Papers

This work proposes a semidefinite programming (SDP) model that synthesizes static output control feedback for networked linear time-invariant multi-agent systems (MAS). We theoretically prove that the output feedback control of MAS can be constructed using the Nash equilibrium previously derived for state-feedback control. The Nash equilibrium considers that agents seek to minimize their individual cost functions subject to the system dynamics of the MAS. The resulting control from the pro- posed SDP model ensures the stability of the closed-loop system and dissipativity when a disturbance is present in the MAS. A set of four case studies is implemented. The first three numerical cases serve to validate and compare cost effectiveness and time-domain performance of the proposed control framework with other state-of-the-art frameworks. The fourth case designs a frequency regulation scheme for a modified IEEE 24-bus power grid. Using the proposed framework, distributed controllers are synthesized for the inverters which are considered agents in the grid. Each inverter uses partial state information of the neighboring inverters. The resulting control scheme is implemented in a high-fidelity non-linear model of 261 states that describes the electromagnetic transient dynamics of the power grid. Hence, compared to prior studies that usually validate their frameworks in small-scale numerical models, this fourth case demonstrates a virtue of our pro- posed SDP: its scalability and practical utility in real-world networked MAS. We implement the proposed framework in an open-access toolbox in Python and MATLAB, enabling practitioners and researchers to readily implement it in their applications.

Agents are increasingly given open-ended research tasks: discovering an empirical law from self-designed experiments, improving a heuristic whose optimum nobody knows, or beating a standing record. Their execution logs record every step of this research, yet the runs are still judged by their outcome score. That score alone does not establish whether an agent's claims follow from executed experiments, and a reference answer may be unavailable. We evaluate the agent's epistemic process: how it forms hypotheses, tests them, and revises them in response to evidence. We introduce OEB (Open-Endedness Bench), a benchmark-agnostic methodology that reads only the agent's execution record and never a reference answer or an outcome score. OEB compiles the record into a unified epistemic event graph whose edges connect the propositions the agent states to the executed actions that test them; each node carries an exact excerpt that code verifies against the record. One principle governs scoring: prose can state a proposition, but only evidence returned by an executed action can support or refute it, so OEB checks what the agent writes against what it actually ran. From the graph, OEB scores four competence axes (evidence, experiment, revision, and no reward hacking), mostly as the share of opportunities for sound research that the agent took, and profiles six subjective persona traits that describe the agent's research habits. We score 119 existing runs over 12 tasks from three benchmarks: LLM post-training, chip design, and a training-speed record. Against logged results, only 16-29% of the improvements agents claim are real. On 9 of 10 tasks, the best run tries more new ideas in its second half than the worst run. The persona readings follow the model: for every trait, the model that ran explains more of its variance across runs than the task (a median of 43% against 7%).

Elusive dwarf galaxy groups provide a rare opportunity to study hierarchical galaxy assembly at the lowest-mass scales, but whether such associations are gravitationally bound systems or transient configurations remains uncertain. We present the largest catalog of isolated dwarf galaxy groups, identified using the DESI Survey, comprising 74 systems with three to six members per group. The groups have a median line-of-sight velocity dispersion of $\sim53$ km s$^{-1}$ and a median crossing time of $\sim0.37$ Gyr. We find that stellar mass alone is generally insufficient to bind the observed configurations. On the other hand, when dark matter is included using independently inferred stellar-to-halo mass relations and NFW potentials, 69 of 74 systems ($93\%$) favor a bound interpretation after propagating observational and halo-model uncertainties. For the 69 bound-favored systems, dynamical-friction estimates indicate that substantial orbital decay within a Hubble time is plausible if their members retain significant dark-matter subhalos. Our results suggest that many of these groups are systems caught during the assembly of more massive descendants, representing some of the lowest-mass sites of ongoing hierarchical galaxy growth in the nearby Universe and accessible laboratories for a mode of assembly expected to have been prevalent at early cosmic epochs.

