Machine Learning
Genesis Steering: Better Activation Steering Through Training Data
Tora Bodin, Eric Aubinais, Ilaria Torre et al. (+1 more)
Activation steering controls large
language models at inference time, but its effectiveness depends on the choice of reference data. We introduce genesis steering, which derives a genesis vector from contrastive examples in the model's own fine-tuning data. When these data are unavailable, we draw on the membership inference literature to identify likely training examples and construct a proxy
dataset. We establish two theoretical results on optimal displacement to the decision boundary and reduced proxy-estimation bias. We evaluate genesis steering and membership-based proxy construction on Tülu-3-8B, whose fine-tuning data are publicly available, and Llama-3.1-8B-Instruct as a control. We find that safety data from Tülu-3-8B's training mixture yield stronger steering than external
datasets on both models. On Tülu-3-8B, membership scores largely agree in identifying training examples, and vectors built from top-ranked examples achieve competitive performance against random subsets of the same size. In contrast, the results on Llama-3.1-8B-Instruct hint that a MIA method will only have a positive effect if the mean direction shift with MIA-filtered in-distribution data is much less severe than the direction shift with OoD data.
Read the full abstract ▾
All authors Tora Bodin, Eric Aubinais, Ilaria Torre, Chih-Hong Cheng
Computation and Language
From Copying to Composition: How Language Models Develop In-Context Learning
Minglai Yang, Xinyu Guo, Renxiong Wang et al. (+3 more)
In-context learning (ICL) allows
language models to adapt their behavior from instructions and demonstrations in a
prompt, but accuracy alone does not reveal how contextual information controls their predictions. We study the acquisition and trainability of ICL with a controlled diagnostic that separates copying a nearby demonstration from integrating information across the
prompt. Across 17 models, about a third prefer the copied answer, a few are undecided, and the rest prefer the rule-consistent answer, which goes with higher accuracy and with control by the requested value rather than the closest conflicting example. Checkpoint analyses show that models begin prioritizing task-relevant
prompt tokens over conflicting examples long before their accuracy improves beyond chance level. In matched training experiments, synthetic documents that define endings and then query them shift models toward compositional responses on the controlled diagnostic, whereas fine-tuning on definition exercises built from rewritten real documents improves accuracy when
prompts reassign label meanings. The results suggest that ICL is not acquired as a single, uniform capability. Instead,
prompt control, answer preference, construction without candidates, and transfer can develop along distinct trajectories.
Read the full abstract ▾
All authors Minglai Yang, Xinyu Guo, Renxiong Wang, Sunjie Hou, Daniel Yue Zhang, Yunzhong He
Materials Science
Engineering Correlated Transparent Conductors via Metal-Insulator Transition Proximity
Amit Cohen, Nofar Tuval, Maria Baskin et al. (+5 more)
Transparent conductors are key functional materials in modern optoelectronic devices. Their performance is governed by an inherent trade-off between optical transparency and electrical conductivity. Correlated metallic oxides offer an attractive workaround to this compromise, where narrow partially-filled bands provide high conductivity, while the absence of visible-range transitions enables optical transparency. Despite this promise, transparency and conductivity remain coupled through carrier density and electronic coherence. Here, we use proximity to a metal-
insulator transition (MIT) as a design principle for engineering this balance in correlated-oxide transparent conductors. Epitaxial Ti-substituted $SrVO_3$ films serve as a model platform where the optical window and
metallicity are tuned simultaneously. We demonstrate and explain a systematic broadening of the optical window to regimes relevant for silicon photovoltaics. This gain is counterbalanced by carrier depletion and disorder, which drive the loss of coherent transport near the composition-driven MIT. A Haacke figure-of-merit analysis identifies the optimal composition range near the MIT, where the loss of metallic transport is accompanied by persistent near-Fermi-level spectral weight, indicating disorder-assisted localization together with correlation-driven spectral-weight redistribution. These findings establish proximity to an MIT as a general materials-design strategy, using electronic filling as a tuning knob for the transparency-conductivity balance in correlated-oxide transparent conductors.
Read the full abstract ▾
All authors Amit Cohen, Nofar Tuval, Maria Baskin, Tien-Lin Lee, Tom Tiwald, Scott A. Chambers, Sophie D. Beck, Lior Kornblum
High Energy Physics (Phenomenology)
Pythia-Contrib: a modular contribution framework for the Pythia 8 event generator
Benoit Assi, Christian Bierlich, Rikab Gambhir et al. (+6 more)
Pythia is a general-purpose Monte Carlo event generator used in particle, astroparticle, and nuclear physics, by theorists and experimentalists, to simulate particle collisions. Because of the size and diversity of the user community, it is challenging to address all of the needs for new and varied physics models. This paper describes the Pythia-Contrib system, which allows for third-party developers to contribute code to Pythia without introducing external software dependencies to the Pythia code base.
