Boyang Zhao, Yifan Zhou, Alban Puech, Tilman Bockhacker, Hendrik Hamann
原文摘要
Grid Foundation Models (GridFMs) are emerging as a powerful framework for learning and transferring knowledge across power grids and grid analysis tasks. This work extends GridFMs from steady-state analysis to dynamic simulation. We present DynGridFM, which uses component-wise tokenization and a structured latent operator to learn shared dynamic representations across heterogeneous grids and enable transfer to unseen systems. Pretrained on nine grids, DynGridFM is adapted to a held-out 48-machine grid, where fine-tuning on 64 trajectories achieves accuracy comparable to training from scratch on 128. On a single GPU, it achieves 924x to 1232x speedup over numerical simulators on one CPU core.
Sota Ogawa, Shojiro Yamabe, Daiki Nishiyama, Jun Sakuma
原文摘要
Fine-tuned (FT) derivatives of large language models (LLMs) are now released in large numbers, yet their provenance relies on self-reports and cannot be verified. Misattributed lineage leaves a verifier unable to audit outsourced fine-tuning, confirm license compliance of derived models, or trace vulnerabilities that a parent passes to its children. All three require the direction of derivation, not merely its presence. Existing output-only methods are limited to pairwise derivation tests or undirected lineage estimation, while recovering a directed tree requires white-box access to model weights. We formalize the problem of reconstructing a directed model tree, including internal nodes, under gray-box access, where only teacher-forced log-likelihoods are observed and weights are never inspected. Fine-tuning raises the log-likelihood of the training data, and a child inherits this increase from its parent. Building on this property, we define an asymmetric per-task log-likelihood deviation between any two models as a trace of fine-tuning history, and derive from it an edge weight for each ordered pair of models. We propose TRACE-Edge, which exactly solves for the maximum spanning directed tree under these edge weights, with neither hyperparameters nor tree-structure constraints. We evaluate it on five configurations totaling 25 model trees built from three base models and seven tasks. TRACE-Edge achieves a parent-match accuracy of 0.74 when probed with the fine-tuning examples, and on average outperforms an output-similarity baseline even with held-out probes drawn from the same distribution as the training data.
Assistive robots and search-and-rescue platforms require reliable human state sensing in smoke, darkness, and privacy-sensitive environments, where cameras and LiDAR may be unreliable Millimeter-wave radar offers an alternative that works without ambient light, penetrates smoke, and does not capture visual appearance. However, the compact arrays that fit robot size, power, and cost budgets have limited apertures, producing joint-localization errors on the order of decimeters at a few meters, large enough to compromise the safety margins and tracking accuracy a robot depends on. Existing pipelines reduce pose error through frame fusion, multiple viewpoints, return selection, or stronger estimators, but examine one technique at a time, without relating its effect to a shared measurement limit or separating gains from added information from gains from estimator completion. We address this gap with a framework that models the human and environment as one time-varying scattering field and states a Cramer-Rao type error floor for joint localization as a function of bandwidth, aperture, signal-to-noise ratio, fused frame count, phase coherence, and visibility of retained returns. We then use existing frame fusion, multi viewpoint, return selection, and pose estimation methods as matched probes of this floor rather than as competing pipelines. Under the evaluated conditions, experiments on three datasets and a live deployment show these strategies move joint localization error in the predicted directions, for example lowering mean per-joint position error from about 97 to 65 mm with a second viewpoint, while a stronger estimator mainly lowers overall error by completing joints the measurement cannot resolve. These results give the robotics community a principled basis for deciding whether to prioritize observation design or estimator capacity when deploying reliable mmWave human sensing.
Post-training of language models is commonly framed as a sample-score-update loop implemented by gradient descent. A recent line of work, exemplified by RandOpt, relocates this loop to weight space, sampling Gaussian perturbations around a pretrained model and ensembling the top-K rewarded specialists at inference. While competitive with PPO and GRPO under matched training compute, this prediction-level ensemble incurs K forward passes per test example and does not extend cleanly to free-form generation. We ask whether the rewarded population can instead be folded into a single deployable model, replacing the inference-time ensemble with one consolidated update. A split-half analysis over 25 model-task pairs reveals reproducible low-rank structure in every case. We turn this geometry into CoRP (Consolidating Rewarded Perturbations), a gradient-free operator that combines reward-weighted aggregation, compatibility-aware reweighting, and a held-out validation gate, with no gradient flowing through the language model. Across five language models from 0.5B to 8B and five tasks covering math, code, and creative writing, CoRP improves the base model by 8.1 points on average. Using one tenth of RandOpt's perturbation budget, CoRP exceeds single-inference RandOpt by 6.5 points and recovers more than half of the gain of the 50-pass majority-vote ensemble, at one forward pass per test example.
