Computer Vision
Human-Inspired Asymmetric Collaboration for Dual-Student Semi-Supervised Medical Image Segmentation
Huiyi Ren, Ziming Fang, Jiaming Cai et al. (+2 more)
Dual-student learning enriches predictive diversity in SSMIS by introducing a strongly augmented view alongside a weakly augmented view. Insufficient strong augmentation may homogenize the students, whereas excessive augmentation can corrupt anatomical evidence and destabilize collaboration. This challenge echoes feedback-regulated human instruction, where learners receive differentiated tasks and the task difficulty is continually adjusted according to their evolving responses. Inspired by these, we propose Human-Inspired Asymmetric Collaboration (HAC), a closed-loop dual-student framework that maintains weak–strong learning through adaptive difficulty calibration and geometry preservation. Specifically, Reinforcement-guided Difficulty Calibration (RDC) formulates strong-augmentation control as a window-level reinforcement learning process, using augmentation history, realized perturbations, and training responses to dynamically adjust the augmentation magnitude, operation number, and magnitude range. Meanwhile, the Gramian Geometry Preservation Constraint (G2PC) preserves the early-stage three-branch geometry through global volume and pairwise area constraints, preventing shared supervision from homogenizing the two students. Comprehensive evaluations across LA, Pancreas, and BraTS-2019 validate both the segmentation effectiveness and capability-aware difficulty calibration of HAC.
Read the full abstract ▾
All authors Huiyi Ren, Ziming Fang, Jiaming Cai, Rui Xu, Haojie Li
Metric Geometry
Excess inertia of finite metric spaces
Kiyonori Gomi
This paper classifies the possible inertias that are realized by an excess
matrix of a finite metric space with at most nine points. For the classification, we make use of relationships between finite metric spaces and non-obtuse and acute sets in classical geometry. As an application, we prove that the small-scale limit of the magnitude function of any six-point metric space is at least one whenever this limit exists.
Read the full abstract ▾
Quantum Physics
Spectrally enhanced gravitational redshift signals in photonic interferometry
Jerzy Paczos, Piotr T. Grochowski, Natalia Bruno et al. (+3 more)
Quantum mechanics and general relativity are remarkably successful, yet their joint predictions remain beyond direct experimental reach. Photonic interferometers operating in a gravitational field offer a promising route to test this regime: gravitational
redshift differentiates
photon propagation between vertically separated arms, leading to a predicted phase shift and reduction in interference visibility. These effects are minute on laboratory scales, motivating existing proposals to consider large interferometer areas to accumulate them geometrically. Indeed, such schemes typically require kilometre-scale optical paths to produce appreciable interferometric signals. Here, we propose a strategy that replaces
geometric accumulation with spectral enhancement by coupling the two interferometer arms to optical scatterers based on ultranarrow atomic transitions. These transitions serve as local spectral references: gravitational
redshift changes the
photon's detuning from each resonance, producing large differences in reflection amplitude and hence substantial visibility loss and interferometric phase shift already for centimetre-scale height differences. We quantify the experimental requirements and assess the feasibility of implementation with current or near-term technology. Our proposal thus opens a route to tabletop optical tests of the interplay between
quantum mechanics and general relativity.
Read the full abstract ▾
All authors Jerzy Paczos, Piotr T. Grochowski, Natalia Bruno, Emanuele Distante, Stephan Welte, Magdalena Zych
Quantum Physics
Distance Prediction for Quantum Codes
Nadav Schneider and Eliya Nachmani
The distance of a
quantum code is a fundamental metric that dictates its error-correction capability. While
quantum low-density parity-check (qLDPC) families such as Bivariate Bicycle (BB) codes offer a promising pathway to low-overhead fault-tolerant
quantum computing, computing their distance is challenging. Current exact solvers and even search heuristics that provide a distance upper bound suffer from severe computational bottlenecks, restricting the discovery of optimal codes. In this paper, we construct a
dataset of BB codes with corresponding distance upper bounds obtained through high-budget heuristic searches. We introduce BB-Net, an
attention-based neural network architecture for predicting the distance of BB codes. It achieves 96.39% accuracy on the reference distances in the test set, compared with 94.86% for state-of-the-art heuristics with a budget of up to $10^{4}$ times more computational runtime. On a subset with certified exact distances, BB-Net achieves 98.94%. By relieving the distance-calculation bottleneck, BB-Net enables faster screening of large code spaces, particularly in ML-driven search pipelines, and thereby can accelerate the discovery of novel
quantum error-correcting codes, an important capability for advancing scalable fault-tolerant
quantum computing. Our code is publicly available: https://github.com/NadavSc/BBNet
Read the full abstract ▾
Operator Algebras
Extensions of Operator Systems, Hyperrigidity and Quantum Double Suspension
Vibhor Bhatt, Satyajit Guin and Palash Sharma
We introduce a class of extensions of unital operator systems satisfying a certain Toeplitz type property. We show that these extensions induce corresponding extensions of the $C^*$-
algebras they generate. We prove that hyperrigidity is preserved under these extensions, yielding new examples through iteration. We introduce the
quantum double suspension of operator systems as a special class of these extensions and show that its iterates yield the Hong-Szymański
quantum double suspensions at the $C^*$-
algebra level, encompassing, in particular, the Vaksman-Soibelman
quantum spheres. We construct hyperrigid operator systems underlying these
quantum spheres. As a further application, we show that every counterexample to Arveson's hyperrigidity
conjecture gives rise to a sequence of counterexamples. We identify the $C^*$-envelopes of these extensions and use this identification to obtain a $K$-theoretic condition ensuring that each such sequence contains infinitely many pairwise non-isomorphic counterexamples. We also observe that the validity of the
conjecture in the commutative setting is linked to its validity for an infinite family of noncommutative operator systems obtained through successive extensions. Finally, we compute the propagation number, an invariant of operator systems introduced by Connes and van Suijlekom, for these extensions.