Read the full abstract ▾
All authors Benoit Assi, Christian Bierlich, Rikab Gambhir, Philip Ilten, Tony Menzo, Stephen Mrenna, Manuel Szewc, Michael K. Wilkinson, Jure Zupan
Number Theory
A new way integral for the $n$-fold cover of $SL_2$
Tianxiang Liu
Let $\pi$ be a genuine irreducible cuspidal automorphic representation of the $n$-fold cover $SL_2^{(n)}(\mathbb{A})$ of $SL_2(\mathbb{A})$, $n \geq 3$ odd, and let $\tau$ be the genuine representation of $GL_1^{(n)}(\mathbb{A})$ obtained from a
unitary Hecke character of $\mathbb{A}^\times$. We prove that, for a certain choice of data, a new way integral for $SL_2^{(n)}(\mathbb{A})$ represents the partial $L$-function $L^S(ns+\tfrac12,\pi\times\tau)$. The integral is derived from the generalized doubling integral of Cai, Friedberg, Ginzburg and Kaplan, following the idea of Ginzburg and Soudry.
Read the full abstract ▾
Quantum Physics
Post-Optimization Subspace Corrections to Quantum Ansätze
Prince Frederick Kwao, Bharath Sambasivam, Jason Saroni et al. (+3 more)
Circuit depth and measurement count are the main limitations for chemical applications of the leading variational
quantum algorithms on
quantum computers. We show that both resource requirements can be reduced by augmenting adaptive variational
quantum algorithms with post-optimization corrective techniques. We focus on two such techniques—
Hamiltonian-moment Krylov subspace diagonalization and the identity-augmented
quantum self-consistent equation of motion (q-sc-EOM)—applied to optimized ADAPT-VQE states without further ansatz growth or nonlinear optimization. Across molecular systems studied with up to 20
qubits, first-order Krylov correction reduces the optimized iterations (and the corresponding number of adaptive layers) needed to reach the $1.6~\mathrm{m}E_h$ threshold by up to approximately sixfold, with a general trend toward larger reductions at higher determinant Rényi
entropy. We also find measurement-cost reductions of up to approximately fortyfold using a suitable resource-cost model. Identity-augmented q-sc-EOM improves molecular potential-energy curves obtained from truncated ADAPT-VQE ansätze with limited expressivity and recovers accuracy from states produced by sampled operator ranking without additional gradient-based operator selection, providing an advantage when small pool gradients make selection measurement intensive. Together, these post-optimization corrections offer routes to improve energy accuracy and reduce adaptive
circuit growth in
quantum algorithms as
quantum hardware advances.
Read the full abstract ▾
All authors Prince Frederick Kwao, Bharath Sambasivam, Jason Saroni, Srivathsan Poyyapakkam Sundar, Edwin Barnes, Ayush Asthana
Computer Vision
Towards a Data-Centric Understanding of Spatial Reasoning in Multimodal Large Language Models
Mahdieh Mallahnezhad, Masoud Jafaripour, Mohamadali Tofigh et al. (+2 more)
Multimodal Large
Language Models (MLLMs) have been widely applied in embodied
agents and robotics for
vision tasks. However, MLLMs suffer from limitations in spatial
reasoning and perception. Post-training techniques have shown promise for improving these capabilities; yet, there are limited data-driven surveys of such approaches. In this survey, we consider approaches to fine-tuning pretrained MLLMs for spatial
reasoning tasks, ranging from supervised to reinforcement learning, with an emphasis on the
datasets used. We investigate themes of
dataset construction, data modality and annotation, and evaluation of the resulting models. Based on these themes, we categorize research into three groups:
dataset design,
dataset construction and evaluation. We find four main methods for constructing fixed spatial annotations: rules and templates, LLM prompting, procedural generation, and human-in-the-loop. We also observe substantial variation in evaluation approaches, including different scoring methods, tolerance metrics, and
benchmark types. Finally, we identify opportunities for future research in unified dynamic annotation, efficient spatial
reasoning, standardization of evaluation, expanded spatial range, and human-centered spatial interactions.
Read the full abstract ▾
All authors Mahdieh Mallahnezhad, Masoud Jafaripour, Mohamadali Tofigh, Sheila Schoepp, Osmar R. Zaiane
Medical Physics
Can $^{225}$Ac be imaged on a long axial field-of-view PET/CT with standard clinical settings?
Lorenzo Mercolli
Purpose: The $^{225}$Ac decay chain includes $^{209}$Tl, whose prominent 1567 keV gamma line can create electron-positron pairs. We investigated whether a long axial field-of-view (LAFOV) PET/CT can image $^{225}$Ac using standard clinical acquisition and reconstruction methods. Methods: A syringe placed next to a vial (5.89 and 10.29 MBq of $^{225}$Ac activity) and a NEMA IEC body phantom (sphere activity concentration 4.56 kBq mL$^{-1}$) were scanned for 45 and 60 min, respectively, on a Biograph
Vision Quadra (Siemens Healthineers). To support our findings, FLUKA simulations scored the annihilation locations around a $^{225}$Ac point source. Results: The contours of the syringe and vial were visible despite high noise, but the NEMA spheres could not be distinguished from the background. As illustrated by the FLUKA simulations, the number of annihilation events decreases exponentially according to the mean free path of the 1567 keV
photon in the surrounding medium. Conclusion: A LAFOV PET/CT can detect positrons created by pair production of the main $^{209}$Tl gamma line. However, these positrons are created where the
photons interact with the surrounding medium rather than at the location of the $^{209}$Tl decays, which rules out standard PET imaging of $^{225}$Ac.