Chenliang Zhou, Yifan Wang, Andrzej Banburski-Fahey
原文摘要
We present Align3D, an image-conditioned diffusion model that generates 3D indoor layouts that are both faithful to a single input photograph and physically plausible beyond the observed view. From the photograph we infer a confidence-calibrated scene graph and train a set-diffusion layout denoiser with image-structure co-alignment where confidence-weighted modality fusion balances image and scene-graph alignment using predicted confidences, while observability-aware fidelity-preserving guidance projects conflicting scene-quality gradients and reweighs them with an object-level confidence mask so under-observed degrees of freedom absorb physical priors without undoing the image-graph match. On 3D-FRONT, Align3D improves image-alignment and scene-quality metrics over recent baselines; once trained, it maps previously unseen photographs to layouts in a single feed-forward pass.
Large language model (LLM) agents are increasingly deployed as personal assistants. Existing evaluations, however, mostly use short, self-contained requests in static environments. Everyday life assistance is different. A task runs for weeks rather than minutes. The world keeps changing while the agent is not being prompted. Many constraints are never stated outright. An agent that merely answers the request in front of it will fail at such a task. What is needed instead is an agent that stays proactive and consistent. It decides on its own when to act, when to ask, and when to stay silent. It notices changes that nobody announced. It keeps one plan coherent from the first day to the last. No current benchmark measures this. We introduce VibeLifeBench, a benchmark of 200 long-horizon tasks across ten everyday-life domains. Each task is a scripted multi-week timeline in a simulated world of 22 mock services. The world advances on its own clock, and many of its changes are silent, so only an agent that re-inspects the world discovers them. Every task is graded by fine-grained, weighted checks that read only what the agent actually left behind, covering the end state, the timeliness of its actions, and whether it upheld the implicit constraints. We evaluate seven frontier models. All of them score low, which shows how far current agents are from assisting with real life. We will open-source all tasks, environments, and the evaluation framework.
Simulation-Based Inference (SBI) serves as a vital framework for parameter inference in scientific fields where simulators involve intractable likelihoods, yet while amortized generative models offer rapid posterior estimation, they are often restricted by the specific priors used during training, thereby limiting their flexibility as prior knowledge evolves. To address this prior dependency, PriorGuide was introduced as an inference-time guidance method, but due to its intractable formulation, it relies on Gaussian approximations of the reverse transition kernel and Gaussian mixture model fitting for the prior ratio, both of which introduce systematic bias. Motivated by these limitations, we propose an unbiased test-time guidance framework that leverages Density Ratio Estimation (DRE) to learn a score guidance term, effectively decoupling the inference process from the prior training. Moreover, our framework remains agnostic to the specific density ratio estimators, making it a general and flexible framework for handling prior changes. Experimental results across multiple tasks demonstrate that our method matches or outperforms PriorGuide on C2ST and MMD in most tasks while maintaining robustness even under limited overlap between the training and target priors. Furthermore, we apply our method to Bayesian updating for parameter inference from planetary light-curve data, where it also demonstrates strong effectiveness and robustness. Code is available at https://github.com/a-chenchen/dre-based-prior-guidance .
Mojdeh Saadati, Juan Panelo, Gustavo Visentini, Soumik Sarkar, Carlos Messina, Baskar Ganapathysubramanian
原文摘要
Climate change threatens maize yields worldwide, yet identifying which genetic traits sustain productivity under future conditions requires exploring an enormous space of genotype, soil, and climate combinations that process-based crop models are too costly to simulate at scale. Here we develop an integrated probabilistic emulation framework that transforms a mechanistic crop simulator (APSIM) into a fast, uncertainty-aware discovery engine. Built from two million simulations and coupled to a generative weather model that produces realistic climate sequences, the emulator reproduces thirteen maize growth outputs with high fidelity while returning calibrated confidence for every prediction, at a fraction of the original computational cost. Applying it across 100,000 trait configurations, six locations, and future climate projections extending to the year 2100, we identify 1,114 trait combinations that remain resilient under the 2100 future conditions and a stricter subset of 246 configurations that remain resilient across four future projection periods: 2040, 2060, 2080, and 2100. These resilient configurations occupy a distinct region of trait space, combining high radiation use efficiency with characteristic shifts in developmental and reproductive traits. Trait-importance analysis further identifies radiation use efficiency as the dominant driver of yield under future climate, while projected yields vary widely across locations and climate models, indicating strong spatial heterogeneity in the maize response to climate change.
