Algebraic Geometry
A uniform lower bound for the Zhang-Kawazumi invariant and applications to the Bogomolov conjecture
Robert Wilms
We prove that the Zhang-Kawazumi invariant $\varphi(X)$ of a compact and connected Riemann surface $X$ of genus $g\ge 2$ is strictly larger than \[\frac{g(g+2)-(2g+1)H_g}{g-1},\] where $H_g=\sum_{k=1}^g \frac{1}{k}$ denotes the $g$-th harmonic number. If $X$ is hyperelliptic, we give the stronger bound $\varphi(X)>\frac{g}{2}(H_g-1)$. The proof relies on a new expression of $\varphi(X)$ in terms of a quadratic form on the space of smooth Hermitian
matrix-valued functions on $X$, evaluated at certain projector
matrices. As an arithmetic application, we deduce new lower bounds for the self-intersection number $\omega_a^2$ of the admissible adelic metrized canonical bundle $\omega_a$ of a smooth projective curve of genus $g\ge 2$ over a number field and hence new uniform height bounds in the Bogomolov
conjecture.
Read the full abstract ▾
Optimization and Control
From detectability of abstract linear systems to exponential output-to-state stability and back
Qiaoling Chen, Andrii Mironchenko and Fabian Wirth
We study exponential output-to-state stability (eOSS) of linear infinite-dimensional systems in
Banach spaces with bounded output operators. It is shown that eOSS is equivalent to the existence of a coercive eOSS Lyapunov function in implication form. Also exponential detectability guarantees the existence of a coercive eOSS Lyapunov function in dissipation form and therefore eOSS. Counterexamples demonstrate that the converse implications fail in general: exponential zero-detectability does not imply eOSS, eOSS does not imply the existence of an eOSS Lyapunov function in dissipation form, which, in turn, does not imply exponential detectability. If the unstable subspace is finite-dimensional, zero-detectability implies exponential detectability and yields equivalent characterizations of eOSS. The results are illustrated with a
parabolic equation.
Read the full abstract ▾
Computer Vision
GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space
Kerui Ren, Tao Lu, Linning Xu et al. (+5 more)
Novel view synthesis from sparse images must reconcile faithful reconstruction of observed regions with plausible completion of unseen content, while maintaining world consistency across viewpoints. Existing geometry-based methods preserve observed scene structure but often struggle to complete unseen regions, whereas
video generative models offer rich appearance priors but accumulate inconsistencies during sequential view generation. We propose GeoVerse, a framework that synthesizes world-consistent novel views by performing generation within the
geometric latent space of a pretrained 3D foundation model and injecting appearance priors from a
video generative model. Specifically, GeoVerse extracts multilevel features from Wan2.2 VACE and injects them into the
geometric latent
diffusion model via a ControlNet-style adapter, incorporating
video-learned appearance priors to enhance structural completion. To enforce cross-view coherence, a global spatial
memory continuously aggregates observed and synthesized content, reprojecting target-aligned guidance to anchor subsequent predictions to a shared scene representation. Extensive experiments across diverse
datasets demonstrate improved visual quality and
geometric consistency, with a 2.23 dB higher PSNR on DL3DV and 32.4% lower ATE on Mip-NeRF360 compared to GLD.
Read the full abstract ▾
All authors Kerui Ren, Tao Lu, Linning Xu, Changjian Jiang, Mu Huang, Chunhua Shen, Mulin Yu, Bo Dai
High Energy Astrophysical Phenomena
Fast radio burst - persistent radio source systems III. The relation between PRS luminosity and FRB rotation measure
D. Pelliciari, G. Bernardi, L. Bruno et al. (+5 more)
Fast
radio bursts (FRBs) are millisecond-duration
radio transients of extragalactic origin whose physical origin remains uncertain. A small fraction of FRBs are known to repeat, and some of them are associated with persistent
radio sources (PRSs), interpreted as synchrotron-emitting nebulae surrounding the FRB source. In the context of magnetar-based models, the rotation measure (RM) of an FRB is expected to correlate with the spectral luminosity of its associated PRS, providing a probe of the physical properties and
evolution of the nebula. We investigate the relation between FRB RM and PRS spectral luminosity, constrain the characteristic size of the PRS nebulae, and use the intrinsic scatter of the relation to investigate evolutionary scenarios for FRB–PRS systems. We analyse a sample of 50 FRB sources with known RM, including both PRS detections and spectral luminosity upper limits, using a
Bayesian MCMC framework that jointly accounts for detections and non-detections. We model the luminosity–RM relation as $L_\nu \propto \zeta_e \gamma_c^2 R^2 {\rm RM}^\beta$, constraining the characteristic nebular size $R$ and the RM scaling index $\beta$. For the full sample, we obtain $R = 0.016^{+0.115}_{-0.014}$ pc at $1\sigma$ confidence for a free $\beta$, while fixing the RM dependence to the canonical linear scaling, $\beta = 1$, gives $R = 0.008^{+0.058}_{-0.007}$ pc. The data mildly favour super linear RM scalings, although the inferred values of $\beta$ remain consistent with $\beta = 1$ within $2\sigma$. We find a substantial intrinsic scatter in the luminosity–RM relation, significantly larger than previous estimates based on confirmed PRSs alone. Interpreting this scatter within evolutionary models, our results favour scenarios involving efficient particle acceleration and/or rapid nebular expansion.
