary between trusted first-party interfaces and untrusted, potentially attacker-authored content. Actions are equally ambiguous: a click may land on an application toolbar or a button controlled by an attacker. Treating the user as the source of trust and confirming each sensitive decision produces confirmation fatigue, with the risk of missing harmful actions. We introduce SA-CUA, a dual-LLM framework that generates a reliable trust boundary for CUAs by deriving conservative deterministic rules from the accessibility tree for unmasking trusted regions from the privileged LLM's screenshot, keeping trusted and untrusted content separated. Human interaction is only required for proper task specification in the beginning. During the task a robust endorsement channel allows values to pass from the untrusted to the trusted domain by verifying information via trusted channels. This allows the agent to autonomously handle tasks involving untrusted information. SA-CUA maintains much higher utility than CaMeL-CUA while keeping the attack success rate at zero.

Clinical trial protocols serve as the source of truth for study design, but are mostly written and used as unstructured PDF documents. Digitizing protocols turns static content into structured, machine-readable data that clinical software can analyze, validate, and execute. Coding agents, which select tools, execute code, and revise their approach as they work, are increasingly applied to such long-document tasks, yet it remains unclear how reliably they turn complete documents into accurate structured outputs. We study this question using protocol digitization as a testbed. We introduce TrialPD-Bench, an expert-annotated benchmark of 150 publicly available protocol PDFs spanning four task categories. Beyond final-answer performance, our evaluation measures evidence recall from agent traces, agent behavior, and cost. Across nine proprietary and open-weight models in five coding-agent harnesses, even the strongest agents do not fully solve protocol digitization. Our analysis further shows that agents usually locate the required evidence but fail in the last mile of transcribing and structuring it. We further develop TrialScribe, a domain-specific harness that equips coding agents with task skills and a verification loop. TrialScribe improves a locally served open-weight model by 6.2-59.4 points across eight metrics, enabling it to match or surpass proprietary models such as GPT-5.6-Sol.

Multiplayer social dilemmas on hypergraphs show a variety of emergent phenomena, which has been attributed to group interactions. Here we test whether pairwise dynamics can reproduce them as well. We derive the mesoscopic Langevin equation of a testbed model and find the operator content of absorbing phase transitions: a directed-percolation line, a tricritical point, and a first-order region whose spinodal is the explosive transition. In our analysis, three-player games only set the values of the Landau coefficients. We construct a strictly pairwise model with short-lived memory, in which every event involves at most two entities, that reproduces these coefficients term by term and has exactly the same mean-field fixed points. Simulations confirm that the two dynamics share mean-field critical and tricritical scaling and the size of the critical mass, and that a purely three-player game on a two-dimensional triangular lattice lies in the directed-percolation class. For these games, the explosive transition and the critical mass therefore do not require group interactions, and the critical behavior is that of the pairwise theory.

We review the accretion-driven channel of stellar black-hole growth and spin-up in gaseous star clusters, developed within the framework of our MSCA project "protoBH". Our detailed calculations involving semi-analytic modelling, dynamical and hydrodynamic simulations suggest that the growth and spin-up of initially low-mass stellar black holes is an integral element of the formation process of dense massive star clusters, such as the proto-globular clusters observed at high redshift. The black holes accrete residual gas, remaining from the first star formation event, and form accretion disks generated by gas angular momentum induced by the black-holes' own curved orbits. Assuming initial black hole masses below the pair-instability mass gap $\leq 55\,{\rm M}_{\odot}$, the gap gets populated with black-hole masses growing to $\sim 10^2\,{\rm M}_{\odot}$, and even reaching several hundred solar masses for sufficiently dense and massive proto-clusters. We predict a consistent spin-mass relation with low black hole spins $\lesssim 0.2$ for low black hole masses $\lesssim 20\,{\rm M}_{\odot}$, spanning all intermediate spins for higher masses and saturating to high spins $\gtrsim 0.8$ for high masses $\gtrsim 70\,{\rm M}_{\odot}$. We find evidence in favour of such an accretion-driven subpopulation of black holes in the GWTC-5.0 released by the LIGO-Virgo-KAGRA collaboration.

Recently, Morris et al. introduced TinyLoRA, a strategy for LLM fine-tuning which introduces arbitrarily few parameters, and demonstrated accuracy improvements using less than $10$ parameters rather than the ${\gtrsim 10,000}$ typically demanded by LoRA-based methods. However, that study continued to use gradient-based local search methods of the kind deployed for training multi-billion parameter LLMs in the first place. This article is predicated on the observation that, in the low dimensional setting, global optimization methods become tractable, unlocking Large language Model Bayesian Optimization (LlMBO). LlMBO is sample-efficient, gradient-free, memory-light, parallelizes across devices without synchronization, and, unlike policy-gradient methods, can optimize greedy-decoding accuracy directly. We describe numerical experiments comparing LlMBO against GRPO on fine-tuning tasks for a variety of models on a math benchmark suite which show that LlMBO can be a suitable alternative in this regime, matching or exceeding GRPO's fine tuning gains, while using a fraction of the compute and memory.