Read the full abstract ▾
Statistical Mechanics
Organisation and aggregation of calcium carbonate ions in water via molecular simulation
Elizaveta Sidler, Rita S Dias and Raffaela Cabriolu
Crystallisation of calcium carbonate (CaCO$_3$) in aqueous solutions is essential for carbon dioxide sequestration and the sustainable production of calcium carbonate materials, which are widely used in industry and increasingly explored for novel applications. While crystallisation is often analysed within the framework of classical nucleation theory, CaCO$_3$ and other ionic
crystals often form through amorphous precursor phases, motivating the need for alternative descriptions. In this work, we investigate the relationship between the formation of the first CaCO$_3$ coordination sphere, a fundamental unit of local ordering, and ionic aggregation using two complementary simulation approaches: on-the-fly probability-enhanced sampling (OPES) and asynchronous and infinite-swap replica exchange transition interface sampling ($\infty$RETIS). OPES was used to explore free-energy landscapes, whereas $\infty$RETIS was used to determine reaction kinetics. We studied systems at two levels of supersaturation, between which the formation of the first coordination sphere exhibited distinct kinetic behaviour. These differences were found to be linked to concentration-dependent ionic aggregation. At lower concentration, aggregation remains limited, whereas higher concentration promotes extensive spinodal-like aggregation. For both concentrations, the first coordination sphere forms within the pre-formed clusters through one of two distinct pathways: either within a cluster close to the minimum size required for its formation, or within larger amorphous aggregates whose size and structure depend on concentration. The coexistence of these pathways at both levels of supersaturation reveals the multifaceted nature of early-stage CaCO$_3$ crystallisation and highlights the importance of differentiating ionic aggregation and organisation.
Read the full abstract ▾
Artificial Intelligence
Monitoring latent reasoning in looped language models
Pawel Olszowiec, Paulina Tomaszewska, Maciej Żelaszczyk et al. (+2 more)
Looped
language models (LMs) reuse one block of layers several times per
token generation, promising more latent computation at no additional storage cost. Despite their rising popularity, the analysis of looped LMs is still in its infancy: little is known about whether the iterative transformations actually improve their capabilities, or whether these models remain stable under
prompt perturbation. We construct an inspection protocol rooted in Mechanistic Interpretability comprising five diverse lenses. We apply it to examine eight looped LMs from four families. Models differ in many aspects but the key one is whether the
prompt is transmitted only within the looped state (carry loops: Ouro, LoopUS) or re-supplied at every iteration (re-read loops: Parcae, and Rec-OLMo & Huginn from the Raven family). The results of our study provide valuable insights concerning safety and efficient use of looped
language models.
Read the full abstract ▾
All authors Pawel Olszowiec, Paulina Tomaszewska, Maciej Żelaszczyk, Michal Byra, Neo Christopher Chung
Artificial Intelligence
Fresh at Decision, Stale at Action: Routing New Information to LLM Decisions Already Made
Ayoub Merimi, Souhail Bakkali and Abdeljalil Sakat
Freshness shapes an LLM decision at the moment it is made; the action it commits to is latent, carried out later, in a changing world. We therefore introduce a timestamp theory of decision validity and a decision ledger that realises it. Six timestamps define a decision; a freshness ceiling bounds what up-to-date evidence guarantees at action time, placing the remaining risk in information published afterwards. Accuracy then factorises into two capabilities: routing, which delivers that information to affected decisions before their revision deadlines, and temporal judgment, which reads it for the time of action. The ledger is persistent state holding each decision's times, evidence and background knowledge, indexed by watch conditions written at decision time. Controlled decision streams with exact labels confirm four predictions across five LLMs, from 8B to frontier scale: freshness alone yields exactly the accuracy the ceiling predicts; models act on decisive news once it arrives, so accuracy follows routing; the ledger's index sets what routing reaches, and watch conditions reach indirect dependencies at low exposure; and the remaining errors concern when a change takes effect. Freshness makes a decision right when it is made; routing and temporal judgment keep it right when it is carried out.