A typed decision model answers a fixed question about an input by returning a probability for each of several caller-defined options. Each option carries a short label and a written definition, which is where a developer states the rule the model should apply. Jev introduced this interface for routing, moderation and triage, open implementations followed, and the same operation occurs whenever a language model is used as a classifier by scoring label strings. We study the open implementations, whose weights we can inspect and patch, and ask whether the probability follows the definitions or the labels. A preference for the label we call option-label bias. Across four open-weight typed decision models, three ways of reading an answer from a Qwen2.5 backbone, eleven classification tasks and PolicyBench, a synthetic routing suite we introduce in which the rule appears only in the definitions, the answer is mostly the labels. Deleting every definition leaves accuracy unchanged (laya-td: 0.8559 against 0.8487), although those definitions support 0.7971 on their own, and renaming the options to A and B raises accuracy by +0.1511 [+0.1377, +0.1646]. One system, von, is unaffected, and the two code bases differ in one expression: laya writes each option as "{label}: {definition}", while von writes only the definition. Changing that expression in both directions, with no weight changed, makes all three laya checkpoints exactly invariant (+0.0000 [+0.0000, +0.0000]) and creates the effect in von, whose accuracy falls from 0.8511 to 0.2281 when a label contradicts its definition. Earlier work attributed this failure to the constrained decision head these models use in place of a text decoder; our results locate it in the prompt rendering. We give a two-call test that tells a practitioner which case applies to their model, and measure what four mitigations are worth.

The off-diagonal Gram map for h vectors in an M-dimensional quadratic space is flat with reduced fibers of rational singularities (FRS) in characteristic zero if and only if M > max_{2 <= r <= h} (2h - r + 1 - 2h/r). Equivalently, this gives the exact FRS convolution threshold of x |-> (x_i x_j){i < j}. For its zero-fiber ideal I{M,h}, we also prove lct(I_{M,h}) = min{ binom(h,2), min_{0 <= k <= h-2} ( binom(k+1,2) + M(h-k)/2 ) }. In the strict range, all truncated coefficient maps have geometrically integral complete-intersection fibers and are faithfully flat over Z[1/2], uniformly over graphs on h vertices. The proof uses a sharp weighted estimate for hollow symmetric matrices over finite local principal ideal rings, paired with isotropic lower bounds. Within the established point-counting approach to rational singularities, this estimate also gives an explicit modulus of continuity for Gram-map densities over every local field of odd residue characteristic.

Sparse Mixture-of-Experts (MoE) models combine learned routing with selection from a large expert pool. We isolate the contribution of dynamic expert combination using a dense analogue: $K$ SwiGLU experts, all active for every token and combined by a softmax gate, at fixed total FFN width. With no larger pool or discrete selection, the dense baseline is the $K=1$ case. Validation loss varies non-monotonically with $K$: $K=2$ improves over the baseline by $0.0048$, whereas $K=4$ and $K=6$ worsen it by $0.0053$ and $0.0197$, respectively. Routing generally remains soft and expert usage balanced, except in the first layer of the $K=4$ model, where routing is nearly one-hot. Forcing this gate to uniform increases loss by $2.4$ nats on a diagnostic subset, indicating that its concentrated routing is functionally important. We further show that the architecture is exactly a dense SwiGLU with token-dependent, simplex-constrained group scaling and contains the dense baseline in its function class. These single-run results suggest a granularity sweet spot for always-active, softly gated FFNs at fixed width, with $K=2$ performing best among the configurations tested.

Dwarf-dwarf interactions are a major driver of galaxy evolution in the early Universe, but observational constraints on their impact across mass, separation, and environment remain scarce. Using the Dark Energy Spectroscopic Instrument Early Data Release, this study constructs the largest sample of dwarf galaxy pairs, spanning stellar masses $6.5 \leq \log(M_\star/M_\odot) \leq 9.5$ and redshifts $0.015 \leq z \leq 1.5$. We analyze the evolution of star formation rates and multiwavelength colors ($NUV-r$, $U$-$V$, $V$-$J$, $r$-$K$, and $g$-$r$) as a function of projected separation, stellar mass, and external tidal environment. Low-mass/low-redshift pairs exhibit interaction-triggered star formation at intermediate separations. At the smallest separations ($\lesssim$10 kpc), these systems display divergent behavior: isolated pairs show the onset of dust-obscured star formation, those embedded in weak external tidal fields undergo extended, dust-unobscured star formation, while those embedded in strong external tidal fields experience rapid quenching of star formation. High-mass/low-redshift pairs follow a more uniform evolutionary track, with generally weaker dependence on environment. Interaction-triggered star formation occurs at intermediate separations, followed by increased dust obscuration and rapid quenching at the smallest separations. At higher redshifts, high-mass dwarf systems display an initial suppression of star formation, followed by renewed activity at smaller separations, consistent with early-stage gas removal and later interaction-driven inflows. These results demonstrate that dwarf-dwarf interactions do not follow a single evolutionary pathway. Instead, we reveal a complex picture of dwarf galaxy evolution, with significant implications for dust-obscured star formation, abrupt quenching, and the formation of compact, dusty, intensely star-forming systems.