We construct non-holomorphic Siegel modular forms whose coefficients encode linking numbers in arithmetic hyperbolic manifolds of simplest type. Our proofs rely on an improved understanding of the transgression forms appearing in Kudla–Millson theory, and along the way we establish some structural properties including in particular their expected current equations.

Two-level overlapping Schwarz preconditioners with Generalized Dryja-Smith-Widlund (GDSW) and Reduced GDSW (RGDSW) coarse spaces for high-order finite element methods (FEM) and virtual element methods (VEM) are investigated and applied to the Poisson problem in 2D. The GDSW coarse space is a well-established robust choice for preconditioners in domain decomposition, while the RGDSW coarse space offers a competitive alternative by reducing the coarse problem dimension while maintaining robustness. On the theoretical side, a condition number bound for both preconditioners is proved for conforming $\mathbb{P}_k$ and $\mathbb{Q}_k$ finite element spaces with explicit dependence on the polynomial degree $k$. Through numerical experiments on structured triangular and quadrilateral meshes for FEM and on polygonal meshes for VEM, it is demonstrated that both GDSW and RGDSW maintain scalability properties across polynomial orders up to $k=9$ for selected degrees. The RGDSW approach achieves a reduction in the coarse problem dimension by a factor of approximately three compared to GDSW while maintaining competitive iteration counts.

Social norms are a fundamental component of society, shaping how we communicate, cooperate, and avoid conflict. Despite their ubiquitous presence and profound influence on behaviour, questions remain about how they spread from person to person and whether they are adopted. Using an evolutionary game-theoretic model, we study whether simple or complex contagion processes characterise how social norms spread. Crucially, contagion type is endogenous and emerges from the model. We also vary cultural conformism, the cost of following a norm, and network topology to assess the generality of our results. We show that social norms in our model can spread through simple contagion when a new norm is low in cost and cultural conformism is low, while norms spread through complex contagion when the new norm is costly and cultural conformism is high.

We develop a macroscopic fluctuation theory for diffusive classical lattice gases subject to continuous coarse-grained monitoring. Treating measurements in the typical-outcome (quenched) regime, we show that observation adds a quadratic confining term to the MFT action that pins the density to its hydrodynamic profile. We derive sufficient conditions under which, in the presence of monitoring, the current large-deviation function remains convex, the additivity principle holds, and dynamical phase transitions are excluded. For particle-hole-symmetric one-dimensional models we find that monitoring suppresses dynamical phase transitions and can change their order from second to first, as we demonstrate in the Katz–Lebowitz–Spohn model. In the strong-measurement limit, we find that the current distribution becomes Gaussian and determine the scaling of its cumulants: they follow simple power laws when the mobility remains nonzero throughout the system, but exhibit anomalous logarithmic scaling when the mobility vanishes at the boundaries.

Egocentric video understanding spans diverse capabilities, modalities, and evaluation formats, yet these are often studied in isolation despite their tight coupling in real-world interaction. We propose the Ego-Action-World interaction loop as a unifying principle: the wearer perceives the world, forms goals, acts, changes the world, and uses feedback to guide subsequent behavior. Based on this view, we introduce EgoOmniWorld, a benchmark organized around four complementary interaction structures: Node captures ego, action, and world states; Edge captures directed relations among them; Chain captures linked dependencies; and Loop captures feedback-closed interactions. We develop an interaction-first, evidence-grounded data pipeline that extracts structured interactions from egocentric videos, grounds them in visual and auditory evidence, and generates questions according to their interaction structure. Required modalities and single- or multi-turn formats follow naturally from the supporting evidence and dependencies. EgoOmniWorld contains 24,166 QA groups, including 20,284 training and 3,882 test samples. Experiments on representative open- and closed-source multimodal large language models reveal substantial and uneven limitations across interaction structures. Models trained on EgoOmniWorld improve in-domain interaction understanding and generalize to existing egocentric benchmarks. EgoOmniWorld provides a unified foundation for evaluating and advancing egocentric video understanding. Data, code, and models will be released.