Read the full abstract ▾
Computer Vision
H-DDCM: Hadamard Codebooks for Fast Generative Image and Video Compression
Or Ronai, Vladimir Kulikov and Tomer Michaeli
Training-free
diffusion-based compression methods repurpose the generative priors of a
diffusion or a flow model to compress data without retraining them as task-specific codecs. Recent codebook-based approaches such as Turbo-DDCM achieve strong perceptual quality at ultra-low bitrates, but their reliance on dense Gaussian codebooks remains costly in both
memory and computation, and scales poorly with the large latent dimensions of modern image and
video generators. Their combinatorial bitstream encoding further becomes computationally heavy at this scale. We introduce H-DDCM, a scalable generative compression framework that replaces Gaussian codebooks with implicit randomized Hadamard dictionaries. The fast Walsh-Hadamard transform allows for implicitly synthesizing full-dimensional dictionaries with $O(d \log d)$ compute for projection and reconstruction and $O(d)$
memory, without materializing the full-dimensional dictionary. We further introduce an efficient table-free lexicographic coding scheme that makes combinatorial encoding practical at larger dimensions. Beyond images, we adapt the framework to an autoregressive
video flow model and develop a streaming compression scheme that periodically re-encodes recently reconstructed frames into a fresh KV-cache prefix, preserving temporal context across segments at no additional transmitted bits. This enables long-form, close to real-time generative
video compression on a single GPU. Across image and
video benchmarks, H-DDCM achieves competitive rate-distortion-perception performance while substantially reducing the
memory and computational overhead of codebook-based training-free generative compression.
Read the full abstract ▾
Quantum Physics
Local error correction prolongs time-crystalline memory
Benedikt Fauseweh
Discrete time
crystals are non-
equilibrium phases of matter with spatiotemporal order and a response that breaks the time-translation
symmetry of a periodic external drive. Interactions, spatial disorder and correlated random driving can in theory stabilize this order. In experiments, however, coupling to the environment limits the time
crystal lifetime. Here we use
stochastic local bit-flip error correction to prolong the time-crystalline response of a driven, spatially disordered two-dimensional Ising model on a superconducting
quantum processor, using a repetition code with distances three and five on a honeycomb
lattice with up to 68 physical
qubits. Even though ancilla
qubits, midcircuit syndrome measurements and decoding increase the required
quantum and classical resources, the
stochastic drive increases the
memory decay time by a factor of up to two. At late times we observe that recovery stabilizes spatial order across the
lattice. The
spin-configuration distributions explain this separation, as recovery suppresses disagreeing bonds while broadening occupies completely reversed states. Syndrome measurements without correction accelerate the
memory decay, demonstrating the importance of conditional correction. Our results show that local bit-flip error correction can preserve time-crystalline
memory over more drive cycles than the bare dynamics.
Read the full abstract ▾
Robotics
MARTS: A Modular Agricultural Robot for Targeted Spraying of Diseased Tomato Plants in Greenhouses
Rajmeet Singh, Manveen Kaur, Yusuf Adamu et al. (+2 more)
Broadcast spraying remains the dominant method of disease control in greenhouse tomato production, yet it treats healthy and diseased tissue alike, exposes workers to hazardous aerosols, and depends on operator skill. This paper presents MARTS (Modular Agricultural
Robot for Targeted Spraying), a low-cost robotic platform that closes the loop from plant-level disease detection to leaf-localised spray delivery in commercial greenhouses. MARTS integrates a Clearpath Husky A200 mobile base, a UR5 manipulator, a novel modular spraying end-effector built around a single 12 V micro-diaphragm pump, and a wrist-mounted Intel RealSense D435 RGB-D camera. Disease detection reuses our previously published Tomato Leaf Disease
Vision Language Model (TLDVLM; Grounding DINO, SAM-2 and LoRA-tuned BLIP-2, 97.27% accuracy over 10 classes); this work focuses on actuation, control and system integration. A ROS pipeline lifts detections into 3D targets, generates collision-aware nozzle poses at a 120 mm stand-off using fitted leaf normals, and scales pump pulse duration to lesion area. Water-sensitive paper calibration showed a linear pulse-volume relationship (R^2 = 0.994). In MuJoCo simulation, an adaptive base-repositioning strategy achieved 100% coverage of 18 waypoints with a maximum orientation error of 1.1 deg. Deployed at the SILAL commercial greenhouse (Al Ain, UAE) on cherry tomatoes under natural infection, MARTS treated 312 diseased leaves in 12 autonomous trials with 92.4% targeting success, 18.7 mm mean spray-centre error, 6.1% adjacent-leaf overspray, and a 71% reduction in pesticide volume versus broadcast spraying. Mean cycle time was 4.8 s per leaf, and the spraying subsystem costs USD 412, 6-10x lower than comparable systems, showing that accurate leaf-level targeting is achievable on commodity hardware.