Read the full abstract ▾
All authors D. Pelliciari, G. Bernardi, L. Bruno, M. Pilia, P. Esposito, L. Beduzzi, A. Geminardi, O. Smirnov
Artificial Intelligence
Failure-Transparent Agents: Benchmarking Post-Failure Reporting in Tool-Using Language Models
Junru Zhu, Shiming Xie, Aime Lu Fan Chen et al. (+4 more)
Tool-using
agents can fail twice: a required tool can fail, and the
agent can then report success without the evidence needed to justify it. Existing
benchmarks often entangle this reporting failure with tool selection, recovery, and environment dynamics. We introduce Failure-Transparent
Agents (FTA), a controlled
benchmark that fixes the failed observation and required evidence state before generation, making post-failure claims directly auditable. FTA contains 100 tasks with deterministic failure traces spanning five failure families, a neutral control, and four user-pressure conditions, and evaluates unsupported claims alongside useful recovery. Across six models, three response policies, and 3,600 human-annotated responses, false-success rates are 22.8% under the baseline policy, 9.3% with a transparency instruction, and 0.8% with a structured evidence contract. Fabricated-detail rates decrease from 28.3% to 14.3% and 0.8%, while useful responses increase from 74.9% to 89.2% and 98.8%, respectively. The tested evidence-contract policy is associated with substantially lower post-failure reporting errors while useful-response rates remain high within this blocked-task
benchmark.
Read the full abstract ▾
All authors Junru Zhu, Shiming Xie, Aime Lu Fan Chen, Xiaoqing Ding, Chunxin Tang, Ruoyu Qi, Yulang Fei
Differential Geometry
Partial Regularity of Stable Stationary Harmonic Maps into Compact Lie Groups
Jacob Krantz
Let $M$ be a compact Riemannian
manifold, and let $G$ be a compact Lie group with bi-invariant metric. We show that the singular set of any stable stationary harmonic map $u : M \to G$ has Hausdorff codimension at least four. This is sharp.
High Energy Astrophysical Phenomena
Fast radio burst - persistent radio source systems II. A faint PRS associated with the nearby FRB 20181030A?
D. Pelliciari, G. Bernardi, B. Margalit et al. (+10 more)
Persistent
radio sources (PRSs) are the continuum counterparts of fast
radio bursts (FRBs), the latter being extragalactic transients of millisecond duration and Jy-level flux density. An FRB-PRS system is thought to be a flaring magnetar surrounded by an highly magnetized,
baryon-loaded nebula. We aim to constrain the size of 20181030A-S1, a new PRS candidate, potentially associated with the repeating FRB 20181030A. The latter is localized with $\sim 1'$ uncertainties in the outskirts of NGC 3252, which is a spiral
galaxy at a luminosity distance of $20$ Mpc. We report very long baseline interferometric (VLBI) observations using the European VLBI Network at $1.7$ GHz of this PRS candidate at an angular resolution of $20$ milliarcseconds. Our observations reveal the presence of an unresolved
radio source (20181030A-S1) at the position of the PRS candidate, confirming its compactness at milli-arcsecond angular scales. A fit to the position of the point-source yields a peak flux density of $280 \pm 30$ $\mu$Jy and a transverse physical size constrained to be $R < 0.5$ pc at $68\%$ confidence level (CL). This flux density converts to a spectral luminosity of $(9 \pm 1) \times 10^{25}$ erg s$^{-1}$ Hz$^{-1}$, $\sim 3$ orders of magnitude lower than confirmed PRSs, making 20181030A-S1 the closest and faintest PRS candidate known. Its low luminosity and modest rotation measure are consistent with the $L_\nu-$rotation measure (RM) relation followed by confirmed FRB–PRS systems, supporting a common physical origin in magnetar-powered nebulae. We show how a magnetized wind nebula powered by an initially weak ($B_\star \simeq 10^{15}$ G) and young ($t_{\rm age} \simeq 15 - 150$ yrs) magnetar can account for both the observed spectral luminosity and RM of the system. Other possible origin scenarios for 20181030A-S1, in the case in which it is unrelated to the FRB source, are also discussed.
Read the full abstract ▾
All authors D. Pelliciari, G. Bernardi, B. Margalit, B. D. Metzger, C. Nanci, L. Bruno, M. Pilia, L. Beduzzi, C. Spingola, C. Stanghellini, P. Esposito, A. Geminardi, M. Giroletti
High Energy Astrophysical Phenomena
Fast radio burst - persistent radio source systems I. A 1.2 GHz search and catalog definition
D. Pelliciari, L. Bruno, G. Bernardi et al. (+5 more)
We present new $1.26$ GHz observations of 24 fast
radio bursts (FRBs; 9 repeaters and 11 one-offs), conducted with the upgraded
Giant Meter Wave
Radio Telescope (uGMRT) at arc-second angular scale resolution, and combined them with literature data to construct an extended sample of 75 sources with either persistent luminosity measurements or upper limits. Our catalog is complete down to a $10^{29}$ erg s$^{-1}$ Hz$^{-1}$ luminosity threshold, which corresponds to the value of the first-discovered PRS, associated with FRB 20121102. We detect compact
radio emission in six FRB fields, two of which are previously known PRSs and four are new candidates. The four new candidates are associated with three new repeating FRBs and one apparently one-off source. We computed the PRS occurrence using the whole catalog of 75 FRBs and found that PRSs more luminous than $10^{29}$ erg s$^{-1}$ Hz$^{-1}$ are rare. Unless otherwise stated, all uncertainties and upper limits are reported at the $95\%$ confidence level (CL). If we only consider confirmed PRSs, we find an occurrence of $f_{\rm all} = 0.06^{+0.09}_{-0.03}$, while including PRS candidates as well yields $f_{\rm all} = 0.14^{+0.22}_{-0.08}$. We also compute the PRS occurrence in the two classes of FRBs: repeating and non-repeating. In the conservative case, i.e., when none of the new PRS candidates are considered to be associated with their FRBs, we find no preferred association between the PRSs and FRB class, with an occurrence of $f_{\rm r} = 0.14^{+0.20}_{-0.09}$ for the sample of repeating FRBs and $f_{\rm nr} \leq 0.12$ for the sample of one-off sources. In the inclusive case, i.e., when all the new candidates are considered confirmed PRSs, we find $f_{\rm r} = 0.27^{+0.21}_{-0.14}$ and $f_{\rm nr} = 0.03^{+0.14}_{-0.02}$, with marginal ($90\%$ CL) evidence that PRSs are preferably associated with repeating FRBs.