Jingwen Wang, Jiao Li, Yuhang Qian, Yuyin Wang, Min Zhang
原文摘要
Multimodal emotion recognition in conversations (MERC) integrates textual, acoustic, and visual cues to identify utterance-level emotions. However, existing approaches generally do not explicitly adapt multimodal processing to the degree of inter-modal predictive disagreement. In real conversations, semantic incongruity, concealed emotion, noisy observations, and ambiguity can affect modality-specific emotional evidence differently, while the reliability of individual modalities varies across samples. Rather than identifying each source of inconsistency separately, we model their common consequence for emotion recognition as predictive disagreement among unimodal emotion distributions over the same emotion-label set. Such disagreement makes indiscriminate fusion prone to propagating unreliable or competing evidence. Inspired by the Brunswik Lens Model, we propose MiD-ERC, a divergence-centric framework that quantifies inter-modal predictive disagreement using a normalized divergence score and estimates sample-specific modality reliability. MiD-ERC uses these signals to coordinate divergence-guided cross-modal gating, divergence-modulated reliability weighting, and divergence-conditioned adaptive LLM prompting. The framework is trained end-to-end using only standard emotion annotations, without requiring additional conflict labels, auxiliary data construction, or external retrieval modules. Experiments on IEMOCAP and MELD demonstrate that MiD-ERC achieves state-of-the-art performance, with particularly pronounced improvements on high-divergence samples.
Quantitative strategy optimization remains largely manual, requiring domain experts to identify weak signals, tune risk-control rules, and repeatedly validate iterative revisions. Large language models can accelerate this process, but directly relying on them to rewrite trading strategies often introduces hallucinated edits, strategy drift, and backtest overfitting. We propose EVOQUANT, a self-Evolving Verifier-guided framework for strategy Optimization in Quantitative trading. Our method utilizes LLMs to deeply diagnose performance bottlenecks, generates semantically controlled candidate edits, selects the best strategy through a multi-stage verification pipeline, and distills optimization experience into reusable knowledge for continual self-improvement. We evaluate our method using seven representative strategies: four from the A-share market and three from the Crypto market. Experimental results show that our method significantly improves the Sharpe ratio across all tested strategies: the average test Sharpe increases from -0.298 to 0.538, and the best-performing strategy achieves a 199% relative improvement. Ablation studies and stress tests under stricter conditions further validate the effectiveness and robustness of the framework. Overall, this work transforms quantitative strategy optimization from costly manual trial and error into an automated and verifiable iterative paradigm, offering a new path for applying large language models to financial strategy research.
Clarissa Miranda-Pena, Andrew Reeson, Cécile Paris, Josiah Poon, Jonathan K. Kummerfeld
原文摘要
Despite extensive research, Large Language Models continue to hallucinate when generating code, particularly when using libraries. On NL-to-code benchmarks that require library use, we find that LLMs generate code that uses non-existent library features in 8.1-40% of responses. One intuitive approach for detection and mitigation of hallucinations is static analysis. In this paper, we analyse the potential of static analysis tools, both in terms of what they can solve and what they cannot. We find that static analysis tools can detect 16-70% of all errors, and 14-85% of library hallucinations, with performance varying by LLM and dataset. Through manual analysis, we identify cases a static method could not plausibly catch, which gives an upper bound on their potential from 48.5% to 77%. Overall, we show that static analysis methods are cheap method for addressing some forms of hallucination, and we quantify how far short of solving the problem they will always be.