Read the full abstract ▾
All authors Rajmeet Singh, Manveen Kaur, Yusuf Adamu, Lochan Kshetrimayum, Irfan Hussain
Computation and Language
Cheap, Consistent, Calibrated, but Biased? Typed Decision Models as Judges
Fabian Lukassen, Christoph Weisser and Alexander Silbersdorff
Typed decision models (TDMs) return a probability for each allowed answer instead of generating text, which makes them candidates for cheaply replacing LLM judges that compare two answers. Whether they share the biases of LLM judges is untested. We run the CALM
benchmark's bias tests on eight TDMs and seven LLM judges of 2024 and 2026. Every TDM is biased, in three to eight of twelve tests. Emotional tone affects them most, and rewriting the better answer in such a tone lowers the accuracy of a typical TDM from 99% to 42%. TDMs are more biased than 2026 LLM judges. All eight change verdicts when the answers swap places, which the 2026 LLM judges hardly do, and several side with an answer that most people allegedly prefer or that a famous model allegedly wrote. Yet both kinds of judge are tripped up by the same
prompts. The TDMs' probabilities show that most tests mainly flip verdicts the TDM was unsure of, and that good calibration does not protect against bias. As TDMs are more consistent and 37 times cheaper than 2026 LLM judges, our results suggest they can be attractive judges if their biases are tested for the task.
Read the full abstract ▾
Computation and Language
Narrative Lookahead in LLMs: Event-Level Decodability, Predictive Advantage, and Steering
Donghoon Jung, Min Kim, Seojin Choi et al. (+2 more)
Do
language models carry information about narrative events before writing them, and how far ahead? We examine narrative lookahead through salient, state-changing events rather than
tokens or response-level attributes. We extract 23,355 LLM-annotated events from 2,000 stories generated by four open-weight LLMs, labeling them along three nonexclusive dimensions of change: within
agents, between
agents, and in the story world, with human validation on a subset. Event segmentation and plot-independent labels enable comparisons across stories at horizons measured in events rather than in
tokens. At event ends, we probe hidden states for upcoming event types and compare predictions with text-only predictors using matched prefixes. Across four models, future event types are decodable above prevalence baselines for at least four subsequent events, but probes significantly outperform both primary text-only predictors only for the next event. In exploratory steering with OLMo-3.1-32B, a probe-derived direction shifts the proportion of the target event type in both intended directions at higher strengths. Our event-level approach links the prediction of future narrative developments to interventions that reshape subsequent event composition.
Read the full abstract ▾
All authors Donghoon Jung, Min Kim, Seojin Choi, Songeun Chae, Seohyon Jung
Audio and Speech Processing
Self-Supervised Source-Filter Model of Speech Production
Nani Meka, Raghavan Ramesh and Sri Rama Murty Kodukula
Deep learning models of
speech seldom represent the source-filter process by which
speech is produced, relying instead on latent features with no physical correspondence to articulation or phonation. We propose SoFilNet, a self-supervised, differentiable Source-Filter Network that estimates the parameters of
speech production directly from the raw waveform: a source, specified by fundamental frequency (F0), voicing and the aperiodicity of five frequency bands, and a filter, specified by ordered formant poles, additional poles that shape the rest of the spectral envelope, and antiresonance zeros. The model is trained only to reconstruct the input
speech; no formant or pitch annotation is used. Since per-bin spectral losses favor blurred spectra wherever the model is uncertain, we train F0 and voicing against autocorrelation pseudo-labels refined with the model's own filter, aperiodicity against the periodic fraction of each band, and formant sharpness against spectral contrast. On VTR-FORMANTS, the estimated formants attain the lowest mean error among the reported methods, including supervised trackers, and on the unseen PTDB-TUG corpus the F0 frame error matches the best classical tracker and is lower than that of all learned trackers. The parameters are also consistent with
speech production theory: aperiodicity rises with frequency and orders phone classes by the
turbulence of their excitation, and the zeros deepen on nasal consonants, the lower one rising with place of articulation.
Read the full abstract ▾
Logic in Computer Science
Certifying Almost-Sure Reachability in Unknown MDPs
S Akshay, Suguman Bansal and Gaurav Mahla
Specification-Guided Reinforcement Learning (RL) has largely concerned itself with learning an optimal policy against a formal specification in an unknown Markov Decision Process (MDP). We take up the dual question. Given a policy already in hand, can one verify that it is correct against a formal specification in an unknown MDP, using sampling access alone? Such a question arises naturally in safety-critical systems. We investigate it for the almost-sure reachability property, whose correctness is witnessed by a ranking supermartingale (RSM) certificate. Our first finding is negative. In the absence of any further information about the MDP, one can neither learn such a certificate nor verify a candidate one with probabilistic guarantees. On the positive side, we show that, under a mild restriction, a known lower bound $p_{\min}$ on nonzero transition probabilities, a correct certificate can be learned with PAC-style guarantees. To the best of our knowledge, this is the first study of learning and verifying certificates for almost-sure reachability in unknown MDPs from sampling access alone. Finally, our experiments show that the $p_{\min}$-based sufficient condition, while sound, is conservative: the learned certificate often remains valid on the true MDP even when the condition fails.