Read the full abstract ▾
All authors D. Pelliciari, L. Bruno, G. Bernardi, M. Pilia, P. Esposito, L. Beduzzi, A. Geminardi, O. Smirnov
Computer Vision
FlowAct-R2: Beyond Talking Avatar via Streaming Multimodal References and Proactive Agent Planning
Ziyao Huang, Zhengkun Rong, Shiyang Qin et al. (+5 more)
We present FlowAct-R2, a framework for interactive
humanoid video generation that combines continuous
multimodal control with proactive
agent planning. Our method consists of two coupled components. First, a Streaming
Multimodal Reference
Diffusion Transformer adapts the pretrained Seedance 2.0 Mini reference-to-
video backbone to accept rolling action
prompts, streaming audio, and dynamically updated image, audio, and
video references.
Video-driven rotary positional
embeddings align reference chunks with the generation timeline, while reference-plus-image conditioning and partially noised historical motion frames preserve appearance and avoid accumulated drift. Second, a Proactive Interaction
Agent separates pre-online planning from online scheduling and response: it prepares a persona, a long-horizon agenda, and reusable
multimodal skills in advance, then autonomously schedules behaviors, responds to audience input, and handles interruptions during a live session. FlowAct-R2 supports real-time 720p generation and hour-scale streaming across entertainment streaming, live shopping,
video chatting, and live vlogging.
Read the full abstract ▾
All authors Ziyao Huang, Zhengkun Rong, Shiyang Qin, Shuang Liang, Wentao Hu, Yuxuan Luo, Yuan Zhang, Mingyuan Gao
Logic
Unifying Conservation as Translation for General Calculi
Giulio Fellin
We present an abstract framework for conservation and translation
theorems between logical calculi. Unlike previous approaches, our setting does not require the underlying consequence relations to satisfy structural properties such as cut, allowing in particular for cut-free calculi. Moreover, we study translations between arbitrary calculi rather than only from a stable extension to the original calculus. Translations are formulated at the level of sequents by pairs of functions acting on antecedents and succedents, and the framework is developed for multi-succedent calculi while subsuming the single-succedent case. This yields a uniform treatment of classical results including negative translations, minimality
theorems, Orevkov's conservation classes, and a characterisation of the least logic satisfying Kuroda's double negation
theorem.
Read the full abstract ▾
Computer Vision
Impact of Patient Orientation in Single- and Multi-View Camera Environments for AI-based Rehabilitation Monitoring
Miriama Jánošová, Andreas Lang, Petra Budikova et al. (+1 more)
Automated quality assessment of rehabilitation exercises relies heavily on accurate human pose estimation from
video data. Although numerous RGB-based pose estimation methods have been proposed, the impact of camera placement on detecting clinically relevant movement errors remains insufficiently explored. To address this gap, we introduce REHAB26-ViewAngles, a
dataset comprising correct and incorrect rehabilitation exercise executions captured from a wide range of camera angles. Furthermore, we propose a novel separability metric to quantify an
algorithm's ability to distinguish between valid and faulty exercise repetitions. Using these tools, we analyze how various RGB-based pose-estimation strategies are suitable for exercise quality assessment under varying camera placements. In particular, we analyze single-camera 2D and 3D pose estimation and four multi-camera strategies: a combination of two orthogonal 2D views, 3D triangulation, weighted 3D fusion, and an AI-based pose-estimation
transformer model specifically trained from two synchronized cameras. Our findings reveal that an optimally placed 2D camera can improve the separability by 16.9% over the commonly used 0{\deg} frontal view and frequently outperforms single-camera 3D estimation, while combining two views can further improve accuracy by up to 13.1%. These results offer practical guidance for deploying rehabilitation monitoring in both home and clinical settings.
Read the full abstract ▾
All authors Miriama Jánošová, Andreas Lang, Petra Budikova, Jan Sedmidubsky
Computer Vision
Superquadric Primitive Decomposition of 3D point clouds via Geometric-Aware Inlier Refinement
Alessandro Rinaldi, Edoardo Tedesco, Andrea Ferraris et al. (+5 more)
The decomposition of 3D point clouds into interpretable
geometric primitives remains a longstanding challenge in Computer
Vision and Computer Graphics. Among the available representations, superquadrics offer a compact and expressive model capable of capturing a wide range of shapes. However, their estimation is inherently challenging, as it requires solving a non-linear optimization problem and is particularly sensitive to noise, outliers, and overlapping structures. While robust estimation methods such as RANSAC and its variants achieve strong performance, they rely primarily on spatial proximity and residual-based criteria, often leading to incorrect inlier assignments across adjacent or complex arrangements of primitives. In this work, we introduce a
geometric-aware framework for primitive decomposition that explicitly incorporates local surface properties into the fitting process. Specifically, we propose an inlier refinement step formulated as an energy minimization problem and solved via graph-cut optimization. Our formulation integrates
geometric priors, such as normal consistency, enabling more reliable inlier selection beyond purely residual-based criteria. The approach naturally applies to both single-model estimation and multi-model decomposition. By leveraging
geometric information beyond point-wise residuals, our method reduces erroneous inlier propagation and stabilizes parameter estimation. Experiments on synthetic and real
datasets show consistent improvements in
geometric accuracy, robustness to noise and outliers, and convergence efficiency compared to state-of-the-art RANSAC-based methods.