Speakers can differ not only in how much silence they produce but also in where interruptions fall, information that global pause counts obscure. We analyzed 115 publicly available Speech Accent Archive recordings of the same read passage: 57 speakers whose archive metadata listed Mandarin as their first language and a mainland-China birthplace, and 58 American English first-language speakers born in the U.S. Midwest. Word-level forced alignment was used to extract interword acoustic gaps and classify each position as punctuation-marked or unpunctuated. At a 250 ms threshold, punctuation-marked gaps were similar for the Mandarin and English groups (means 5.56 and 5.47), whereas unpunctuated-position gaps were more frequent in the Mandarin group (means 3.67 and 0.43; adjusted rate ratio 11.98, 95% confidence interval 6.53-22.00). A crossed-effects logistic model with random intercepts for speakers and passage positions confirmed that the contrast was disproportionately concentrated at unpunctuated positions (interaction odds ratio 9.44, 95% confidence interval 5.23-17.02). The pattern persisted at a 500 ms threshold, after removing positions adjacent to punctuation, after excluding severe transcript-deviation cases, and when analysis was restricted to acoustically confirmed silence. Nearly half of the Mandarin group's unpunctuated gaps occurred at positions classified as within-phrase in an exploratory single-coder annotation. In this fixed-passage corpus, acoustic gap placement relative to textual and syntactic structure distinguished the groups more clearly than punctuation-marked pausing. The results support placement-sensitive measurement as a reproducible dimension of breakdown fluency while not establishing a causal mechanism, proficiency difference, or perceptual consequence.

Metropolitan concentration has motivated policies for more balanced regional development. We use agent-based simulation to examine trajectory diversity, the variety of people's recurring activity orientations toward neighborhoods, urban centers, and wider networks. The model maps 2010 population shares to 2026 distributions in five country cases, China, Russia, Japan, the United Kingdom, and the United States, and simulates change to 2036. Trajectory diversity declines by 52.6-79.0% over 2026-2036. Russia and Japan show the largest increases in the share of the most common trajectory, while US diversity contracts despite a slight decline in this share. Cost burdens rise throughout, whereas welfare falls in the Russian and Japanese cases but rises in the other three. These findings suggest that regional development should be assessed through activity diversity, cost burdens, and welfare together.

Physical AI Smart Spaces is, to the best of our knowledge, the first benchmark to simultaneously provide large-scale, multi-class, and multi-camera 3D perception data for indoor smart spaces. It contains over 280 hours of synchronized 1080p footage captured by nearly 1,800 cameras in warehouses, hospitals, retail venues, and similar settings, together with automatic annotations for multi-camera identities, 2D bounding boxes, 3D bounding boxes, camera calibration, and depth where available. The benchmark spans Isaac Sim synthetic generation, Cosmos Transfer appearance augmentation, and real-world Sim2Real evaluation. For the real-world target, we include two warehouse deployments with time-synchronized streams, automatic VGGT-based calibration, and a 3D labeling interface that projects world-frame 3D boxes into each view for cross-camera verification. We describe the dataset scope, annotation and calibration schema, generation workflow, benchmark protocols, and official evaluation system, which standardizes submission format, and leaderboard reporting. A central contribution is a 3D instantiation of Higher Order Tracking Accuracy (HOTA), extending the usual 2D box-based tracking evaluation to 3D locations and 3D boxes. We further report empirical baselines from the AI City Challenge leaderboards, showing how methods evolve from person-only 3D location tracking to multi-class 3D box tracking under realistic smart-space constraints. The release is available at https://huggingface.co/datasets/nvidia/PhysicalAI-SmartSpaces.