Read the full abstract ▾
Robotics
ForeTac: Predictive Contact Guidance for Generative Robot Policies
Zhiyuan Chen, Shuai Chen, Haoran Zhou et al. (+7 more)
Visually similar actions can produce contact loss, jamming, slippage, or irreversible damage, while tactile feedback may arrive after a critical transition. Contact-rich manipulation therefore requires generative policies to anticipate and evaluate whether candidate actions will produce desirable contact outcomes, then adjust action generation before execution. We present ForeTac, a predictive contact-guidance framework organized as a represent-predict-evaluate-guide loop. ForeTac encodes the current tactile state, predicts the tactile consequences of a candidate action sequence, and evaluates the predicted trajectory using a contact-quality objective learned from successful and failed physical interactions. At deployment, its gradient propagates through the tactile predictor to guide action generation while the pretrained model parameters remain frozen. We evaluate ForeTac on four real-
robot tasks (planar and curved-surface wiping, card swiping, and fragile-object grasping) with
diffusion and flow-matching policies. Across these tasks, ForeTac raises the average task success rate from 36.3% to 71.3% and the average contact success rate (CSR) from 20.0% to 66.3% compared with
vision-only
Diffusion Policy. Component analyses support the tactile representation and action-conditioned foresight, and show that predicted touch preserves contact-quality ordering. Project page: https://foretac.github.io/.
Read the full abstract ▾
All authors Zhiyuan Chen, Shuai Chen, Haoran Zhou, Yizhe Zhao, Kaiyuan Yang, Zhixiang Duan, Shaoqing Xu, Ran Cao, Diyun Xiang, Yangwei You
Functional Analysis
Density of polynomials whose Aron–Berner extensions attain their norms
Javier Falcó
We establish a
polynomial analogue of Lindenstrauss'
theorem in full generality. More precisely, we prove that, for every real or complex Banach space X and every positive integer n, the n-homogeneous scalar-valued
polynomials whose Aron-Berner extensions to the bidual X** attain their norms form a dense subset of the space of continuous n-homogeneous
polynomials on X. This resolves a problem posed by Aron, Garcia and Maestre in 2003, extending their result for quadratic
polynomials to arbitrary degrees without imposing any structural assumptions on the underlying Banach space. We further establish the corresponding
theorem for vector-valued
polynomials with arbitrary Banach space codomains. Our approach relies on a new perturbation scheme based on averaged linear contractions, which provides simultaneous control of the multilinear variables and yields norm attainment through iterated limits in the bidual.
Read the full abstract ▾
Optics
Chirality Inversion and Giant Optical Activity in Scalable Self-Assembled Metamaterials
Antonio Carone, Aso Rahimzadegan, Birgül Kalkan et al. (+8 more)
Chiral materials manipulate light in ways that hold promise for sensing, imaging, and photonics, but producing such materials with strong and reproducible optical activity at large scale remains challenging. Here, we develop a scalable strategy to assemble hierarchical helicoidal metamaterials from metallic nanoparticles and chitin nanocrystals, a biosourced polysaccharide. Spontaneous co-assembly produces ordered chiral structures with strong plasmonic optical activity, reaching a dissymmetry factor of 0.55. The unprecedented ability to vary the helicoidal pitch allows control of the optical response, enabling inversion of optical handedness without changing the handedness of the template, effectively decoupling geometrical chirality from optical chirality. Remarkably, the gold nanoparticles can also be recovered from dry films and redispersed as colloids, introducing an appealing circularity concept for sustainable chiral plasmonic materials.
Read the full abstract ▾
All authors Antonio Carone, Aso Rahimzadegan, Birgül Kalkan, Thomas G. Parton, Tony Jin, Zihao Lu, Clemens Schmitt, Emeline Raguin, Bruno Frka-Petesic, Kevin Vynck, Silvia Vignolini
Neurons and Cognition
Sense-Making phase transition framework for Affect
Miguel Bosch, Laura Fanfarillo, Gustavo Diez et al. (+3 more)
Contemporary accounts of affect emphasize different perspectives, ranging from the regulation of bodily states in allostatic approaches to the relational and sense-making character stressed by enactive theories. Here, we show that affect can be made explicit within the sense-making phase transition framework previously introduced for the perception of an inner sensation. Within this same formal structure, each stable minimum of the sense-making potential defines an affective state: its position determines valence, while its depth determines arousal. Affect therefore is not introduced as an independent ingredient, but emerges as the associated state in the valence-arousal space, providing a link between allostatic and enactive approaches to affect. This formulation distinguishes valence intensity from arousal, allows for dominant and secondary affective states, and naturally accommodates mixed-valence configurations beyond a strictly bipolar description. Revisiting representative landscapes of the model, we show how neutral, uncertain, polarized, and mixed-valence regimes arise as a function of contextual sensory information, learned historical evaluation, and externally supplied information. We further show that the resulting valence-arousal relation provides a theoretical basis for the upper-open V-shaped organization discussed in the affect literature, while also opening a natural route toward a dynamical description of affect and emotions.