We initiate the systematic study of quantum interactive proofs of proximity (QIPPs). These are quantum interactive proof systems with sublinear communication, where the verifier accepts inputs in a property $Π$ and rejects inputs $\varepsilon$-far from $Π$, while only querying the input a sublinear number of times. 1. We show that all logspace-uniform $\mathbf{NC}$ properties admit QIPPs with query, communication and time complexity all $\widetilde{O}(n^{2/5})$. This improves on the best known classical total complexity of $\widetilde{O}(\sqrt n)$, which is optimal in the worst case for polynomial-time verifiers under the PRG assumption of Kalai and Rothblum (CRYPTO, 2015). This resolves an open question of Dall'Agnol, Gur, Roy Moulik, and Thaler (Quantum, 2022) in the affirmative. 2. Assuming the existence of a post-quantum subexponentially secure PRG computable in logspace-uniform $\mathbf{NC}$, we prove that, for every nondecreasing, logspace-computable query bound $q(n)\le n^{1/2-Ω(1)}$, there exists a property $Π$ in logspace-uniform $\mathbf{NC}$ and a constant $\varepsilon>0$ such that no polynomial-time $\varepsilon$-QIPP for $Π$ can achieve query complexity $q(n)$ and communication complexity $c(n)$ satisfying $q(n)+c(n)=o(n^{1/3})$. 3. We establish a bi-directional connection between QIPPs for logspace-uniform $\mathbf{NC}$ and query-model QIPs for parity. A QIP for $k$-bit parity with total complexity $k^{2/3-Ω(1)}$ yields QIPPs for every logspace-uniform $\mathbf{NC}$ property with total complexity $n^{2/5-Ω(1)}$. Conversely, achieving total complexity $o(n^{2/5})$ within the standard $\mathsf{PVAL}$-based framework requires improving the known query-communication tradeoff for parity. Additionally, we show that almost every size-$2^{βn}$ property ($0<β<1$) requires linear total QIPP complexity at constant proximity.
Learning locomotion directly on a humanoid robot is a promising direction to avoid the sim-to-real gap. However, locomotion rewards in Reinforcement Learning (RL) typically track the base velocity, which cannot be measured onboard without an external motion capture system or well-calibrated state estimator. This is why we introduce \method{}, which learns humanoid locomotion from minutes of samples with a reward that only requires the onboard Inertial Measurement Unit (IMU). Instead of tracking the base velocity, STRIDE rewards the robot for following a commanded acceleration profile. To learn efficiently from this reward, we use the off-policy RL algorithm XQC and combine a high gain setting for the joint position controllers with left-right symmetry augmentation. In simulation, STRIDE learns to walk on the Unitree G1 and Booster T1 humanoids within 14 minutes of samples. Compared to learning without these two components, it reduces the falls during training by 81%, while without them the robot learns only in over an hour.
A source who has been right before deserves attention, but about what? We ask how far a demonstrated record of reliability travels in language models when the new claim cannot be checked, and whether the answer depends on the kind of claim. Three behavioral studies across six models approach the question from three angles. Study 1 adapts the Einstein-effect paradigm and finds that models, unlike the human sample, favor a spiritual leader over a scientist for the same obscure statement, and that the preference largely reverses when the statement is reworded in the vocabulary of physics. Study 2 adapts a paradigm in which a trusted expert calls a scientific or religious explanation unknown or a mystery; models reproduce the human domain asymmetry in exaggerated form. Study 3, the central test, gives two fictional informants different records on checkable river forecasts and then has them endorse matched claims about hidden natural and spiritual causes of the same village events. Most models carry the record to natural causes and barely at all to spiritual ones; one model transfers it weakly to either. Telling the model that the reliable informant forecasts by spiritual guidance rather than by measurement lowers that informant's influence on natural claims in every model and produces no comparable gain on spiritual claims in any. Demonstrated reliability thus has a scope in these models, and the account a source gives of how they came to be reliable can narrow it. Whether this reflects a judgment about relevance or the low prior credence of spiritual claims is the question the design leaves open, and we outline the experiment that would settle it.
Jacob Hobbs, Ashray Gupta, Dennis Ulmer, Geetanjali Aich, Nolan Golden, Yanan Long, Ian Rios-Sialer, Jennifer Mickel 等 13 位作者
原文摘要
While there is an abundance of research into large language model (LLM) biases against LGBTQ+ communities, far less research has examined how and why LGBTQ+ people choose to make use of AI chatbots. In this work, we examine how LGBTQ+ people choose (not) to use AI chatbots across three domains in which LGBTQ+ users have unique needs: relationship support, mental health support, and gender or sexual health support. We surveyed 73 people (60 LGBTQ+; 13 non-LGBTQ+) on how they interact with AI chatbots and find that participants' access to external resources, privacy concerns, and perceptions of chatbot capabilities shape whether and how they choose to make use of AI chatbots. For instance, some LGBTQ+ participants felt that chatbots might actually be less biased against LGBTQ+ people than the people they know offline. We also find that LGBTQ+ participants report using AI chatbots less frequently than non-LGBTQ+ participants across 15 specific personal health and relationship advice scenarios, citing concerns about AI sycophancy and adverse cognitive side-effects. Recognizing AI chatbots as vital (albeit imperfect) resources for LGBTQ+ people, we recommend researchers consider systemic issues that drive users to AI chatbots and design technologies (AI or otherwise) that prioritize privacy, anonymity, and real human connection.