In this paper, we provide a novel approach to limiting Heesch numbers of a specific family of tiles. In particular, we show how a certain family of three–dimensional prisms has a finite Heesch number. In so doing, we initiate a structural study of ratio $q$ tilings by two prototiles, in which, loosely speaking, one tile appears $q$ times more often than the other. The method works as follows: we take planar intersections with coronas around the prism in two perpendicular directions and compare the frequencies of the resulting prototiles in those patches. If every tiling by these prototiles has limiting ratio at least $q >1$, then sufficiently large coronas of the prism would require incompatible tile counts in the two directions. We also provide some computational results and exhibit $6$ infinite families of notched prisms that have finite Heesch numbers greater or equal to $2$. Finally, we provide some negative computational results about the non–existence of certain prisms with given Heesch numbers.

To commit to buying external data or participate in collaborative learning, one must decide whether the additional data will improve prediction enough to justify the cost. This comes with several challenges: (i) the decision often relies only on aggregated statistics available publicly, rather than individual-level data; (ii) covariate and model shifts can induce negative transfer, so the additional data deteriorates rather than improves performance; (iii) if the data is sensitive, its privatization requires the injection of noise, which can also offset the benefit of a larger sample size. In this paper, we model the problem of dataset selection through high-dimensional regression with multiple heterogeneous sources and a weighted ridge estimator. Our approach uses only summary statistics and it gives privacy guarantees either on labels only or jointly on features and labels, in terms of $\rho$-zero-concentrated differential privacy. The main technical contribution is a deterministic equivalent of the test error, which captures the interactions between sample size, covariance structure, model shift, regularization and privacy noise. Our theory allows to optimize hyperparameters (weights and ridge regularizers) and, more broadly, to decide when private external datasets are useful without accessing the data itself but only relying on population-level quantities. This provides a theoretically tractable foundation for private transfer learning, which we support via experiments on both synthetic and real-world datasets.

Most published neural-receiver studies train on synthetic channel models, although real over-the-air measurements can improve performance. We examine how much of this benefit transfers from other sites to a distinctly different deployment site and what adaptation to that site costs elsewhere. The study uses a dataset from a 5.88 GHz measurement campaign covering eight sites and four environment types. An industrial site measured entirely from a vehicle is withheld as a deliberate transfer stress test. One receiver architecture is trained under four regimes: simulation alone; seven other sites; all sites jointly; and other-site training followed by deployment-site finetuning. The last two use identical measurements through different optimization protocols. Across five independently trained seeds, sensitivity gains are evaluated at uncoded bit error rate (BER) targets of $10^{-1}$ and $10^{-2}$, representative of moderate- and higher-rate coded links, respectively. On the deployment site, training with other-site data rather than simulation improves sensitivity by 0.3 dB at $10^{-1}$ and 0.1 dB at $10^{-2}$. Adding deployment-site data contributes less than 0.1 dB at $10^{-1}$; at $10^{-2}$, total gains over simulation reach 0.4 dB with joint training and 0.6 dB with finetuning. Joint training shows no measurable aggregate loss on the other-site test set, whereas finetuning yields the larger deployment-site gain but reduces other-site sensitivity by about 0.1 dB at $10^{-2}$, indicating spatial forgetting.

Systematic reviews condense clinical trials into evidence tables, yet clinicians can interrogate these tables only through database queries, and many questions concern attributes that the table does not record, such as a drug's target class or a harmonised endpoint. Here we introduce FD-SCoPE, a language-model framework that answers both kinds of question, exposes the query, the selected trials and the derivation rule behind every answer, and learns from expert corrections. On an oncology evidence table of 159 immune checkpoint inhibitor trial records, FD-SCoPE completed all 140 clinician-style tasks (alternatives, 90.7-97.9%). For questions needing derived attributes it retrieved 99.3% of relevant trial records at a positive predictive value of 89.8% and outperformed four alternative approaches (derived-value F1 77.7% versus 64.8-73.4%). Corrections on 299 questions, simulated from reference answers, raised F1 on 1,201 unseen questions from 77.9% to 84.9%. Language models coupled with executable queries, verified programs and expert feedback can give clinicians auditable access to trial evidence.