Read the full abstract ▾
All authors Miguel Bosch, Laura Fanfarillo, Gustavo Diez, Victor Gómez Mayordomo, J. Ricardo Arias-Gonzalez, Belén Valenzuela
Earth and Planetary Astrophysics
3D Global Climate Simulations of Synthetic Mars Warming
Tom N. Landefeld and Harald Lesch
Artificial warming via engineered aerosols presents a novel pathway to bypass volatile scarcity on modern Mars. Utilizing the Generic Planetary
Climate Model, we evaluate the physical feasibility of aerosol-induced terraforming over a full Martian year (669 Sols). Unlike previous approaches that added aerosols to background dust or assumed continuous nanoparticle injection, we simulate the complete substitution of natural dust with a single-injection profile of radiatively active aluminium nanoparticles without continuous dumping. This tests whether initial infrared greenhouse forcing can overcome the severe ice-albedo cooling induced by a massive regional water-ice sheet ($3400 \text{ kg/m}^2$) localized in the Hellas Basin. Comparing this scenario against an unperturbed control run, we analyze surface temperature dynamics, sublimation fluxes, and water vapour concentration. Our 3D global
climate simulations quantify the radiative-dynamical feedback thresholds and determine whether engineered aerosol forcing can trigger a self-sustaining water-vapour greenhouse effect or remains fundamentally limited by regional ice-albedo suppression in Hellas.
Read the full abstract ▾
Machine Learning
When Missingness Follows the Graph: Beyond Feature Propagation for Missing Features
Enrico Chialà, Francesco Ferrini, Antonio Longa et al. (+4 more)
Graph Neural Networks (GNNs) commonly assume that node features are fully observed, despite missing features being pervasive in real-world applications. By leveraging the dependence between node features and graph structure, existing graph-specific methods handle missingness by propagating observed feature values along graph edges. However, these methods have mainly been evaluated in overly optimistic settings, where feature entries are removed independently and uniformly across the graph. This overlooks realistic scenarios in which missingness propagates among connected nodes, as can occur during localized sensing failures in sensor networks. We formalize this complementary notion as graph-structured mask dependence, where the joint distribution of missingness indicators is shaped by the graph, and introduce a controlled mechanism that instantiates this setting. We study propagation-based methods theoretically and empirically in this setting, showing that their effectiveness changes substantially compared with the uniformly random regime. Motivated by this limitation, we propose PEMix, designed to remain effective under challenging graph-structured missingness. Rather than propagating observed values, PEMix uses global structural representations to condition a probabilistic model of incomplete node features before message passing. Experiments under controlled graph-structured missingness and naturally occurring missingness show that PEMix remains robust across complex missingness regimes, with particularly strong performance as missingness becomes severe.
Read the full abstract ▾
All authors Enrico Chialà, Francesco Ferrini, Antonio Longa, Bruno Lepri, Filippo Maria Bianchi, Manfred Jaeger, Veronica Lachi
General Relativity and Quantum Cosmology
Kantowski Sachs Cosmological Model with Radiation and Generalized Chaplygin Gas
G. Oliveira-Neto and R. Hincho Laucata
In the present paper we study the
evolution of a homogeneous and anisotropic Kantowski Sachs cosmological model coupled with radiation and a generalized Chaplygin gas. Our main objective is verifying if this model evolves to a homogeneous and isotropic Friedmann Robertson Walker cosmological model, which is the more accepted model to describe our present Universe. If this is the case, we may imagine that the Kantowski Sachs cosmological model could had been an early stage of our Universe.
Read the full abstract ▾
Solar and Stellar Astrophysics
Dynamical and thermal evolution of localized reconnecting current sheets under varying guide-fields in the solar corona
Saksham Pande, Abhishekh Kumar Srivastava, Fabio Bacchini et al. (+9 more)
Although
magnetic reconnection in coronal current sheets (CSs) is a fundamental energy release mechanism in the solar corona, the role of guide-fields in governing its formation and self-consistent
evolution is not fully understood. Through 2.5D resistive magnetohydrodynamic (MHD) simulations, we investigate how guide-fields affect the formation and
evolution of a localized reconnecting coronal CS. Strong guide-fields prevent oppositely directed field lines from approaching sufficiently to form a CS, delaying and suppressing its subsequent
evolution. The reconnection rate exhibits a rising phase resembling a power-law, with exponent $\delta_1$ decreasing with rising guide-fields, indicating slower temporal growth of the reconnection rate. The aspect ratio of the CS follows a similar power-law trend with the exponent $\alpha_1$ decreasing with rising guide-fields. A subsequent slower growth phase of the reconnection rate shows a power-law exponent $\delta_2$ that increases with rising guide-fields, as does the corresponding aspect-ratio exponent $\alpha_2$. As the CS evolves, the latter phase is achieved due to a sharp decremental transition of the guide-field within the CS after a time when the increasing inflows are inefficiently resisted by the guide-field itself. In the reconnection regime during prior phases, the power-law exponents of the maximum temperature and the outflow velocities also decrease with rising guide-fields demonstrating that stronger guide-fields suppress the growth of these indicators also. This study suggests that a self-consistently formed CS displays phases of efficient and inefficient quenching of the reconnection mechanism, promoted by non-zero guide-fields. We also quantify this behavior for indicators of the reconnection process and discuss their applicability to observations of apparent CSs and probing background coronal conditions.