Read the full abstract ▾
General Relativity and Quantum Cosmology
Hussain Gohar
Gravity from
entropy (GfE) identifies its Lagrangian with the
geometric quantum relative
entropy (GQRE) between the spacetime metric and a metric induced by matter and
curvature, treating the metric itself as a
quantum operator. We extend this construction to the Tsallis, Rényi, and Sharma-Mittal families, which we call
gravity from generalized
entropy (GfGE), and determine their common Einstein limit, weak-field spectra, auxiliary-field formulation, and horizon
entropy. The
entropy choice controls the leading higher-
curvature terms and additional modes. We confirm that the linearized GfE spectrum contains a healthy scalar and a Planck-scale tachyonic
spin-2 ghost. We identify the conditions that remove this ghost:
curvature must enter through its scalar trace, or the
entropy generator must be flat to quadratic order at the identity. The latter gives exact general relativity for Tsallis at $q=0$ and ghost-free $f(R,L_m)$ theories for Rényi and Sharma-Mittal. On shell, the auxiliary G-field is proportional to the tensor in Wald's
entropy formula. For analytic
entropy functions, horizons with a single
curvature scale have an area term followed by constant and inverse-area corrections. We also extend the volume-constraint cosmological constant to the full family; its leading
entropy dependence is set by the scalar mass. A finite set of
entropy-function derivatives controls the leading weak-
curvature departures from general relativity.
Read the full abstract ▾
Solar and Stellar Astrophysics
A Distinct Type of Standstill in Ultra-Compact White Dwarf Binaries
Jan Kára, Liliana Rivera Sandoval, Iwona Kotko et al. (+3 more)
AM CVn
stars are ultra-compact, semi-detached binaries consisting of a white dwarf primary and a hydrogen-poor secondary. In this work, we analyse the AM CVn-type systems MGAB-V270, ZTF18aaxuusk and ASASSN-19ct, which have similar properties and show unusual post-superoutburst behaviour. We focus on the occurrence of standstill-like phases. In MGAB-V270, superoutbursts are followed by a series of echo outbursts that evolves into a standstill-like phase before the system returns to quiescence. ZTF18aaxuus shows similar post-superoutburst activity, with one echo-outburst series also developing into a standstill-like phase. ASASSN-19ct exhibits comparable echo outbursts but no standstill-like phases. Periodicity analysis using TESS data shows that superhumps can persist during both echo outbursts and standstill-like phases. Although standstills have previously been observed in AM CVn
stars, the standstill-like phases we observe appear to represent a distinct phenomenon associated with post-superoutburst activity. We compare our results to other AM CVns that may exhibit similar behaviour.
Read the full abstract ▾
All authors Jan Kára, Liliana Rivera Sandoval, Iwona Kotko, Jean-Marie Hameury, Luis E. Salazar Manzano, Ryan J. Oelkers
Machine Learning (Statistics)
Neural Supervised Clustering for Learning Individualized Treatment Rules with Many Treatments
Daixi Huang, Yixin Yuan, Yifan Dai et al. (+1 more)
Learning an individualized treatment rule (ITR) with many treatments requires estimating treatment-specific effects from limited observations per arm. Treatments with similar conditional responses can share information, but linear contrast models may miss nonlinear heterogeneity. We propose nonlinear supervised clustering fusion (NSCF), a neural Q-learning framework with a shared nonlinear
covariate representation and low-dimensional treatment
embeddings. Pairwise fusion groups treatments with similar learned contrasts, yielding an ITR and an interpretable partition. An alternating
algorithm combines neural training,
embedding fusion, validation-based partition selection, and unpenalized grouped refitting. NSCF lowers regret relative to a strong neural reference in nonlinear exact-sharing simulations; perturbations characterize the limits of hard sharing. In a melanoma xenograft application, NSCF improves average policy value over the neural reference. These results support nonlinear representation learning with supervised treatment fusion when treatment effects exhibit shared structure.