Realistic models of polarized Galactic dust emission at high angular resolution are essential for component separation in next-generation cosmic microwave background (CMB) $B$-mode experiments. However, existing dust polarization templates are constrained only to ~1 degrees, and it remains challenging to develop models that can generate realistic small-scale complexity that is also coherently connected to the well-measured large-scale structures. In this work, we develop a data-driven super-resolution framework that predicts small-scale dust polarization structure directly from multiple tracers observed at high resolution. We combine the Planck dust optical depth $\tau_{353}$, \ion{H}{1}-based Stokes templates, and coarse Planck GNILC $Q$ and $U$ maps in a transformer-based model trained to recover native-resolution GNILC polarization from a $4\times$ beam-smoothed version of itself, so that the network solves a deconvolution problem informed by the structure of multiple tracers rather than generating structure to match a target statistic. Applying the model to the uniform-resolution GNILC maps, we produce $Q$ and $U$ predictions at 20' over the high-latitude (|b|>30 degrees) sky. Using scalar and tensorial Minkowski functionals and scattering transform statistics, we find that our maps exhibit more coherent, anisotropic, and filamentary small-scale structure than the latest PySM dust model. Our results establish data-driven learned deconvolution as a promising route to foreground models whose small scales are inferred from data rather than prescribed.

High-fidelity solubility prediction is fundamental to pharmaceutical development and environmental partitioning, where accurate modeling must couple molecular structure with thermodynamic behavior across diverse chemical environments. However, recent advancements have been dominated by deep learning architectures that often sacrifice physical interpretability for predictive power. We challenge this trend by showing that state-of-the-art performance does not require such non-transparent architectures. To address this, we introduce DISSOLVR, a transparent framework for molecular solubility prediction. In addition, we perform a comprehensive literature review and a benchmarking study against various methods. We show that DISSOLVR approaches the aleatoric limit of experimental uncertainty and achieves OOD generalization through structural invariance, derived by mapping molecules to physically-grounded descriptors. Then, we present an LLM-assisted post-hoc explanation pipeline that bridges the gap between symbolic model artifacts and chemically grounded narratives. Finally, a comparative benchmark of a survey involving 22 expert chemists reveals that expert evaluators provide deep insights.

In this paper, we consider the control of Markov jump linear systems when the active system mode is not exactly known. We rather assume that this true mode is only known to belong to an observed cluster of modes, making both the dynamics and the associated transition probabilities uncertain. We construct a min-max optimal control problem in this setting, with the objective of robust regulation. Using results in regularized least mean-square optimization, we describe how the solution of this problem can be reduced to a Riccati recursion, such that the controller depends only on knowledge of which cluster the mode lies in. We subsequently establish convergence of the recursion and stability of the resulting closed-loop system. A numerical example illustrates the effectiveness of the proposed methodology.

Recent work on neural sparse retrieval has demonstrated strong relevance by leveraging Large Language Models (LLMs) for semantic term expansion. However, learned models paired with previous sparsification techniques still yield overly long document and query vectors partly due to a large LLM vocabulary, imposing a serious challenge to retrieval time and space efficiency. This paper proposes a scheme for optimizing model sparsity through a synergy of adaptive strategies, including learnable soft top-K, per-term thresholding, and FLOPs regularization to increase the sparsity of query and document vectors. Experimental results with Lion-SP model on the MS MARCO and BEIR datasets demonstrate that the proposed scheme can outperform the baselines by significantly reducing the average query and document lengths. Our scheme can achieve much shorter retrieval latency and lower storage cost while maintaining highly competitive relevance.

As AI-assisted programming becomes increasingly mainstream, the environmental impact of AI-generated software has emerged as an important consideration. This motivates evaluating LLM-generated code beyond functional correctness by considering execution efficiency and energy consumption. However, despite substantial advances in code generation, frontier LLMs are rarely evaluated based on the energy efficiency of the code they produce. In this work, we conduct a comprehensive evaluation of 21 prompting strategies for energy-efficient code generation and identify 8 strategies for evaluation across 10 widely used open-weight and proprietary LLMs. We evaluate their effectiveness for both Python and C++ code generation relative to a baseline prompt. Across the evaluated models, the selected prompting strategies achieved energy reductions of up to 25% for Python and 17% for C++ code generation. At the model level, Python energy reductions reached up to 50% for Granite-4.0-H-Small, 39% for Claude 4.5 Haiku, and 28% for MiniMax M3, while C++ reductions reached up to 57% for Granite-4.0-H-Small and 6.4% for Qwen3-Coder-480B-A35B-Instruct. These results demonstrate that prompting strategies can substantially influence the energy consumption of LLM-generated code, although their effectiveness varies across models and programming languages. Our findings highlight the importance of incorporating energy efficiency into the evaluation and optimization of LLM-based code generation and provide practical insights into designing prompts for more sustainable AI-assisted programming.