Most FDR-controlled feature selection methods are designed for coordinate-wise hypotheses, where each feature has a single weight or importance score. This abstraction fails in sequential and grouped models, where one original feature is represented by a block of sub-features, such as lags, recurrent states, or attention-based interactions. We propose a grouped-feature FDR control framework for such settings. For grouped linear models, we construct null-symmetric block-level mirror statistics with matrix-valued perturbations. For neural sequential models, we combine Permutation SHAP derivatives as model-agnostic block-level importance scores with kernel-based dependence measure. The framework is model-agnostic across network architectures, does not require specifying the covariate distribution, and reduces to Gaussian Mirror or Neural Gaussian Mirror when the block size is one. We prove FDR control for low- and high-dimensional grouped linear models and asymptotic symmetry of smoothed Permutation SHAP derivatives under fixed fitted nonlinear models. Experiments on simulated and real-world datasets show reliable FDR control and improved power under correlated grouped-feature signals.
Shaswat Patel, Daniel Aarao Reis Arturi, Kevin Gomez, Jaume de Dios Pont, Jorge Poco, Eduardo V. Teixeira, Sam Westrick, Claudio Silva 等 9 位作者
原文摘要
Agentic systems now formalize research-level mathematics in Lean, but they are built for autonomy: the mathematician states a target and inspects a result. The alternative, a general-purpose coding harness, is domain-neutral and must be assembled into a formalization workspace before any mathematics begins. Mathematicians, however, report wanting to keep high-level control over how an argument is decomposed and established. We present $\textbf{Lea}$, a Lean-specific agent backbone for mathematician-facing software rather than a single interface. Lea exposes runs, sessions, and typed events through one application-neutral API, and ships with Lean tooling (blueprints, LSP, Loogle, SafeVerify) already wired. We report two applications built on it: $\textbf{LeaChat}$, a web client for interactive proof authoring, and $\textbf{LeaOverleaf}$, which formalizes theorems annotated directly in an Overleaf document; a session begun in either continues in the other. In a case study, a mathematician outlined a proof of the Durfee square generating function in a few natural-language messages, and Lea produced a complete Lean development from Mathlib alone, following the intended decomposition closely and requiring substantive correction at a single point. Lea and both applications are released as open source so that further mathematician-facing tools can be built on the same backbone.
Aoran Zhang, Tianyao Wei, Maria J. Guerrero, César A. Uribe
原文摘要
Recovering latent structure from count data has received considerable attention in network inference, particularly when one seeks both cross-group interactions and within-group similarity patterns in bipartite networks, which are widely used in ecology research. Such networks are often sparse and prone to imperfect detection. Existing models mainly focus on interaction recovery, while the induced similarity graphs are much less studied. Moreover, existing models for under-counted data place sparsity, if at all, on the detection model rather than on these graphs. To address these issues, we propose a nonnegative low-rank factorization framework that jointly estimates detection probabilities and promotes sparse within- and cross-group graphs through $\ell_{1/2}$ regularization. For the resulting nonconvex, nonsmooth problem, we develop an ADMM-based algorithm with adaptive penalization, and establish asymptotic feasibility and KKT stationarity of cluster points under stated assumptions. We also give sufficient conditions for exact recovery of within- and cross-group graphs. On synthetic data with up to 20% missing entries, the method recovers the graphs and detection probabilities with competitive accuracy. Across three ecological datasets, it matches the baseline held-out predictions while estimating detection probabilities and within- and cross-group graphs.
Large language models (LLMs) are increasingly used for text annotation in social science, but standard performance metrics do not explain why errors occur. Low F1 may reflect limitations of the annotation system, defects in the written codebook, or inconsistency in reference labels. We propose a pre-deployment codebook audit for LLM-assisted annotation. The audit uses class-specific metrics and confusion matrices to identify problematic classes and boundaries, structured document-level review to attribute error sources, and targeted codebook revision followed by held-out evaluation. We apply the audit to two published multi-class codebooks from distinct domains. In both applications, the audit identifies under-specified or conflicting coding rules, and targeted revisions reduce the diagnosed errors on held-out data. Analyses with two LLM annotators show that some problematic boundaries and codebook revisions transfer across models, whereas other revision effects are annotator-specific. The results show that LLM annotation errors can help identify weaknesses in the written measurement instrument, not only in the model. Codebook auditing complements prompting, retrieval, and fine-tuning by clarifying whether persistent errors call for model improvement, codebook revision, or renewed scrutiny of reference labels.