Read the full abstract ▾
All authors Daixi Huang, Yixin Yuan, Yifan Dai, Yufeng Liu
Machine Learning
Contextual Credit Policy Optimization
Huiwen Han, Ziyao Zeng, Zihao Zhu et al. (+6 more)
Group-based reinforcement learning enables large
language model
agents to learn from sparse outcome rewards in long-horizon tasks. Recent methods derive step-level credit by comparing observation-matched experiences across trajectories. However, identical observations can occur in different contexts, and matched peers are not equally informative for a given decision. Moreover, favorable terminal outcomes can obscure redundant actions or detours within successful trajectories. To address these two challenges, we propose Contextual Credit Policy Optimization (CCPO). For the first, CCPO weights observation-matched peers from other trajectories by contextual similarity, computed from frozen LLM representations and accumulated trajectory statistics, allowing peers from similar contexts to contribute more to a step's baseline. For the second, CCPO pairs historical credit, which compares the realized return against the weighted peer baseline, with future credit, which measures the change in contextual outcome value over a short realized continuation. A step is thus credited for the progress it makes, not only for its trajectory's eventual outcome. Both credits are computed in hindsight from rollouts already collected for training, enabling CCPO to address these challenges without an auxiliary value critic. Across two Qwen2.5-Instruct backbones (1.5B and 7B), CCPO achieves up to 97.4% success on ALFWorld and 79.4% on WebShop and outperforms GRPO by up to 21.6 and 18.5 points. Ablations show that contextual weighting outperforms uniform weighting, and that historical and future credit serve as complementary learning signals.
Read the full abstract ▾
All authors Huiwen Han, Ziyao Zeng, Zihao Zhu, Sisong Bei, Hao Wang, Hongyu Xu, Yan Han, Degui Zhi, Zhiwen Fan
Computer Vision
The Gap Is in the Geometry: A Convergent Meta-Analysis of Multi-Facet Aerial and Remote Sensing Video Understanding
Georgios Angelidis, Aggelos Psiris, Maria Vakalopoulou et al. (+3 more)
Aerial and remote sensing
video is increasingly evaluated like web
video, by posing multiple-choice questions to
multimodal large
language models (MLLMs) over sampled frames and reporting accuracy. This paper argues that this recipe measures the wrong aspect, and supports the claim with a convergent meta-analysis of a broad set of published results, organized in four evidence tiers, in which a claim becomes a position only when collected evidence agrees. Parameter scale exhibits a median of +3.4 accuracy points per tenfold increase on aerial
benchmarks, far below a human gap of at least 19 points, whereas acquisition conditions control perception, as 315 of 318 evaluations in 19 within-paper contrasts lose accuracy when conditions such as altitude, view, object size or time horizon become harder. No aerial
benchmark in the corpus combines
video input with per-task human and random baselines, competence collapses from answering to acting in closed-loop navigation, and the same MLLMs score higher on the aerial
benchmark than on ground spatial ones, but lower on studies that test cross-view geometry. From these findings the current study derives five falsifiable principles and proposes the AVU-Card, a reporting card for aerial
video benchmarks
Read the full abstract ▾
All authors Georgios Angelidis, Aggelos Psiris, Maria Vakalopoulou, Georgios Th. Papadopoulos, Giorgos Sfikas, Iraklis Varlamis
High Energy Physics (Phenomenology)
Quasiparticle Modeling of Strongly Magnetized Strange Quark Systems
Yogesh Kumar, Manas Lohani, Poonam Jain et al. (+5 more)
We examine
spin-polarized Strange
Quark Matter and the complex structure of Strange
Quark Stars in the existence of a strong
magnetic field. Our framework uses a quasiparticle model, en hanced by a QCD-informed effective
quark mass dependent on both chemical potential value and
magnetic field strength. The resulting Equation of State yields an energy per baryon below the iron limit, supporting the hypothesis that the actual ground state of matter is strange matter. The Tolman-Oppenheimer-Volkoff expressions are solved and find that the maximum stellar mass decreases with the degree of
spin polarization, reaching as high as 2.35 Mo with a related radius 12.33 km. Furthermore, the model predictions for mass, radius, and surface
redshift are in excel lent conformance to observational limitations for the massive
pulsar PSR J0952-0607. These results highlight the crucial role of quasiparticle medium effects and strong
magnetic fields in describing dense matter.
Read the full abstract ▾
All authors Yogesh Kumar, Manas Lohani, Poonam Jain, Pargin Bangotra, Vinod Kumar, T. A. Nahool, Om Prakash, Sanjay Tyagi
Artificial Intelligence
GroundGUI: GUI Grounding by High-Quality Data and Distillation-to-Reinforcement Learning
Qingsong Xie, Jiahao Tian, Jiacheng Tang et al. (+5 more)
Constructing robust GUI
agents relies on
multimodal visual grounding: the ability to precisely map natural
language intents to actionable on-screen components. Despite abundant desktop, web, and mobile
benchmarks, high-fidelity resources tailored for diverse environments are surprisingly scarce. To bridge this gap, we introduce the Screen Referring Grounding
Dataset (ScreenRef), a comprehensive GUI grounding
dataset consisting of GUI screenshots paired with natural
language referring expressions and box or point annotations. Spanning desktop, web, and mobile domains, ScreenRef comprises 586K screenshots and over 633K referring expression instances, covering tasks such as text recognition, visual matching, spatial understanding, and refusal scenarios. To boost the model's grounding performance, we propose the
Distillation-to-Reinforcement Learning (D2RL) method to address the limited generalization of supervised fine-tuning and the sparse reward problem in reinforcement learning. Utilizing ScreenRef, we train Qwen3-VL-8B-Instruct via sequential SFT and D2RL training, yielding GroundGUI-8B. GroundGUI-8B achieves 95.2% on ScreenSpot-V2, 63.9% on ScreenSpot-Pro, 65.4% on OSWorld-G, 86.8% on MMBench-GUI L2, and 38.3% on UI-
Vision. Among 8B-scale models, it produces substantial gains over models trained on open-source
datasets, with performance comparable or even superior to models trained on closed-source data. Code, models, and
datasets are available at https://github.com/PrentisAI/GroundGUI.