Read the full abstract ▾
All authors Alessandro Rinaldi, Edoardo Tedesco, Andrea Ferraris, Filippo Leveni, Daniele Baieri, Filippo Maggioli, Simone Melzi, Luca Magri
Algebraic Geometry
Abelian canonical covers of minimal rational surfaces and $\mathbb F_1$ up to degree $8$
Alexandre Dorothée and Francisco Javier Gallego
We classify abelian canonical covers of degree at most $8$ of smooth surfaces of minimal degree, and of $\mathbb P^2$ and Hirzebruch surfaces not embedded as surfaces of minimal degree. Let $\pi: X \longrightarrow W$ be an abelian canonical cover of a smooth surface of minimal degree $W$ with group $G$ and degree $n$. Degrees $n \le 4$ were classified earlier; we complete the picture for $5 \le n \le 8$: - No such covers exist if $n$ is
prime and $n > 3$, nor if $G=\mathbb Z_8$. - For $G=\mathbb Z_6$, $\mathbb Z_2 \times \mathbb Z_2 \times \mathbb Z_2$ and $\mathbb Z_2 \times \mathbb Z_4$ we give the complete list of covers. Only $W =\mathbb P^2$, $\mathbb F_0$ and $\mathbb F_1$ arise. The list contains several families with unbounded $p_g$, most of them new; $X$ is smooth for general covers with $G=\mathbb Z_2 \times \mathbb Z_2 \times \mathbb Z_2$, but singular if $G=\mathbb Z_6$ and, with one exception, if $G=\mathbb Z_2 \times \mathbb Z_4$. When $W$ is $\mathbb P^2$ or a Hirzebruch surface not embedded as a surface of minimal degree, we show that abelian canonical covers exist only for $n = 2$ or for $G =\mathbb Z_2 \times \mathbb Z_2$, and we classify the latter completely: they form infinitely many families, most new, with unbounded $p_g$ and, most of them, with unbounded irregularity. The slopes $K^2/\chi$ of all these $X$ accumulate at $6$, at 8, and at $8\alpha/(\alpha +1)$ for every integer $\alpha \ge 2$. These results rest on more general ones, valid for $X$ normal, locally Gorenstein and of arbitrary dimension, and $W$ smooth with $p_g(W) = 0$: we determine which branch divisors of an abelian canonical cover must vanish, and, for canonical covers whose $\pi_*\mathcal O_X$ splits as a sum of line bundles, we describe the $\mathcal O_W$-module structure of $\pi_*\mathcal O_X$, computing it completely when $W$ is $\mathbb P^2$ or a Hirzebruch surface.
Read the full abstract ▾
High Energy Astrophysical Phenomena
Empirical determination of the Galactic neutron star–black hole merger rate using StarTrack models
Terrence Pierre Jacques, D. R. Lorimer, Nihan Pol et al. (+2 more)
The discovery of a
pulsar in a binary system with a black hole would provide a unique laboratory for testing general relativity in the strong-field regime and offer vital constraints on massive
star evolution. We assess prospects for the detectability of such systems by utilizing the
radio pulsar population and survey models within the modeling package PsrPopPy and, as a proof of principle, two evolutionary models with different treatments of common-envelope phases from the
stellar population synthesis code StarTrack. For these two models, assuming the
radio pulsar beaming model proposed by Tauris and Manchester and accounting for the effects of orbital motion on
pulsar detectability, we calculate upper limits on the number of
neutron star–black hole systems in the
Galaxy and the corresponding merger rate given the lack of any
radio-detected
pulsar–black hole binary systems in current
radio surveys. For the most constraining model, we find a
Galactic merger rate $< 4.3~\text{Myr}^{-1}$ (at 95% confidence). These results are extrapolated to the expected horizon distance for the fifth LIGO observing run, leading to LIGO detection rate estimates that are consistent with the latest values reported by the LIGO-Virgo-KAGRA collaboration. Our results indicate that the increased sensitivity of FAST, MeerKAT, and DSA offers the best prospects yet for uncovering this elusive population.
Read the full abstract ▾
All authors Terrence Pierre Jacques, D. R. Lorimer, Nihan Pol, M. A. McLaughlin, K. Belczynski
Lattice Field Theory
Multilevel Plaquette-Space Sampling for Lattice Gauge Theories with Local Constraint Solves
Ankur Singha, Jacob Finkenrath, Karl Jansen et al. (+2 more)
Lattice gauge theories are an important class of physical models with high-dimensional structured distributions that underpin first-principles calculations in particle,
nuclear, and condensed-matter physics. Recent advances in generative modeling have opened new avenues for sampling Boltzmann distributions in
lattice gauge theories, offering the potential to alleviate limitations of traditional Monte Carlo methods, including critical slowing down and topological freezing. However, generative samplers for
lattice gauge theories are typically constructed in link space, where correlations become increasingly long-ranged toward weak coupling, posing a challenge for learning. Plaquettes provide a more natural representation, as the action is local in these variables. However, exact Bianchi constraints restrict them to a lower-dimensional
manifold, complicating direct generative modeling. We introduce a multilevel normalizing-flow construction that samples directly in plaquette space while satisfying these constraints exactly. The key idea is a coarse-to fine factorization that transforms a globally coupled constraint problem into a sequence of local solves: at each refinement, every determined plaquette depends on at most four newly generated variables, while the size of the constraint problem remains independent of the
lattice size. We validate the construction for $U(1)$ in two and four dimensions and for $SU(2)$ in two dimensions. Our multilevel Plaquette-Space Sampler (PSS) substantially outperforms link-space baselines, with the advantage increasing toward weak coupling, where the gauge coupling becomes small. This regime is particularly challenging for generative sampling and, in asymptotically free gauge theories, is relevant to continuum studies.