External safety filters can prevent unsafe requests from reaching a large language model (LLM), but may also reject legitimate queries. We present an end-to-end empirical evaluation of Jev, a probabilistic decision model, as a safety guardrail for a Qwen2.5-1.5B-based pipeline. Three configurations are compared on 970 benchmark prompts: an unfiltered LLM, Jev input filtering, and tiered Jev input filtering with selective output review. Across 720 prompts labeled harmful (520 AdvBench and 200 XSTest-unsafe), the fraction recorded as answered decreases from 527/720 (73.2%) without filtering to 17/720 (2.4%) with input filtering. On 250 XSTest-safe prompts, however, non-answer/refusal increases from 3/250 (1.2%) to 27/250 (10.8%). Tiered filtering yields 40/720 (5.6%) harmful-prompt answers and 10/250 (4.0%) benign-prompt refusals, with 79/970 (8.1%) of prompts requiring output review. Joint optimization of the tiered thresholds improves macro-F1 from 0.9351 to 0.9512, while reducing harmful-prompt answers and output-review frequency at the same benign-refusal rate. Matched-budget comparisons reveal modest additional gains over a retuned single gate on the evaluation data, although these gains are not consistently observed on a held-out split.
Many engineering failures in orientation-dependent systems are geometric failure modes: changing the material delays the failure; changing the geometry can eliminate it. The mono-monostatic property (exactly one stable equilibrium under gravity) is mathematically proven to exist in convex homogeneous bodies, but, to our knowledge, no verified smooth geometry had been openly published when v1 appeared (April 2026). We introduce an Equilibrium Count Score (ECS) oracle that measures stable equilibria via drainage basin analysis on the center-of-mass height landscape. v2 correction: v1 reported that Sloan's (2023) analytical Gomboc parameterization shows 4-11 local minima "at all published parameter values." That claim is withdrawn. The beta values swept were this work's own grid points, not Sloan's (he states convexity bounds beta <= 0.15 and beta <= 0.17); they lie inside the strictly convex sub-regime beta < ~0.036, where a body at beta = 0.023 was later certified mono-monostatic by interval arithmetic; and the oracle's two stages are biased in opposite directions, so its count can be wrong either way. We construct one strictly convex mono-monostatic body by extending the Sloan phase function with a Fourier term and optimizing via differential evolution, with ECS=1 confirmed across merge thresholds from 0.5% to 10%; two more reach ECS=1 under the oracle but are not strictly convex. Conventional geometries cannot achieve ECS=1 through ballast alone: cylinders and cubes retain multiple stable equilibria even at 30% bottom-weighted mass. Applications: an IMU calibration housing (349x orientation precision improvement), aerial reforestation seed pods (20-67% germination loss from incorrect landing orientation), and marine buoy self-righting. Cross-layer scoring shows the Gomboc is 11.8x worse than the cylinder on contact distribution while optimal on equilibrium stability.
Chao Xia, Xi Zeng, Alexey Vdovin, Huanxia Wei, Simone Sebben, Carl Lewerth, Qing Jia, Zhigang Yang
原文摘要
The wheel region contributes substantially to vehicle aerodynamic drag, and rim geometry strongly influences the flow in this region. However, publicly available automotive aerodynamics datasets offer limited variation in rim design. We present DrivAerRim, a large-scale dataset of full-vehicle computational fluid dynamics simulations covering 904 DeepWheel rim designs with diverse spoke layouts, opening areas and topologies. All cases use the same DrivAer Estateback, operating conditions and numerical setup, with rim geometry as the sole design variable. Wheel rotation is modelled using moving reference frame regions around the rims and rotating wall conditions on the deformed tyres. The numerical setup is assessed through mesh sensitivity analysis and comparisons with wind tunnel measurements. The resulting 9.32 TB release includes wheel geometries, vehicle and component force coefficients, accumulated drag, surface pressure, wall shear stress, and flow fields around the vehicle and wheels. Together, these data support surrogate model training and evaluation across diverse rim geometries, as well as aerodynamic optimisation and generative design to accelerate wheel rim development for vehicle drag reduction.