Read the full abstract ▾
All authors Qingsong Xie, Jiahao Tian, Jiacheng Tang, Zhifeng Li, Madalina Ciobanu, Navan Preet Singh, Qingqing Mao, Ritankar Das
Game Theory
Maximin share allocations for two types of items
Yu Chen, Kai Jie Wong and Xiaorui Zhang
We study the fair allocation of indivisible items among $n$
agents with additive valuations (or costs) using the
fairness notion of maximin share (MMS). Since MMS allocations do not always exist when $n > 2$, one direction in the literature has been to identify special classes of instances in which exact MMS allocations can be guaranteed. We show that, when items can be divided into two types of chores, MMS allocations always exist and can be computed in
polynomial time using a novel
geometric approach inspired by prior work in job scheduling. We further discuss how our technique extends to instances with two types of goods, as well as mixed instances consisting of one type of objective good and one type of objective chore.
Read the full abstract ▾
Systems and Control
Global Parameter Identification and Uncertainty Quantification of reduced-order ADM1 using Maximum Likelihood Estimation and Bayesian Inference
Simon Hellmann, Leander Lerch, Terrance Wilms et al. (+6 more)
Anaerobic digestion (AD) models can be challenging to calibrate against real data due to limited practical identifiability, data scarcity, and model complexity. Weighted least squares objective functions and gradient-based solvers using finite differences are often used in literature, but result in suboptimal parameter estimates and large uncertainty bands. This paper identifies the parameters of a reduced-order ADM1 by comparing the frameworks of maximum likelihood estimation (MLE) and
Bayesian inference (BI), and quantifies uncertainties of parameters and outputs. Multirate experimental data are taken from a lab-scale AD reactor fed with agricultural substrates and operated at organic loading rates of 3$\,$g VS $\mathrm{m^{-3}d^{-1}}$. Global optima are found via global and gradient-based local search
algorithms using forward sensitivities. Enhanced scatter search provides the best global search results. These are refined by the Gauss-Newton
algorithm as the best local search, which delivers $R^2$ values of 0.95 for acetic acid and 0.77 for biogas production. BI yields substantially narrower uncertainty bands than MLE and avoids the need for a global search by incorporating prior knowledge about parameter distributions. The search space is captured efficiently by log10 scaling of parameters. Algorithmic details on its implications are provided. Despite a high-resolution lab-scale
dataset, the ADM1-specific inhibition parameter $K_\mathrm{I,nh3}$ is practically not identifiable with MLE, and shows a low output sensitivity, whereas a parameter for the effective ion concentration shows superb practical identifiabilty at high confidence with both MLE and BI. The present work lays a scalable base for calibration of nonlinear AD models by providing reliable simulations. It thus advances model-based monitoring and control applications of multirate processes such as agricultural AD plants.
Read the full abstract ▾
All authors Simon Hellmann, Leander Lerch, Terrance Wilms, Moritz Reumschüssel, Alberto Meola, Félix Delory, Steffi Knorn, Stefan Streif, Sören Weinrich
Robotics
Benchmarking 3D Reconstruction for Under-Ice Robotic Perception in Arctic Environments
Klára Orbán, Ashwin Nedungadi, Ana-Ştefania Kogălniceanu et al. (+5 more)
As the Arctic transitions toward a seasonally ice-free state, capturing the complex morphology of submerged sea ice is crucial for advancing models of ocean-ice-atmosphere heat exchange and monitoring under-ice habitats. Currently, sub-surface ice mapping is predominantly achieved through acoustic sensors such as multibeam sonar. While sonar provides robust large-scale geometry, it lacks spectral and high-resolution textural information required to record fine morphological features - a critical limitation in under-ice environments, where sparse visual features and degraded sensing conditions challenge the capture of detailed structural and optical surface properties. We present a
benchmark for
vision-based under-ice 3D reconstruction using real-world data from the MOSAiC expedition and synthetic data generated in an Unreal Engine-based underwater robotics simulator. The
benchmark evaluates Structure-From-Motion (SfM) and monocular depth estimation, with quantitative comparison enabled through cross-modal registration to upward-looking multibeam sonar.