Read the full abstract ▾
All authors Ankur Singha, Jacob Finkenrath, Karl Jansen, Vipul Arora, Shinichi Nakajima
Earth and Planetary Astrophysics
Correcting stellar contamination in transmission spectroscopy with contemporaneous monitoring: Application to GJ 1214 b
Ò. Porqueras-León, M. Perger, J. C. Morales et al. (+19 more)
Transmission spectroscopy is an essential tool for characterising the atmospheres of transiting
exoplanets. However,
stellar surface inhomogeneities contaminate transmission spectra, posing a major challenge to high-precision observations such as those from the James Webb Space
Telescope (JWST). We develop and apply methods to correct for
stellar contamination, improving the accuracy of
exoplanet atmospheric characterisation. We obtained coordinated multi-band photometry and high-resolution spectroscopy around JWST observations of GJ 1214 b and applied two frameworks: (i) a StarSim inversion, which infers
stellar activity parameters from the monitoring data and propagates them to corrections; and (ii) UnSPOTTER, in which neural networks trained on StarSim simulations learn the mapping from monitoring data to corrections. Both frameworks agree within uncertainties and indicate that the JWST observations of GJ 1214 occurred at a favourable rotation phase with low
stellar contamination. For UnSPOTTER, the inferred corrections can be constrained at approximately 10-30 ppm across the JWST wavelength range, implying post-correction residuals of the same order. If the same system were observed near activity maximum, the predicted contamination would be strongly chromatic and affect atmospheric
retrievals, highlighting the importance of coordinated monitoring and robust activity correction techniques. The two approaches are complementary but serve different roles. The inversion yields physically interpretable parameters and surface maps, whereas UnSPOTTER marginalises over surface-configuration degeneracies and provides tighter predictive corrections within its validated simulation domain. Our pipeline can potentially be generalised to other active
stars: with multi-band and multi-technique monitoring around JWST visits, it can provide activity forecasts for scheduling and activity corrections.
Read the full abstract ▾
All authors Ò. Porqueras-León, M. Perger, J. C. Morales, I. Ribas, J. Blanco-Pozo, H. Tabernero, E. Herrero, E. Pallé, H. Parviainen, F. Murgas, C. Rodríguez-López, P. H. Hauschildt, A. Schweitzer, N.-E. Nèmec, D. Baroch, N. R. C. Gomes, G. Anglada-Escudé, J. Aceituno, N. Narita, A. Fukui, M. Mori, Y. Smolyarova
High Energy Astrophysical Phenomena
SN 2025qe: A case study of diversity in low-luminosity thermonuclear explosions
Hrishav Das, Anirban Dutta, Devendra K. Sahu et al. (+3 more)
We present optical photometric and spectroscopic observations of SN 2025qe, a Type Iax
supernova in IC 0529, monitored from $-10$ to $+150$ rest-frame days relative to $g'$-band maximum using the GROWTH-India and the Himalayan Chandra
Telescopes. SN 2025qe reaches a peak $g'$-band absolute magnitude of $M_{g'} \approx -16.16 \pm 0.15$ mag, placing it in the sparsely sampled luminosity gap between the brightest and faintest members of the SNe Iax class. The multiband light curves show pronounced wavelength-dependent broadening at late times, with the redder bands remaining brighter than in other SNe Iax. Semi-analytical (Arnett) modeling of the pseudobolometric light curve yields a synthesized $^{56}$Ni mass of $0.02\,M_\odot$, an ejecta mass of $0.45\,M_\odot$, and a kinetic energy of $7.0\times 10^{49}$ erg. Three-dimensional pure-deflagration models of Chandrasekhar-mass carbon-oxygen white
dwarfs reproduce the peak luminosity and $^{56}$Ni yield but underpredict the slow, red post-maximum decline; models including energy input from a bound remnant better match the late-time
evolution, especially in the red bands. We model early-time spectra using TARDIS with a modified deflagration model and identify late-time spectral features using syn++. We qualitatively try to explain the persistent red-band flux with a two-component ejecta structure comprising a dense inner region, possibly including a bound remnant, and a density-enhanced outer shell that sustains redder emission. Together, the photometric and spectroscopic properties of SN 2025qe highlight the diversity of ejecta structure and
evolution among SNe Iax with similar peak luminosities, emphasizing the need for detailed observations and modeling of their ejecta.
Read the full abstract ▾
All authors Hrishav Das, Anirban Dutta, Devendra K. Sahu, G.C. Anupama, Prathamesh S. Kadam, Vishwajeet Swain
Computers and Society
AI Agent Swarms as Researchers: Progress, Challenges, and Open Questions
Sergey Gusev and David E. Bernal Neira
Artificial intelligence (AI)
agents,
language models connected to tools and run in a loop, can now carry out long, multi-step tasks with little supervision. We gave swarms of off-the-shelf coding
agents a short statement of scope, from a narrow topic to a whole field, access to the literature and to computing tools, and one standing instruction: make real, correct, useful progress, and do not stop. We supplied no scientific ideas. Within weeks, the
agents produced a large body of research notes, paper-length drafts, and formal proofs in five areas of optimization theory and physical science, and proposed untested laboratory experiments in a sixth. We do not claim that all of it is correct or new, but it is not noise: in what we have checked so far, we found no major scientific error, and several results are proved in a proof assistant. The
agents produced results faster than we could review them; we estimate that a full review would take us months. Together with two widely discussed 2026 results in mathematics obtained with swarms, our runs suggest that
agents can already do a large part of routine theoretical research, at least in areas that we experimented with. This raises questions we cannot yet answer: how to trust results when review, not production, is the scarce resource; what credit and publication counts mean when the human input is a
prompt, and why institutions would pay researchers rather than buy computing time; and how people can learn a field, add to what
agents do, and stay in control of research they cannot keep up with. Research institutions are not ready: models improve faster than institutions change, so they should decide now how to respond as capabilities increase. We offer tentative positions, release the
agents' unedited output as of 25 September 2026, and invite readers to repeat the experiment in their own fields.