Read the full abstract ▾
All authors Saksham Pande, Abhishekh Kumar Srivastava, Fabio Bacchini, Sripan Mondal, Akash Bairagi, Balveer Singh, Kartika Sangal, Sudheer K. Mishra, Xinping Zhou, Ding Yuan, R. -Y. Kwon, B. N. Dwivedi
Machine Learning
SYNPRED: A Synergistic Approach to Multimodal Learning for Clinical Prediction
Ivana Janíčková, Yen Y. Tan, Thomas H. Helbich et al. (+6 more)
Multimodal medical data captures complementary aspects of disease. It provides a comprehensive basis for diagnosis and the identification of biological associations linked to progression or treatment response. Established contrastive learning approaches exploit shared information across modalities, but fall short in exploiting complementary, modality-specific signals relevant for a task. Here, we propose a synergistic
multimodal variational autoencoder (VAE) framework that separates task-relevant unimodal and
multimodal information from instance-specific variation task-irrelevant, allowing complementary signals across modalities to be leveraged for prediction. The method yields benefits even if modalities are missing at test time. We evaluate the method on a treatment response prediction task using medical imaging and RNA data. Empirical comparison with
alignment-based and
multimodal VAE state-of-the-art approaches demonstrates improved
multimodal prediction performance, and consistent benefits for prediction from a single modality after
multimodal training. The full code can be found here: https://github.com/cirmuw/SYNPRED.
Read the full abstract ▾
All authors Ivana Janíčková, Yen Y. Tan, Thomas H. Helbich, Philipp Seeböck, Julieta Di Marco, Zsuzsanna Bago-Horvath, Ulrike Heber, Thomas Spiegel, Georg Langs
Digital Libraries
Interpreting Paper-to-Patent Citations in AI-Related PMC Articles
Haowei Li and Yi Bu
Paper-to-patent citations connect scientific and technological documents, but their interpretability depends on how patents are used in the citing text. This study examines citation contexts and paper-level outcome associations in AI-related articles in PubMed Central (PMC). Screening 189,073 articles identified 746 with explicitly tagged patent references (0.395%), including 713 with 1,544 identifiable in-text patent-citation instances. Three large
language models classified these instances by represented content, motivation, textual position, and evaluative stance. Regression models examined associations among these dimensions, and propensity-score-matched regressions compared patent-citing papers with XML-screened controls. Background-oriented citations were predominantly neutral and introductory, whereas adoption-oriented citations were more closely associated with methodological positions and supportive evaluations. These patterns suggest heuristic configurations of contextualization, operationalization, and problematization, although classification overlap and exclusion of rare categories limit their interpretation. In matched samples, the patent-citing indicator was associated with lower scholarly-citation measures and more citations received from patents. These associations are observational, not causal. Together, these findings show that paper-to-patent citations encode heterogeneous forms of engagement with patented technology and should be interpreted as selective bibliographic signals rather than direct indicators of technology-to-science knowledge transfer or research value.
Read the full abstract ▾
Machine Learning
A CaLLM Judge is a Better Judge: Mitigating Bias in LLM-as-a-judge
Cesare Barbera, Luca Stradiotti, Lorenzo Cascioli et al. (+6 more)
Large
language models (LLMs) are increasingly used as judges to predict which of two candidate answers a human evaluator would prefer, providing a scalable alternative when human evaluation is too costly. Unfortunately, LLM judgments can exhibit systematic biases. Prior work has documented position, verbosity or model family of the candidate answers influencing the judge decision rather than their quality, while additional biases may remain undocumented. $\textit{Multicalibration}$ provides a natural framework to address this problem: each bias corresponds to miscalibration on an identifiable group, $\textit{e.g.}$, family bias manifests as overconfidence in answers generated by models belonging to the same family of the judge. The challenge is to fix miscalibration simultaneously across many, potentially unknown groups. To this end, we introduce $\mathrm{CaLLM}$, a post hoc framework that multicalibrates an LLM judge, thereby bounding all such biases simultaneously. We prove that $\mathrm{CaLLM}$ bounds the bias of every group whose error structure is captured by an available representation, and that augmenting it with explicit bias groups further reduces the residual bias. We then derive theoretical conditions under which $\mathrm{CaLLM}$ improves the judge's accuracy, rather than trading it for calibration. Finally, we evaluate $\mathrm{CaLLM}$ across pairwise preference
benchmarks, showing that it reduces the multicalibration error on both specified and unspecified bias groups, while improving predictive performance of the judge.