Read the full abstract ▾
All authors Klára Orbán, Ashwin Nedungadi, Ana-Ştefania Kogălniceanu, Bianca-Marcela Sava, Bastian Cieply, Raul-Alexandru Simon, Kuderna-Iulian Benţa, Stefan Lüdtke
Solar and Stellar Astrophysics
Understanding the Complex Morphology of a Coronal Mass Ejection. III: How to Create a Severe Geomagnetic Storm
Cecilia Mac Cormack, Abril Sahade, Erika Palmerio et al. (+4 more)
On April 21, 2023, a complex filament eruption produced an Earth-directed coronal mass ejection (CME) that triggered the first severe geomagnetic storm of solar cycle 25. The primary objective of this work is to understand the physical origin of the severe impact of this CME on Earth. We explore the CME's complex morphology by simulating the eruption using the thermodynamic magnetohydrodynamic (MHD) model CORHEL-CME. We model the pre-eruptive structure as two intertwined
magnetic flux ropes and validate the early
evolution of the event by comparing synthetic observables generated from the model with extreme-ultraviolet and white-light observations. We then compare the simulated CME structure with in situ measurements. From our analysis, we conclude that the CME merged with a stream interaction region (SIR) before reaching Earth. This interaction compressed the CME sheath, producing a sustained negative $B_n$ component that, together with the associated enhancement in solar wind density and velocity, triggered the geomagnetic storm. The internal configuration of the
magnetic flux rope, which also maintained a sustained southward component during its propagation toward Earth, intensified the storm to severe levels. This model-data comparison underscores the power of using MHD simulations, when validated by multi-viewpoint observations, to reveal hidden dynamical processes and improve the understanding and prediction of space weather impacts.
Read the full abstract ▾
All authors Cecilia Mac Cormack, Abril Sahade, Erika Palmerio, Evangelia Samara, Phillip Hess, Robin C. Colaninno, Teresa Nieves-Chinchilla
Computation and Language
Similar Models, Broken Odds: Where Luce's Axiom Fails in Chatbot Arena
Soovadeep Bakshi, Apurba Bose, Aishwarya Bose et al. (+1 more)
Chatbot Arena ranks
language models with the Bradley-Terry model, which gives each model one score and so assumes Luce's choice axiom. Choice theory predicts where this fails: near-copies are told apart more decisively than their scores imply. On 1.8 million Arena votes, votes between versions of one model are 1.37 times as decisive as Bradley-Terry predicts, between sizes of one generation 1.26 times and between models of one lab 1.12 times; a release a year later gives consistent multipliers, and the effect survives alternative tie treatments,
prompt subsets and topics, and timing checks. How alike two models' answers are worded and formatted tracks the effect. A correction with four to six parameters predicts held-out votes better than Bradley-Terry and than models with hundreds of parameters, mostly by modelling ties; frozen and carried to new releases in a test fixed in advance, its gains are concentrated in ties, and better head-to-head odds are not established. The leaderboard order barely moves. With
language-model judges and three-answer menus, a near-copy takes votes mostly from its twin when choices are pooled over
prompts (log-odds shift +0.10, 95% interval 0.05 to 0.16); within a
prompt the shift is smaller and clear for two of four judges.
Read the full abstract ▾
All authors Soovadeep Bakshi, Apurba Bose, Aishwarya Bose, Sandip Nandi
Computational Physics
A deep energy method formulation for phase-field modelling of ductile fracture
Nojus Plungė, Peter Brommer, Rachel S. Edwards et al. (+1 more)
This work extends the Deep Energy Method to phase-field modelling of ductile fracture in elasto-plastic solids. Its key contribution is an incremental DEM-based formulation that couples history-dependent J2 plasticity and plasticity-dependent fracture toughness with separate energy-optimised displacement and phase-field representations. The formulation combines the second-order Ambrosio-Tortorelli phase-field fracture model with small-strain J2 plasticity and linear isotropic hardening. A volumetric-deviatoric split defines the elastic crack-driving energy, while accumulated equivalent plastic strain governs the degradation of local fracture toughness. Plastic internal variables are updated locally through radial return. The reduced energy is minimised alternately with respect to the displacement and phase fields, represented by a ReZero residual neural network and spatially localised Gaussian radial basis functions with trainable centres, respectively. This split addresses the diffuse phase-field profiles observed in preliminary calculations with a shared global network. Alternating optimisation is combined with transfer learning to reuse trainable parameters and optimiser states across loading increments, while the converged constitutive history is carried forward separately at quadrature points. The numerical results demonstrate close agreement with finite-element reference solutions in predicting force-displacement responses, energy
evolution, plastic deformation and crack propagation. The framework captures the main features of ductile fracture, although accuracy becomes more sensitive to loading increments near abrupt fracture events.