Read the full abstract ▾
Computer Vision
Hard Vision, Easy Vision: What GPT-6 Astra Reveals Across Computer Vision
Hanoona Rasheed, Mohammed Irfan Kurpath, Bin Ren et al. (+3 more)
Frontier general-purpose systems are rapidly expanding beyond visual understanding into capabilities traditionally handled by dedicated computer-
vision models. As these capabilities expand, a central question for the computer-
vision community is how far this reach extends, and what remains hard. We evaluate GPT-6 Astra alongside five frontier general-purpose AI systems across 34 capabilities and 55
benchmarks spanning nine areas of computer
vision. We compare their performance with dedicated models and humans where suitable references are available. Astra demonstrates broad visual capability, with substantial gains over other frontier systems in visual and spatial
reasoning and several forms of structured prediction. Across the state-of-the-art systems, a consistent pattern emerges. Capabilities involving semantic interpretation,
reasoning, and object-centric prediction increasingly approach or reach available reference levels. In contrast, larger gaps remain when tasks require metric
geometric accuracy, faithful reconstruction, temporally consistent dense prediction, or specialized fine-grained visual knowledge. Additional
reasoning and specialist tools close selected gaps, but their benefits vary across capabilities. These results map a changing landscape of computer
vision in which increasingly sophisticated visual tasks are accessible through a general-purpose interface, while precise and fidelity-sensitive perception remains an important frontier.
Read the full abstract ▾
All authors Hanoona Rasheed, Mohammed Irfan Kurpath, Bin Ren, Hisham Cholakkal, Fahad Shahbaz Khan, Salman Khan
Robotics
RoboCompiler: Graph-Native Compilation of Closed-Chain Robots for Consistent Modeling, Control, and Simulation
Mehdi Heydari Shahna, Joongheon Kim and Jouni Mattila
Robots with kinematic loops, coupled actuators, and changing contacts require consistent models of configuration, motion, force, and dynamics. Yet these interfaces are often reconstructed separately for control and simulation, making closure and actuation consistency difficult to maintain. This paper presents RoboCompiler, a graph-native framework that compiles a canonical mechanism graph into a shared mechanical interface. From bodies, joints, frames, inertias, and actuator ports, it constructs closure paths and analytic residual Jacobians, then assembles feasible configurations through rank-checked continuation and correction. A tangent lift maps independent velocities to full
robot and task motion, while paired actuator-port maps preserve virtual work. A constraint-
curvature correction extends the reduction to accelerations and projected rigid-body dynamics, including floating-base and support modes. Cycle-local evaluation, generated Jacobians, and dependency-aware reuse enable localized updates when closure inputs change. We evaluate physical loops and task-induced constraints on a industrial excavator, Unitree Go2, Franka Panda, Kangaroo, and a six-UPS Stewart platform. High-precision constrained-dynamics and independent Pinocchio checks confirm mechanical consistency; MuJoCo and Isaac Sim/PhysX executions demonstrate task performance and model reuse under native contact. For Kangaroo, compilation reduces residual-and-Jacobian evaluation time by 96.7% and closed-loop rollout wall time by 66.8%, with dynamics and control held fixed.
Read the full abstract ▾
Numerical Analysis
When does least-squares residual minimization solve a differential equation?
Carlos Esteve-Yagüe, Richard Tsai and Stanley Osher
Least-squares residual minimization approximates a differential equation through an optimization problem. This paper investigates the conditions required to transition from stationary points and minimizing sequences of the residual functional to the exact solution. Starting from the classical orthogonality relation between the residual and the range of the linearized operator, we study how boundary conditions and parameterized variations determine what can be inferred about the residual at a stationary point. Smooth spurious critical points can occur even for well-posed problems; for linear equations, a compatibility condition yields an exact projection characterization. For a scalar conservation law with transonic boundary data, every minimizing sequence converges to a continuous function that does not solve the equation, even when a stationary
entropy shock exists. We give conditions under which approaching this limit forces the parameter norms to diverge. A Hamilton–Jacobi example shows that a function can have zero residual without being the viscosity solution. Favorable results for a viscous model and uniformly convex Monge–Ampère critical points identify conditions that restore the corresponding implications. For piecewise-smooth trial families, the regularity needed to define the residual functional does not by itself justify the sampled gradient used in training. Moving residual jumps can contribute terms absent from that gradient; an explicit example has identically vanishing sampled gradients and a nonzero continuous derivative. Together, these results clarify what stationarity and sampled gradients imply about the residual, and under which conditions minimizing sequences lead to the intended solution.
Read the full abstract ▾
Machine Learning
X-Reset: Scaling Object-Centric Reinforcement Learning via Cross-Embodiment Resets
Prithwish Dan, Chenyang Ma and Wei Zhan
Reinforcement learning (RL) in simulation can train dexterous manipulation policies without
robot demonstrations, but training a single generalist policy with task-agnostic rewards faces a severe exploration problem: approaching, grasping, and reorienting diverse objects with many degrees of freedom is difficult to discover from scratch. Prior works make exploration tractable with high-quality
robot demonstrations, per-task reward shaping, or by restricting policies to narrow modes of behavior. We propose X-Reset, a framework that instead resolves exploration with human hand-object demonstrations. Rather than imitating or tracking retargeted human motion, X-Reset kinematically retargets hand-object states to noisy
robot states, filters out states that are unstable in simulation, and samples the remainder as resets during RL training with general-purpose object-centric rewards. The resulting policy depends only on object state and goal, with demonstrations entering training through the reset distribution. We show that X-Reset trains generalist policies on 20 objects across three embodiments—a 22-DoF hand on two different arms and a parallel-jaw gripper—and resolves the exploration challenges of RL from scratch. X-Reset scales with the number of training objects, generalizes to unseen objects, can learn from imperfect hand-pose estimates, and transfers behaviors zero-shot from sim-to-real.