Read the full abstract ▾
All authors Cesare Barbera, Luca Stradiotti, Lorenzo Cascioli, Lorenzo Perini, Niek Tax, Daniel Haimovich, Jesse Davis, Andrea Passerini, Andrea Pugnana
Instrumentation and Methods for Astrophysics
Charge Response Simulation of Liquid Argon Compton Cameras for Precise MeV Gamma-Ray Event Reconstruction
Shota Arai, Hirokazu Odaka, Satoshi Takashima et al. (+5 more)
We studied the detector response for a
liquid-argon time-projection chamber (LArTPC) Compton camera targeting MeV gamma-ray astronomy by a Geant4-based Monte Carlo simulation. The simulation, implemented within ComptonSoft, incorporates electron-ion recombination, charge
diffusion, and induction on the electrode pixels. We evaluated the angular resolution measure (ARM) and the energy reconstruction error as functions of electronic noise, pixel pitch, and the recombination correction method. An electronic noise level of 4 keV and a pixel pitch of 1 mm yield an ARM (FWHM) of about 3 degrees at 1332.5 keV and an energy reconstruction error better than 4% for incident energies of 1173.2 keV and above. Recombination correction using both scintillation light and charge signals improves the ARM and the energy error by factors of 1.7–2.3 and 2.2–5.4, respectively, compared to the charge-only correction, although the uncertainty of the recombination-fluctuation model remains. Since this simulation relies on idealized assumptions such as the use of the Monte Carlo truth in the event reconstruction, these values represent the best cases achievable under these detector conditions; nevertheless, they show that a large-scale LArTPC is a promising Compton camera for
cosmic MeV gamma-ray observations.
Read the full abstract ▾
All authors Shota Arai, Hirokazu Odaka, Satoshi Takashima, Masashi Tanaka, Ryutaro Tatsumi, Hodaka Kawamura, Kouichi Hagino, Aya Bamba
Group Theory
The one percent property for free groups
Emmanuel Breuillard and Ferdinand Jacobé de Naurois
We prove that for every $\varepsilon >0$, there is $\delta >0$ such that for every probability measure $\mu$ on a free group $\mathbb{F}$, if the operator norm of the action of $\mu$ on $L^2\mathbb{F}$ by convolution exceeds $\varepsilon$, then there is a coset of an abelian subgroup of $\mathbb{F}$ with $\mu$-mass at least $\delta$.
Read the full abstract ▾
Artificial Intelligence
Chain of Thought Monitoring Can Fail Even When CoT Is Necessary
Lennart Bürger, Piotr Komorowski, Moritz Weckbecker et al. (+4 more)
Chain of Thought (CoT) monitoring is used by frontier labs to detect misbehaviour of AI systems. Yet, its reliability remains poorly understood. We attribute this to the lack of a successful organizing principle that leads to generalizable conclusions about monitoring reliability in different regimes. Recent work attributes monitoring failures to the CoT not being necessary for the answer, and argues that CoT becomes monitorable when it is necessary. We show that this binary necessary/not-necessary organizing principle is too coarse by demonstrating that monitoring failures persist even if the CoT is necessary for the answer. In response, we introduce a more fine-grained organizing principle based on a hierarchy of four CoT necessity levels. Building on this hierarchy, we develop a framework for measuring CoT monitorability at the three lowest levels via intervention experiments. We demonstrate monitoring failures at all three of these levels on realistic, hard prediction tasks. On a difficult medical-diagnosis task, patient sex and a clinical attribute influence the answer in at least 19% and 24% of correct responses with a necessary CoT, respectively, yet their influence is almost never articulated. These results show that even when CoT is necessary, it is not guaranteed to be monitorable.
Read the full abstract ▾
All authors Lennart Bürger, Piotr Komorowski, Moritz Weckbecker, Evgenii Kortukov, Aakriti Jain, Sebastian Lapuschkin, Wojciech Samek
Machine Learning
The Effect of Layer Ordering in Hybrid Architectures
John Cooper, Mingchen Ma, Jiaqi Mo et al. (+4 more)
Hybrid
attention is rapidly becoming the dominant architectural choice for frontier models. Despite their successes, such architectures complicate model design by introducing additional choices, including the types and locations of layers. This combinatorially large design space has been studied empirically, but given the vast expense of training and evaluating models with different configurations, a principled understanding of the effects of layer selection and ordering is needed. We provide the first such analysis, showing that the dominant hybrid architecture is highly task-specific. Concretely, for any layer configuration, we prove that there exists a task where we can construct a model with this configuration that provably outperforms any other model of any different configuration under fixed resource constraints. Empirically, we observe that for standard training, learned hybrid architectures exhibit separations between layer orderings that match our theoretical findings. For instance, under the same
memory budget, the optimal architecture predicted by our theory for a segmented
retrieval task consistently achieves perfect accuracy, whereas the next strongest plateaus below 0.5. These results extend to deeper models, where “containing” the dominant architecture for a task indicates whether the model attains high performance. In addition, we investigate the mechanisms learned by hybrids, observing that
Transformer layers contribute the largest share of performance on associative recall across architectures. This work establishes insights into task-dependent hybrid architecture design relevant to the training and scaling of frontier-scale hybrids.
Read the full abstract ▾
All authors John Cooper, Mingchen Ma, Jiaqi Mo, Shengqi Qiu, Nathaniel Bastian, Ilias Diakonikolas, Frederic Sala