Autoformalization (AF) requires more than compilation: formal statements must preserve the meaning of their informal sources, a property known as faithfulness. An emerging frontier is the automated evaluation of this property, yet most existing methods operate at the statement level and produce whole-item verdicts, failing to reveal how individual expressions correspond or which contextual interpretations support the judgment. This omission is consequential because AF is inherently theory-scale: target statements derive their meaning from surrounding mathematical material and unstated conventions. We formulate faithfulness evaluation as a context-aware align-then-label problem: identifying phrase-level correspondences between informal and formal statements and assigning a faithfulness label to each under explicitly specified contextual interpretations and assumptions. We also introduce FaithAlign, the first theory-scale faithfulness benchmark comprising 460 items and 4,982 annotated phrase-level links, alongside FaithAlign-Gap, a companion set of 146 items with informal context and assumptions withheld. Each FaithAlign item includes the target statement pair, relevant context, formal dependencies, and, where needed, assumptions specifying one given interpretation. Across 8 models, GPT-6 Astra achieves the highest defect F1 at 47.8% and identifies all defects in only 4.5% of items with two or more defects. On FaithAlign-Gap, Astra identifies only 28.2% of undecidable items as such. These results show the difficulty of recovering context-dependent correspondences, establishing FaithAlign as a testbed for fine-grained, context-aware faithfulness evaluation.

Tool-call judges evaluate argument values proposed by LLM agents. We find that they accept a correct value less often when their input holds no reference for it, that is, no text that says what the value should be. This happens when the value is a default value from the function schema that the user did not mention, when it has to be inferred, or when the text it comes from was cut from the judge input. We call this drop in acceptance the derivation gap. It appears across all seven judges we evaluate, including System One models CLM and Jev. The gap is not only a truncation artifact: correct default values have the largest gap on a benchmark with no truncation. A controlled source-relocation test shows that judges follow the reference in their input: they favor a visible incorrect value over a correct value whose source is hidden. We introduce DeRef (Derived Reference): a separate model reads the full history without seeing the candidate and writes a reference for the judge. Across three benchmarks, DeRef improves overall argument AUROC for all judges tested. On the controlled test, it raises AUROC from chance or below to between 98.40 and 99.59.

This work provides a framework to treat optimal control problems on separable Banach spaces in the viscosity sense. First, it extends the notion of $B$-viscosity solutions to the Banach setting and relates the extension to the notions of viscosity solutions used in metric space settings. As the main results, the paper proves a strict comparison principle and characterizes the value function of an optimal control problem as the unique viscosity solution to an infinite-dimensional HJB equation.

In order to execute quantum algorithms on bosonic quantum computers, it is necessary to encode their inputs into a desired target quantum state, with a process called state preparation. In real settings, state preparation methods operate in stochastic scenarios, where the prepared state is represented as a mixed state, that is a probabilistic ensemble of pure states. Therefore, existing state preparation methods optimize a variational ansatz by averaging over all possible sequences of photon loss events contained in the prepared ensemble, yielding an output state which approximates the desired target with a certain average fidelity. Such an approach, however, does not take into account directly the distribution of fidelities across states in the ensemble. In this work, we aim to maximize the fidelity of only a subset of pure states sampled from the prepared ensemble, at the cost of reducing the average fidelity overall. This trade-off is relevant when the goal is to obtain high-fidelity states with a certain probability, rather than maximizing the average fidelity of the prepared ensemble. We borrow and adapt objective functions from risk-sensitive optimization, a framework that optimizes specific characteristics of a distribution rather than only its expectation, and that has been successfully applied in Reinforcement Learning. In the considered experimental settings, our results show that the infidelities of some pure states in the prepared ensemble can be reduced by up to four orders of magnitude when using risk-seeking objectives, compared to those achieved using the standard optimization of the average fidelity. Furthermore, we show that it is possible to control the trade-off between the probability of obtaining high-fidelity states and the lowest infidelity threshold reached, providing the flexibility to select the operating point that best suits a certain application.

Objective digital measures of depression severity could complement conventional symptom assessments and facilitate more timely clinical evaluation. Here, we investigated whether task-evoked eye movements provide interpretable behavioral markers of depression severity in adolescents and young adults. In this single-center, cross-sectional study, 155 participants aged 14–20 years completed emotional-spatial free-viewing and memory-guided recall tasks and were stratified by 17-item Hamilton Depression Rating Scale scores. We examined 19 prespecified eye-movement features across six task-informed domains using covariate-adjusted ordinal-trend models. Greater depression severity was associated with lower gaze-allocation entropy (standardized $\beta=-0.599$), fewer inter-area switches ($\beta=-0.526$), longer first target-entry latency ($\beta=0.572$), and lower target dwell proportion ($\beta=-0.678$; all false-discovery-rate-adjusted $q<0.001$). Emotion-specific attentional-bias measures were not associated with severity, whereas associations involving fixation dispersion weakened after adjustment for anxiety symptoms. In a complementary machine-learning analysis incorporating 44 candidate eye-movement features and demographic covariates, a three-stage cascade classifier achieved the highest four-class stratification performance (accuracy, 0.665; macro-F1, 0.670; area under the receiver operating characteristic curve, 0.869), outperforming single-stage and two-stage approaches. Together, these findings identify altered visual exploration and memory-guided gaze control as candidate digital biomarkers of depression severity, rather than generalized negative attentional bias. These results support further prospective evaluation of task-evoked eye tracking as a complementary approach for timely severity assessment and clinical decision support.