Read the full abstract ▾
Statistics Theory
Competitive optimality in testing by betting via Bell-Cover randomization
Aaditya Ramdas
Bell and Cover showed that an investor who multiplies the initial unit of capital by an independent uniform random variable on $(0,2)$, and then uses the log-optimal portfolio, wins a head-to-head wealth comparison with probability at least one half against every independently randomized competitor. We explain very simply how this result transfers to testing by betting: for any composite null $\mathcal P$ and simple alternative $Q$, denoting $E^*$ as the corresponding numeraire e-variable, we show that $UE^*$ exceeds any other e-variable $E$ with probability at least half. Interestingly, we show that this competitive optimality result is actually equivalent to the numeraire inequality $\mathbb E_Q[E/E^*]\leq1$, and in general randomization only helps the numeraire and fails to improve the competitive advantage of an arbitrary e-variable. Under optional stopping with or without knowledge of $U$, we emphasize a key distinction between e-process validity and competitive optimality. We also show that competitive optimality comes at the price of expected log wealth and power: thresholding $UE^*$ at $1/\alpha$ has sharp size at most $\alpha/2$, but the factor of two actually disappears under optional stopping. Even after correcting for this factor of two, the test is dominated in conditional rejection probability by randomizing the testing threshold (randomized Markov's inequality). Thus, Bell-Cover randomization is optimal for a specific competitive objective, at the cost of others.
Read the full abstract ▾
Cosmology and Nongalactic Astrophysics
Tide-and-Seek: Anisotropic scale-dependent bias in the Lyman-alpha forest and designer samples
Yong Sheng Yap, William R. Coulton and Vid Iršič
Searches for
primordial non-Gaussianity are powerful ways of learning new information about the early Universe. Observing
primordial signals in the late Universe is challenging as, in general, it is difficult to disentangle the small,
primordial signatures from the complex physics that describes the distribution of
galaxies and large-scale
cosmic structures. We demonstrate that scale-dependent bias from
primordial non-Gaussianity with non-trivial directional dependence in the squeezed limit (also known as anisotropic
primordial non-Gaussianity) can arise in a much larger class of tracers than previously considered. This is a smoking gun signature of
primordial Universe physics that cannot be mimicked by late-Universe processes. We demonstrate this explicitly with simulations of the Lyman-$\alpha$ forest and
halos selected through their projected shape. The former is a standard cosmological probe and standard analyses could be simply extended to constrain this. The latter provides an interesting new approach: design samples with strong selection effects to provide new sensitivity to
primordial physics. As a demonstration of these concepts, we apply this method to BOSS DR12 data, finding no evidence of anisotropic non-Gaussianity.
Read the full abstract ▾
Materials Science
Catalogue of Collinear Magnetic Structures in Stoichiometric Materials
Liangliang Huang, Houhao Wang, Yuanze Song et al. (+3 more)
Magnetic materials persistently serve as a fertile platform for exploring novel
quantum emergent phenom ena and advancing information technologies. Reliably determining their
magnetic structures (MSs), which is essential for understanding their emergent
quantum properties, usually depends on experiments such as
neutron scattering. Here, noting that around half of the MSs in MAGNDATA (a database of experimentally determined MSs) are collinear, we adopt the
spin space groups (SSGs) to preset the candidate MSs for a given nonmagnetic
crystal structure to be collinear from the outset. On the other hand, more than 80% of the collinear magnets in MAGNDATA are subject to maximal subgroups of nonmagnetic groups which are collinear SSGs, and we then enumerate all collinear MSs formed at Wyckoff positions by the maximal collinear SSG subgroups of the 230 nonmagnetic space groups. With such exhaustive enumeration, we categorize all
symmetry-allowed collinear MSs for the 72,075 nonmagnetic stoichiometric structures in Inorganic
Crystal Structure Database, obtaining 365,639 collinear MSs. We then determine the ground-state SSG
symmetries of 7,824 collinear magnets among the 21,315 selected materials using first-principles calculations. These magnets then undergo a comprehensive topological classification combining SSGs and
magnetic space groups, outputting thousands of new
magnetic topological materials. The experimentally determined MSs in MAGNDATA represent the tip of the iceberg in our computational database of collinear magnets, expected to guide MS characterization by experimentalists and further investigation towards spintronics applications.