Jiwoo Choi, In Ik Lee, Woo Sung Seo, Minho Lee, Yun Young Choi
原文摘要
Estimating advertising performance in cold-start settings is challenging because historical logs are unavailable before launch. Large language models (LLMs) offer a potential alternative by simulating persona-conditioned user responses, but it remains unclear which design choices actually drive LLM-based persona simulation. Using real Google Ads e-commerce campaign data across five demographic segments, we identify three signal-form dimensions that affect simulation quality: (i) a demographic anchor without noisy rich attributes, (ii) a relative qualitative banner-blindness signal rather than quantitative anchors, and (iii) prompt phrasing. On this campaign, our method attains an MAE of 24.9 and a MAPE of 26.1%, a 51.3% lower MAE than the best tested non-LLM baseline and 77.3% lower than an LLM configuration using explicit CTR anchors. Systematic ablations show that minimal demographic descriptions achieve lower mean error than richer persona configurations, while relative qualitative cues substantially outperform the tested quantitative anchors. We additionally compare these configurations with direct-prior and rule-based baselines and with alternative LLM configurations.
Emilia Agasi, Charlie Gourley, C. Simon Herrington, Robert L. Hollis, T. Ian Simpson
原文摘要
Histopathology foundation models provide rich representations of tissue morphology, but technical artefacts in whole slide images introduce noise that may impact diagnostic accuracy. We introduce Artefact Prototype Learning (APL), a novel approach to detecting these artefacts that exploits the geometry of pre-trained embedding spaces without manual labels or fine-tuning. Rather than treating artefacts as isolated outliers, APL computes prototypes of out-of-distribution patches and uses them to distinguish artefact-associated patterns from tissue. Evaluation on a single-centre dataset shows that APL outperforms prior methods GrandQC and HistoART in detecting artefacts such as blur, dark spots, slide edges and pen marks, but struggles with tissue folds and blood. These results show that frozen pre-trained embeddings already carry information that allows artefact detection without expensive labels or vast computational resources.
Wenbo Fu, Chenghui Mao, Haifeng Lv, Chengwen Hu, Zhou Lu
原文摘要
We introduce Radial Kernel Attention (RKA) for long-range modeling in polymer machine learning force fields. Existing approaches often rely on local neighbor graphs, which limit communication across polymer chains and their surrounding environments, whereas explicit all-pair attention incurs quadratic computational cost. RKA maps interatomic distances from the cluster centroid to learned radial features and combines their kernelized overlap with content-dependent queries and keys. Factorized aggregation and readout enable global feature exchange at linear complexity in the number of atoms for fixed feature dimensions. To ensure physically consistent predictions, we insert the invariant scalar update between EquiformerV3 local blocks and obtain conservative forces by differentiating the total energy, including its dependence on radial coordinate . Our models are trained on official OPoly26 training set for 12 epochs and evaluated directly after training on the complete official validation set. A matched 5M comparison shows that RKA reduces per-atom energy and force MAEs by 67.4% and 36.3%, respectively. The 20M RKA model attains a total-energy MAE of 15.81 meV (per-atom MAE of 0.1182 meV/atom) and a force MAE of 2.619 meV/Angstrom,improving all three metrics over UMA-S-1.2 and UMA-S-1.2.1 on the same validation structures. RKA provides a compact pathway for information exchange beyond local graph connectivity in polymer machine learning force fields.
Vineet Jagadeesan Nair, Stephen Peng, Arun Ravindran
原文摘要
Equivalent encodings of an optimization problem can produce different learned predictions. Interpreting these differences requires accounting for numerical error and multiple optima. We construct metamorphic tests for optimization solver surrogates with explicit eligibility conditions, output alignment, and numerical thresholds. Reference-solver agreement and independent constraint checks determine which sampled pairs enter the audit. We study alternating-current optimal power flow (AC-OPF), the traveling salesman problem (TSP), the capacitated vehicle routing problem (CVRP), and maximum independent set (MIS). Across five grids, GridSFM-Open violates each of 13 relations on at least one grid, and 12 by more than twice the threshold. Randomly initialized copies of the grid models fail nearly the same relations, so the failures do not require training. Source patches and input canonicalization remove selected discrepancies, but no intervention establishes preserved accuracy and feasibility within the declared margin on every grid. In routing, Policy Optimization with Multiple Optima (POMO) has a single-pass rotation/reflection violation rate that rises from 12.7% to 89.5% between 20 and 100 cities. Graph-Based Diffusion Solvers for Combinatorial Optimization (DIFUSCO) violates the MIS disjoint-union relation on 35.6% of draws. Post hoc normalization within connected components removes these violations on the tested graphs. These findings support reporting metamorphic consistency alongside optimality, feasibility, and accuracy, and validating repair quality separately from consistency.