Read the full abstract ▾
All authors Nojus Plungė, Peter Brommer, Rachel S. Edwards, Emmanouil G. Kakouris
Combinatorics
Finding non-bisimplicial and triad-free minimal separators
Maria Chudnovsky and Ilya Maier
A $u\hbox{-}v$ separator in $G$ is a set $S\subseteq V(G)\setminus\{u,v\}$ such that there is no $u\hbox{-}v$ path in $G\setminus S$. A minimal separator in $G$ is a set $S$ such that for some $u,v\in V(G)\setminus S$, $S$ is a $u\hbox{-}v$ separator in $G$ but no proper subset of $S$ is. A set $S$ is bisimplicial if it is the union of at most two cliques. Recognizing graphs in which every minimal separator is the union of at most $k$ cliques is
polynomial-time solvable for $k=1$ and NP-hard for every fixed $k\geq 3$. We settle the remaining case $k=2$ by giving a
polynomial-time
algorithm to decide whether every minimal separator in a graph is bisimplicial. A set $S$ is triad-free if its independence number $\alpha(S)$ is at most two. Given a graph $G$ and two vertices $u,v$, the problem of finding a $u\hbox{-}v$ separator $S$ with $\alpha(S)\leq k$ had the same complexity gap: it was known to be
polynomial-time solvable for $k=1$ and NP-hard for every fixed $k\geq 3$. We resolve the remaining case $k=2$ by giving a
polynomial-time
algorithm that finds a triad-free $u\hbox{-}v$ separator or determines that none exists.
Read the full abstract ▾
Numerical Analysis
Neural Spline Flow Diffeomorphisms for Pullback Finite Element Methods on Complex Domains
Ioannis Kalogeris, Anaksagoras Spitas, Konstantinos Atzarakis et al. (+2 more)
We propose a method for solving partial differential equations on complex domains by learning smooth, invertible mappings from a fixed reference domain using Neural Spline Flows (NSFs). Monotone rational-quadratic spline coupling layers ensure a strictly positive Jacobian determinant by construction, eliminating the need for Jacobian-sign penalties. A diffeomorphic pullback transfers the weak form of the PDE to the reference domain, enabling finite element analysis on reusable structured meshes. To generalize across shapes, a shared NSF is conditioned on low-dimensional latent codes learned through an auto-decoder framework; a subsequent encoder predicts these codes for unseen geometries in a single forward pass. For a given learned map, the pullback preserves the mathematical structure of the PDE and is exactly equivalent to solving on the domain represented by that map. Because the learned domain generally approximates the target geometry,
geometric mapping error remains distinct from finite element discretization and quadrature errors. Numerical
benchmarks show that sufficiently accurate learned mappings yield reliable finite element solutions. By avoiding repeated generation of unstructured meshes, the method addresses a bottleneck in simulation workflows. The fixed reference mesh also creates opportunities for efficient assembly,
memory-aligned data access, and precomputed quadrature. The resulting framework assigns
geometric complexity to the neural network while retaining conventional finite element discretization and convergence properties on the learned domain, with overall accuracy also governed by
geometric fidelity.
Read the full abstract ▾
All authors Ioannis Kalogeris, Anaksagoras Spitas, Konstantinos Atzarakis, George Stavroulakis, Vissarion Papadopoulos
Graphics
DiffGIPC: Differentiable Mixed-Coordinate Contact Dynamics on GPU
Kemeng Huang, Bowen Ren, Xinyu Lu et al. (+2 more)
Contact-rich inverse problems require efficient gradients and reliable enforcement of non-penetration. Existing differentiable methods provide fast simulation and rigid-deformable coupling, but contact treatments based on penalties or velocity-level constraints can permit penetration or positional drift. Incremental Potential Contact (IPC) maintains non-penetration through
geometric barriers and {continuous collision detection (CCD)}. However, repeated forward and backward evaluations remain costly in full-space finite-element (FEM) implementations, particularly when nearly rigid components retain many vertex {degrees of freedom (DoFs)}. We present DiffGIPC, a GPU-accelerated differentiable IPC framework for optimizing interacting deformable and nearly rigid bodies. Our affine-body derivatives enable {rest-shape} design, initial-state optimization, and articulated control while retaining only twelve motion DoFs per affine body. These gradients account for how changes in geometry, placement, and actuation affect other objects through collision and friction. We preserve these interactions in one coupled FEM-affine adjoint system per frame, reducing the backward-system size without discarding the dependencies that drive the inverse problem. GPU-optimized backward computation further reduces the cost of repeated gradient evaluations. A reusable PyTorch interface connects supported physical parameters to user-defined parameter maps and losses across design, identification, motion, placement, and articulated control. Experiments demonstrate peak speedups over DiffIPC of approximately 150.0x with full-space FEM and 4133.3x with affine bodies for combined forward simulation and backward differentiation.
Read the full abstract ▾
All authors Kemeng Huang, Bowen Ren, Xinyu Lu, Minchen Li, Taku komura
Soft Condensed Matter
Geometry-Encoded Hidden Orientational Memory
Vladimir A. Baulin
Materials usually remember their past through dissipative changes of structure or through a net order. Simulations of caged colloidal ellipsoids at a fluid interface reveal a purely orientational
memory. The particles stay in place, the layer has no net
alignment, and a pattern written by a field is retained by capillary interactions between neighbors. It lasts hundreds of times longer than a particle takes to reorient, while every particle keeps rotating. The layer also remembers sequence: when two strong fields act one after the other on the same particles, it records which came last.