Read the full abstract ▾
All authors Liangliang Huang, Houhao Wang, Yuanze Song, Ruixi Pu, Xiangang Wan, Feng Tang
Game Theory
On the Power of Determinism in Multi-Item Auctions
Yiannis Giannakopoulos and Johannes Hahn
We study the classical multi-item monopoly setting with a single additive buyer and $m$ heterogeneous items whose values are independent but not necessarily identically distributed. We compare the revenue (REV) of the optimal truthful auction, which in general may be randomized or complicated, with the revenues of three simple deterministic auctions: selling all items separately (SREV), selling them as a single grand bundle (BREV), and choosing the better of the two. Our technical cornerstone is a general upper bound on the optimal multidimensional revenue, parameterized by freely chosen item-wise caps, and expressed in terms of optimal single-item, Myersonian revenues and basic statistics of the value distributions. We derive this inequality through Lagrangian duality for a nonlinear programming formulation of the worst-case approximation ratio of separate-selling. Combining this inequality with carefully chosen caps and a probabilistic analysis of the resulting terms, we obtain improved approximation guarantees for all three deterministic auctions, both for general independent values and for iid values. Notably, we prove that $REV\le 2.5\max\{SREV,BREV\}$, improving the long-standing $5.2$ factor of Ma and Simchi-Levi and the $6$ factor of Babaioff, Immorlica, Lucier, and Weinberg. We complement this upper bound with a construction establishing a lower bound of $2.19$, improving the bound of $2$ by Rubinstein. For two iid items, we determine the exact approximation ratio of selling separately. Our proof constructs tight Lagrangian dual certificates for the aforementioned programming formulation, when item values lie in the grid $\{0,1/K,2/K,\dots,1\}$, for any discretization parameter $K$. Taking $K\to\infty$ yields the tight ratio $1+W(1/e)\approx1.2785$ in the continuous-valued setting, where $W$ is the Lambert-W function, closing the $[1.278,1.368]$ gap left by Hart and Nisan.
Read the full abstract ▾
Computer Vision
Lagrangian–Hamiltonian Flows for Video Prediction and Image Generation: A Symplectic Perspective
Jiawei Hu
We introduce LHFM, a
geometric framework for learning image dynamics. Drawing on structures central to classical mechanics,
symplectic geometry, and
geometric quantization, LHFM represents each image as an exact Lagrangian graph and models its
evolution through image-dependent
Hamiltonian flows, which yield a transport–source parameterization of image velocities. Our primary application is deterministic
video prediction: LHFM-V is a recurrent model that advances frames by integrating predicted transport and source fields, and achieves the lowest reported FLOP count among the compared recurrent models with similar prediction accuracy. The image variant, LHFM-I, shows that the same construction is compatible with flow matching: in a matched experiment, it attains a lower FID than the flow-matching baseline.
Read the full abstract ▾
Robotics
Humanoid Loco-Manipulation With Discrete VLA Model
Wenxin Shao, Siqi Chai, Kun Li et al. (+4 more)
Vision-
language-action (VLA) models using discrete action
tokens have proven effective for controling robotic arms on manipulation tasks. For a
humanoid, however, the whole-body action space – legs, torso, arms, and hands – is far higher-dimensional and heterogeneous, raising tokenization, training, and real-time inference challenges that the previous VLA models do not address. We present Holo-M, to our knowledge the first discrete VLA model for
humanoid loco-manipulation that intrinsically exploits the
language model by extending its vocabulary with action
tokens. In this model, we devise a unified action tokenizer that decomposes the
humanoid action space into four body-part-specific tokenizers – end-effector, body, hand, and kinematics – enabling training across drastically different embodiments and data sources, including
humanoid teleoperation, ego-centric human
video, and simulation. By extending the
language model's vocabulary with these action
tokens, we avoid the knowledge-insulation problem inherent to the models that use separate continuous action experts. To meet real-time control requirements, we decode each body part's action
tokens through grouped discrete
diffusion decoding, rather than using autoregression on the action
tokens. We have conducted extensive experiments on the SIMPLE
humanoid loco-manipulation
benchmark, in which Holo-M achieves the highest success rates in both the generalist and specialist evaluations, leading the second best by significant margins. We will release all the code and model weights.
Read the full abstract ▾
All authors Wenxin Shao, Siqi Chai, Kun Li, Kerou Zhang, Xinzhou Jiang, Wei Xu, Qiang Liu
Computer Vision
Mind the RefGAP: Correcting Reference Attention in Diffusion-Based Visual Editing
Yanan Wang, Shengcai Liao, Guangyi Liu et al. (+1 more)
Reference-guided
diffusion editors struggle to faithfully reproduce user-provided references. We identify a potential bottleneck in
diffusion editors: many methods provide limited reference-
attention allocation. For example, in LoomVideo, edit-region queries assign less than 1% of their
attention mass to the reference. We introduce RefGAP, a training-free correction that determines logit-offset magnitudes online at each layer from the reference-
attention mass measured during the forward pass. Positive offsets to reference logits strengthen reference usage by edit-region queries, while negative offsets for keep-region queries limit reference-induced changes outside the edit. Two global coefficients control the correction; they are selected once on validation data from four development
diffusion editors and held fixed. Across seven
diffusion-based image/
video editors, RefGAP improves identity fidelity in head swapping and face swapping. RefGAP achieves a fidelity-preservation trade-off comparable to separately tuned constant edit-side biases, without per-approach strength sweeps. Additional experiments on virtual try-on and background replacement evaluate transfer beyond identity editing.
Read the full abstract ▾
All authors Yanan Wang, Shengcai Liao, Guangyi Liu, Xiaodan Liang
Machine Learning
ScAn-Bench: Evaluating Scaling Analysis Methodology
Artin Sermaxhaj, Nastaran Alipour, Donat Sinani et al. (+5 more)
Recent progress in machine learning is driven by large-scale foundation models, where scaling laws and finding optimal scaling prescriptions for architecture, data, and hyperparameters are key in advancing the state-of-the-art. Therefore, it is surprising that no systematic study evaluates the methodology to obtain scaling laws and prescriptions across different model types. To shed light on this crucial blind spot and facilitate future research, we introduce the surrogate
benchmarks ScAn-Bench-LLM and ScAn-Bench-VLM based on 4524 and 8024 checkpoints of
language and
vision-
language model pipelines. On our
benchmarks, we perform the first systematic evaluation of both data acquisition and extrapolation methodology for scaling analysis across different data modalities.
Read the full abstract ▾
All authors Artin Sermaxhaj, Nastaran Alipour, Donat Sinani, Johannes Hog, Neeratyoy Mallik, Steven Adriaensen, Jenia Jitsev, Danny Stoll