Semidefinite programming (SDP), particularly the sum-of-squares (SoS) hierarchy, provides $n^{O(q)}$-time approximation and certification algorithms for optimizing degree-$d$ homogeneous polynomials on the unit sphere. For arbitrary inputs, degree-$q$ SoS gives a powered-objective approximation factor $O_d((n/q)^{d/2-1})$ [BGGLT17]. For random order-$d$ tensors, degree-$d$ SoS has an integrality-gap lower bound of $2^{-O(d)}(n/d)^{d/4-1/2}$ relative to the true spherical maximum; lower bounds also extend to growing SoS degree in tensor PCA [BGL17,HKP17]. We establish lower bounds for weaker linear-programming relaxation hierarchies. Our degree-$q$ moment LPs impose the sphere identities and any prescribed, instance-independent family $\mathcal{G}$ of valid polynomial inequalities of degree at most $q$. They include Sherali--Adams-type product hierarchies. For every even $d\ge2$, constant $C>0$, and $q$ satisfying $2d | q$ and $2d\le q\le n$, every such relaxation with at most $n^{Cq}$ tests has instances $P$ satisfying \[ \frac{\mathrm{PowLP}_{q,\mathcal{G}}(P)}{\|P\|_\infty}\ge \frac{c_{d,C}}{\sqrt{q\log n}}\left(\frac nq\right)^{d/2-1/2}.\] Here $\|P\|_\infty$ is the maximum absolute value of the polynomial over the sphere and $\mathrm{PowLP}_{q,\mathcal{G}}(P)$ is the relaxed objective. The exponent $d/2-1/2$ improves the leading dependence on $d$ relative to the $d/4$-type exponents in prior random-tensor SoS lower bounds. Our LP hierarchies impose weaker positivity conditions than semidefinite moment relaxations. It leaves open whether analogous degree dependence can be achieved in lower bounds for SDP hierarchies.
Let $P$ be a set of $n$ points in the plane, and let $C$ be a compact, convex subset of the plane that contains the origin in its interior. We study the problem of finding two points $p$ and $q$ of $P$ with the property that any homothet of $C$ that contains $p$ and $q$ must contain at least a constant fraction of the points of $P$. We give proofs for several cases. When $C$ is strictly convex we achieve a fraction of $1/5$, and when $C$ is a general convex region we achieve a fraction of $1/6e$. We show how the ideas of this second bound can be used to prove a lower bound for a version of the problem in $\mathbb{R}^d$ when $C$ is a $d$-ball. We also consider a few special cases. When $C$ is a triangle we get a lower bound of $1/12$, and when we allow that triangle to be rotated by $π$ radians in addition to being scaled and translated we get a lower bound of $1/768$. In proving this last result, we improve the state of the art on the version of this problem in three dimensions, where $C$ is a parallelepiped. We also consider the algorithmic aspects of finding such a pair of points.
Specialized prompt injection guards are deployed in front of document analysis systems, but no false positive rate has been published for any of them on legal contracts. We measure eleven open-weights guards on two corpora of legal contracts and find that existing guards struggle: document-level false positive rates are as high as 92.39%; two guards fall below chance on paired AUC; and a five-regex baseline is not Pareto-dominated by any guard in the panel. Controlled probes locate model-specific as well as general failure modes. Prompted LLMs resolve the tested failures, including a second threat category (unauthorized scope claims) on which every guard scores zero. A cascaded LLM solution, where the positives of a mid-tier model are verified by a frontier model, is 6.5x cheaper than using exclusively a frontier model and approaches its performance.
In the lower-star filtration (LSF), simplices are assigned filtration values according to the maximum-vertex rule, with each simplex entering the filtration when its highest-valued vertex appears. A limitation of this construction is that some higher-dimensional simplices may enter simultaneously with their boundary faces, causing certain topological features to be filled immediately upon formation and thus not appear in the resulting persistence diagram. To address this limitation, we propose three alternative filtration methods: the \(\varepsilon\)-lower-star filtration (ELSF), the \(L_p\) filtration (LpF), and the softmax filtration (SMF). {ELSF introduces a dimension-dependent delay to the maximum-vertex rule,} allowing higher-dimensional simplices to enter the filtration after their boundary faces, whereas {LpF and SMF use aggregation functions that depend on all vertex values rather} than solely on their maximum. By modifying how filtration values are assigned, these methods allow simplices to enter the filtration more progressively, so that topological features such as loops and voids may persist over longer filtration intervals rather than being filled immediately. Consequently, the resulting persistence diagrams may include persistence features that are not observed under LSF, potentially providing useful information for distinguishing data populations. Experimental results on simulated and real-world datasets demonstrate that the proposed filtration methods are computationally efficient and can provide greater sensitivity to structural differences in the data than LSF.