We derive a weighted Kirchhoff diffusion on a smooth Riemannian open book as a scaling limit of regime-switching Brownian motions on boundaryless manifolds. The pages are compact Riemannian manifolds sharing a common boundary manifold and inducing the same metric on the binding, while their second fundamental forms and interior geometries may differ. For each $\varepsilon>0$, we regularize the metric in an $o(\varepsilon)$ collar, pass to smooth doubles of the pages, run Brownian motion on each double, and allow the page label to switch in a collar of width $\varepsilon$ at rates \(q_\varepsilon(r)Q_{ij},\) where $Q$ is an arbitrary irreducible finite-state Markov generator with invariant law $\pi$, while \(q_\eps\) is bounded and supported on an \(\eps\) collar. Denoting \(\Theta_\eps\) the total mass of \(q_\eps\) and \(\Xi_\eps\) its second raw moment, we prove that, if \(\Theta_\varepsilon\to\infty\) and \(\Xi_\varepsilon\to0\), then the glued processes converge, from every deterministic sequence of starting points, to Brownian motion with weigthed Kirchhoff interface conditions. No reversibility of $Q$, pointwise scaling ansatz for $q_\varepsilon$, or first-moment condition is required. Furthermore, we show that the second fundamental form of the pages contributes no additional interface term: curvature remains only through the bulk Laplace–Beltrami operators.

Coding agents have become increasingly long-horizon, autonomous, reliant on general-purpose shell and maintain their own persistent memory for self-improvement. While these capabilities have made the agents powerful, they have also made them harder to defend against external adversaries. Defenses that restrict this architecture — typed tools, information-flow control, or policy prediction engines — give up too much functionality to be adopted. Agents deployed today (e.g. Claude, Codex) rely on a combination of user-mediated and automode sandboxing as their primary defense. In user-mediated sandboxing, user-maintained policies decay over time and repeated permission requests cause user fatigue, while auto mode's tool-call classifiers learn no user-specific policy and are not meant to defend against adversarial setups. Pincer is a new defense that operates at the resource layer and works alongside existing defenses at the tool-call layer like the auto mode. At the core of Pincer lies a digital twin, an isolated-context model that automatically learns and enforces dynamic user-specific least-privilege policies. The digital twin keeps continually learning the user's preferences allowing it to act as the user's proxy for the agent's permission requests. To emulate the learning phase, we propose a new usercentric dataset with examples following a multi-day transcript of user-agent interaction. Our evaluation shows that Pincer performs strongly on both security and utility in comparison to several baselines which includes variants of LLM judges and adaptations of Conseca (HotOS '25). We highlight attack types where Pincer's design leads to a significant security improvement compared to all other baselines, while outperforming the baselines even for other types of attacks.

Personal advice, including relationship advice, now ranks among the most common uses of generative AI. But language models (LMs) exhibit sycophancy: they affirm users much more often than humans do, which can make people overconfident and less willing to repair their relationships after a conflict. Prior work on mitigating sycophancy has focused on factual settings where a response can be checked against a ground truth answer, while mitigations for social sycophancy (e.g., personal advice, where there is no ground truth) have relied on simple prompting and post-training methods with limited effectiveness. Our insight is that social sycophancy occurs in part because LMs overly center on the user and fail to consider the perspectives of other stakeholders impacted by the user's behavior. To address this problem we propose Pluralistic Preference Optimization (PlurPO): given inputs describing interpersonal conflicts, the LM identifies and simulates the relevant stakeholders, and is then trained to prefer and generate responses acceptable to all stakeholders. PlurPO uses only signals the model produces about its own outputs, without ground-truth labels. PlurPO substantially reduces social sycophancy across four datasets and four model families compared to prior methods. For example, on statements of intent to cause harm, where the users' actions should not be endorsed, PlurPO reduces the endorsement rate by 89% on average across four models. On general advice questions, where the target is to match the endorsement rate of human responses, it closes the gap by more than half, from 17.8% to 8.0% on average. The preference dataset constructed by PlurPO for an 8B model also effectively transfers to mitigating sycophancy in a larger (32B) model. Our results indicate that social sycophancy can be reduced by leveraging a model's own capabilities to simulate a plurality of relevant perspectives.