Read the full abstract ▾
Robotics
DexMimic: A Generalist Dexterous Hand-Object Tracking Policy via Value-Guided Reset Allocation
Zhexi Luo, Zhixuan Xu, Rui Yang et al. (+3 more)
Achieving human-like dexterous manipulation remains a central challenge in robotics. Learning from human hand-object interaction (HOI)
datasets offers a promising pathway to diverse dexterous skills, yet existing physics-based approaches train specialist policies for individual trajectories, thereby limiting skill reuse at scale. We introduce DexMimic, a generalist dexterous hand-object tracking policy jointly trained across diverse HOI trajectories. Efficient joint training requires allocating each rollout to both an undertrained phase and a starting state that supports continued tracking. DexMimic realizes this insight through Value-Guided Reset Allocation, which adapts the initialization distribution online to prioritize difficult phases and select reset states with higher expected tracking returns. We further introduce a
benchmark spanning training, fine-tuning, and generalization. Experiments show improved tracking success and training efficiency, faster fine-tuning, and the potential for zero-shot generalization to unseen trajectories and objects; larger training sets further improve efficiency and generalization. Finally, we validate DexMimic on a real-world dexterous manipulation platform.
Read the full abstract ▾
All authors Zhexi Luo, Zhixuan Xu, Rui Yang, Zeyu Xiao, Yiwen Hou, Lin Shao
Logic
Measures in Nairian models, I
Gabriel Goldberg and Grigor Sargsyan
Assuming $\mathrm{AD}^+ + V = L(P(\mathbb{R}))$, we show that the Ultrapower Axiom holds in $\mathrm{HOD}_s$ for every $\omega$-sequence of ordinals $s$. We use this result to analyze the structure of Nairian models, showing that they contain many measurable cardinals. As an application, we establish the consistency of the existence of infinitely many consecutive measurable cardinals with ZF, starting with a Woodin limit of Woodin cardinals. This solves a question of Henle, Mathias, and Woodin. From the same large cardinal hypothesis, adapting a forcing technique of Adkisson and Ben-Neria, we produce a model of ZF in which every uncountable cardinal is singular. Gitik's original construction requires a proper class of strongly compact cardinals.
Read the full abstract ▾
Computational Engineering
Linguistic Passiveness in Earnings Calls and Stock Returns
Lynn Tao, Yuntao Wu, Charles Martineau et al. (+2 more)
We study linguistic passiveness as the fraction of passive-voice sentences in earnings conference call Q&A sessions and show it is informative for equity
markets. Using 105,642 U.S. calls from 2008 to 2024, passiveness is negatively associated with announcement-day returns among positive-surprise firms: returns fall from 4.38% to 2.74% across passiveness quintiles. This announcement-day discount fully reverses over the subsequent quarter, indicating a temporary mispricing when news is good. For negative-surprise firms, passiveness has no announcement-day effect but predicts negative long-run returns, consistent with investor under-
attention when salient bad news dominates. Passiveness remains informative after controlling for sentiment and standard readability proxies, suggesting that linguistic passiveness captures a distinct yet important dimension of managerial communication.
Read the full abstract ▾
All authors Lynn Tao, Yuntao Wu, Charles Martineau, Vincent Grégoire, Andreas Veneris
Analysis of PDEs
Sharp Convergence Rate for stationary states of the Porous Medium Equation
Igor Pinheiro and Filippo Santambrogio
In this paper we prove the sharp rate of convergence for stationary states of the porous medium equation with an external drift in the p-Wasserstein metric under very mild assumptions on the potential and for any $1\leq p\leq +\infty$.
Metric Geometry
Blaschke–Santaló inequalities with respect to costs
Alexander V. Kolesnikov, Galyna V. Livshyts and Liran Rotem
We study functional Blaschke–Santaló inequalities in which the inner product is replaced by a cost. An inequality on each orthant, combined with Steiner symmetrization, gives a sufficient condition for a cost inequality to hold for even convex functions. We apply this principle to a multiplicative cost defined on all orthants. Its second variation gives a weighted improvement of the Brascamp–Lieb inequality in the directions for which differentiation is justified. We also prove an $L^p$ version of the Fradelizi–Meyer
theorem from a weighted volume-product inequality and derive the corresponding radial cost and Brascamp–Lieb inequalities. Explicit examples include anisotropic power and Cauchy densities, with dilation equality cases and a sharp obstruction below the coordinate exponent two. For the spherical logarithmic cost we compute an exact
entropy–logdet deficit decomposition and the second-order deficit transverse to ellipsoids. We also explain the transport-
entropy formulation and distinguish the restricted spectral question from the scalar $(B)$ property.
Read the full abstract ▾