Diffusion transformers (DiTs) generate high-fidelity images from text and image conditions, but their outputs carry large variance and their faithfulness to a desired target depends heavily on how the condition is supplied. We present DAGS, a lightweight, attention-free, disentangled appearance and geometry conditioning scheme that steers a frozen image DiT to produce high-fidelity, highly faithful, and independently controllable renders. Two small convolutional encoders compute conditioning features once per frame and inject them as a learned, per-layer, element-wise residual into the image tokens, avoiding the quadratic cost of stacking conditions through attention. Because control and temporal handling live outside the frozen backbone, we retain its vast pretrained prior and eliminate backbone-overfitting risk. We further add a small recurrent lighting stabilizer and a training-free temporal guidance term that, coupled with our conditioning, elevate a per-frame image model into a streaming renderer. DAGS produces controllable, high-quality renders at a fraction of the compute of path tracing; it is not real-time, trading compute for controllability and quality. On a matched 1-spp + G-buffer input, per-frame DAGS reconstructs +8.6 dB / +10.1 dB PSNR over the real-time denoiser Intel OIDN and the diffusion renderer RGB<->X while being 2.5-8x more temporally stable perceptually (temporal-LPIPS flicker).

Many physical and biological systems -foams, membranes, elastic shells, or tissue sheets- can be mathematically described as 2D surfaces. These surfaces can host complex dynamics, like reaction-diffusion systems, and undergo drastic deformations. Here, we present triangulax, an open-source Python library for simulations and geometry processing that discretizes surfaces using triangular meshes available at https://github.com/nikolas-claussen/triangulax. It combines two features. First, discrete differential geometry provides coordinate-free discretization, which allows simulating strongly deforming surfaces. Second, the library is built on the machine-learning framework JAX, so mesh-based energies and entire simulations can be differentiated automatically (including simulations with stochastic forces or topological modifications). This simplifies multi-physics simulations by computing forces automatically and allows fitting model parameters to experimental data or design objectives. We implement an algorithm to preserve mesh quality in strongly deforming surfaces and apply it to a model of membrane mechanics coupled to an on-membrane concentration field. We validate our approach against exact solutions and existing software, and demonstrate triangulax across a range of problems, including membrane mechanics, vertex models, and inverse design of a shape-morphing elastomer sheet.

Constraint satisfaction problems (CSPs) with operator assignments to the variables provide a well-structured setting to study the decision complexity of the entangled value of classes of nonlocal games. Due to the CSP dichotomy theorem, the complexity of constraint satisfaction problems with classical assignments can be fully understood by studying the symmetries of the CSP, in terms of the polymorphisms of the underlying relational structure. In this work, we show that the polymorphism-based reductions between CSPs can be generalised to gap-preserving reductions between entangled CSPs based on an entangled analogue of the polymorphisms. This reduction allows us to show undecidability of entangled graph colouring with more than three colours, a problem that has proved resistant to prior hardness reductions based on commutativity gadgets.

We develop a closure modeling framework for identifying missing components of dynamical systems using Kernel Ridge Regression (KRR). The framework addresses two classes of closure problems: difference-equation closures arising in ODE and PDE settings, and algebraic closures arising from moment closure in kinetic equations. For the first class, we derive an error bound in an ODE setting that quantifies contributions from time integration, approximation of unresolved scales, and interpolation required to couple unresolved-scale effects to the resolved solver. Numerical experiments on the Lorenz-63 system and the Kuramoto-Sivashinsky equation demonstrate accurate long-horizon predictions and substantial improvements over an LSTM-based closure model. For the second class, we consider moment closure for a one-dimensional kinetic equation by modeling discrepancies between kinetic and macroscopic fluxes as a function of the resolved macroscopic variables. We compare global KRR models based on PCA coordinates with spatially local models. While the global model performs well for unimodal initial conditions, its accuracy deteriorates for bimodal initial conditions. Spatially local models with appropriate modeling inputs improve robustness and achieve higher predictive accuracy.

We present OpenGameEval, a benchmark and evaluation framework for agentic game development inside Roblox Studio. It runs language models as agents in reproducible, stateful game-engine sessions and scores each run with executable checks, both on the edited scene and in a simulated play session. Most agentic coding benchmarks require exploration but score only final task success. OpenGameEval separates observation tools from editing tools in its eight-tool action space, so exploration can be measured directly. We measure the pass rates and exploration behavior of 13 frontier models on 84 human-curated core tasks, with 16 attempts per task. The tasks are hard for current models. The best model solves 51.7% of tasks on a single attempt and 39.4% five times out of five, and no tested model solves six of the tasks. Models at the frontier reach similar pass rates by solving different tasks: splitting tasks by the kind of work they require spreads the top five by 5.0pp on script-authoring tasks and 12.5pp on scene-change tasks. Exploration behavior predicts whether a run succeeds. Holding task and model fixed, a run that inspects every object a reference solution touches before acting on it passes 13.4pp more often than a run that inspects none of them on scene-only tasks, and 9.8pp more often on script-only tasks. We release the task suite, its place files, the per-task annotations, a plugin that runs the tasks inside Roblox Studio, and an updated leaderboard under the MIT license at https://github.com/Roblox/open-game-eval.

Science-fiction interfaces are often reproduced for their appearance, and their diagnostic logic is rarely rebuilt around real data. Here we describe Santiago's A.T. Field, an interactive visualization of urban accessibility built from a diagnostic interface in Episode 13 of Neon Genesis Evangelion. We translated that interface in five stages: phenomenon, field, progression, interface, and response. A microscopic Angel became a relational urban problem, and a fictional scanning lattice became an accessibility field computed with the E2SFCA method. Every element of the fictional display is tied to a quantity of the model. Three tensions appeared when we made the fictional language operational: the saturated palette of the genre against color-vision deficiency, the motion of the genre against the reduced-motion preference, and the per-frame cost of the animation. The urgency of the fictional command center has its own cost, because emergency aesthetics can cast urban populations as threats. These three tensions are what the translation cost, and they are what transfers to other attempts of this kind.

Camera-based physiological monitoring can produce multiple estimates from several facial regions, extraction methods, and processing settings. Signal quality indices aim to select reliable estimates without a physiological reference, and their potential is often assessed with an oracle that selects the estimate closest to the reference in each window. This retrospective selection can reward chance agreement. We model the effect with order statistics. For K independent candidates unrelated to the reference, the expected oracle error decreases approximately as 1/K. We analyze thermal heart rate estimation on 96 iBVP recordings with 168 candidates per 10 s window. The oracle achieves a mean absolute error of 0.91 bpm, compared with 10.74 bpm for the best fixed configuration, 18.03 bpm for the best quality index, and 8.61 bpm for a constant predictor. With K = 24, a forehead signal from another recording matches the correct one, with 4.62 against 4.61 bpm. Oracle evaluations should report candidate count, valid coverage, and matched null controls.

Combinatorial interaction testing (CIT) is a black-box testing method that has received extensive attention in both research and practice over recent years. Its primary objective is to construct an effective combinatorial test suite that detects software failures caused by parameter interactions. As a fundamental component of the CIT testing process, the evaluation metric plays a critical role in assessing and comparing combinatorial test suites, as well as in evaluating various test generation techniques. For CIT practitioners, selecting an appropriate evaluation metric is both important and challenging, given the wide variety of available options. Nevertheless, no prior work has systematically addressed this problem. To fill this gap, this paper first provides a comprehensive survey of black-box evaluation metrics for combinatorial test suites, offering rigorous definitions, clear classifications, illustrative examples, and complexity analyses. We then conduct an extensive empirical study involving eight open-source projects, encompassing 32 test scenarios and 295,624 combinatorial test suites. In this study, we examine the correlation between each static evaluation metric and fault-detection effectiveness using two correlation measures. Experimental results show that the Value Combination Coverage (VCC) metric serves as a valid predictor for test-suite evaluation. However, distribution-based metrics generally incur lower computational costs than interaction coverage-based ones. The choice of an appropriate metric should also account for the test suite's inherent properties, as these characteristics can substantially influence the effectiveness of the evaluation. Finally, we provide practical guidelines to assist CIT practitioners in selecting suitable evaluation metrics for assessing or comparing combinatorial test suites.