Quantum Physics
Robustness of Quantum Signal Processing and Resource-aware Polynomial Design
Thiago O. Maciel, Giancarlo Camilo, Allan Tosta et al. (+3 more)
Quantum signal processing (QSP) realizes operator functions through
polynomial transformations of a finite degree. In addition to function approximation errors, realistic implementations incur algorithmic errors also when encoding the input operator and in synthesizing gates into a discrete fault-tolerant gate set. We provide an end-to-end analysis of these errors and show that the input error propagation is governed by the operator-Lipschitz constant $L_\star$ of the implemented function. We then derive an efficient classical
algorithm to compute tight lower and upper bounds on $L_\star$, and we show that these bounds induce in turn lower and upper bounds on the total
quantum-resource cost (for instance, $T$-gate count) needed for QSP implementations compatible with a target total error. We further devise a procedure to search for a
polynomial with minimum cost based solely on its coefficients, yielding a resource-aware
polynomial-design procedure for QSP. We illustrate the resulting method for ground state energy estimation via
eigenvalue thresholding for the H$_6$ molecular
Hamiltonian. In the tested hardware profiles, the resource-aware family improves the total-error budget and reduces the modeled $T$-gate count by $51.9\%$–$56.7\%$ relative to the standard analytical
polynomial obtained from an error-function approximation.
Read the full abstract ▾
All authors Thiago O. Maciel, Giancarlo Camilo, Allan Tosta, Daniel Stilck-França, Leandro Aolita, Thais de Lima Silva
Strongly Correlated Electrons
The ALPS project release 3.0: open source software for strongly correlated systems
F. Alet, T. Chen, A. Feiguin et al. (+16 more)
We present release 3.0 of the ALPS (
Algorithms and Libraries for Physics Simulations) project, an open-source software project to develop libraries and application programs for the simulation of strongly correlated
quantum lattice models such as
quantum magnets,
lattice bosons, and strongly correlated
fermion systems. As in previous releases, development is centered on common data formats, on libraries to simplify and speed up code development, and on full-featured simulation programs that let non-experts carry out serial or parallel numerical simulations using the important
algorithms for
quantum lattice models: classical and
quantum Monte Carlo (QMC) using non-local updates, extended-ensemble simulations, exact and full diagonalization (ED), the density
matrix renormalization group (DMRG), and continuous-time QMC solvers for dynamical mean-field theory (DMFT). Major changes in release 3.0 include distribution of the pyalps binary through the Python Package Index (pip install pyalps) and through Spack for HPC systems; migration of development to GitHub with continuous integration and automated testing; relicensing of the package under the permissive MIT license; a rebuilt documentation and tutorial website, including a set of Jupyter-notebook tutorials and localized content; a broad modernization of the C++ codebase (C++17 compliance, Boost and NumPy 2.0 compatibility, and warning and dead-code cleanup) together with a major DMRG update and associated reliability and build-compatibility fixes; the removal of legacy components (the VisTrails provenance integration and the TEBD, MPS, and directed-worm-
algorithm application codes); archival of release 3.0.0 with a Zenodo DOI; and a formal governance and sustainability model developed under the US National Science Foundation (NSF) POSE program. The software is available at https://alps.comp-phys.org.
Read the full abstract ▾
All authors F. Alet, T. Chen, A. Feiguin, E. Gull, S. Iskakov, J. P. F. LeBlanc, F. Lin, A. Mirmira, G. Möller, L. Pollet, M. Rosales, V. W. Scarola, H. Shinaoka, H. Terletska, S. Todo, M. Troyer, M. Wallerberger, P. Werner, T. M. R. Wolf
Quantum Physics
On Information Propagation in Krylov Subspaces
Pablo Martinez-Azcona, Maximilian Meyer-Moelleringhof, Adolfo del Campo et al. (+1 more)
Krylov subspace methods provide an efficient tool for modeling any
quantum initial value problem as a one dimensional chain. A time ordered path formulation, in which the dynamical generator successively splits probability between paths, yields a speed limit within such dynamical subspaces. For power law growing Lanczos coefficients $b_n \sim n^\delta$ we find super ballistic lightcones, with an exponentially spreading lightcone for maximal growth $\delta = 1$. Causality in the Krylov chain constrains the Lanczos coefficients to grow at most linearly. We illustrate the validity of the above in the $SL(2, R)$ and Heisenberg Weyl
algebras, and in a many body skew field Ising model. Moreover, we extend the above bound to the family of $Q$ complexities. Lastly, mapping to a
spin chain gives a Lieb Robinson bound analogue and opens the problem to standard condensed matter methods.
Read the full abstract ▾
All authors Pablo Martinez-Azcona, Maximilian Meyer-Moelleringhof, Adolfo del Campo, Apollonas S. Matsoukas-Roubeas
Instrumentation and Methods for Astrophysics
The Nancy Grace Roman Space Telescope Wide Field Instrument: Characterization of the Brighter-Fatter Effect in a Roman H4RG-10 Detector using the Talbot Effect
Jennie Paine, Ami Choi, Christopher Merchant et al. (+14 more)
The Nancy Grace Roman Space
Telescope will conduct several community-driven surveys using its Wide Field Instrument (WFI), which contains 18 Teledyne H4RG-10 near-infrared detectors in the focal plane. Roman's wide area survey to measure weak gravitational
lensing places particularly stringent requirements on point spread function (PSF) precision for the WFI. The brighter-fatter effect (BFE) is a nonlinear detector effect caused by the deflection of new charge away from pixels with previously accumulated charge during an exposure. As a result, the PSF measured at the detector depends on the charge contrast between pixels, and therefore the effective PSF for astronomical sources is a function of flux contrast and integration time. We present characterization of the BFE observed in a non-flight H4RG by projecting a grid of bright point sources onto the detector using the Talbot Illuminator at the NASA Goddard Space Flight Center Detector Characterization Lab. We find that the effective area of a pixel decreases by $(-2.18\pm0.04) \times 10^{-6}$ per electron of charge contrast with its neighbors, corresponding to a measurable increase in the effective PSF size over the course of an exposure. Despite charge migration, the total signal detected in the vicinity of a bright source is roughly conserved. We also analyze flat field statistics for the same detector to measure the total inter-pixel nonlinearity (IPNL), which is dominated by BFE. We present a method to correct BFE in H4RG images using the flat field IPNL results, adapted from an extant optical correction technique. The correction reduces the observed BFE by about a factor of 25 as measured by the effective pixel area change, but there is remaining nonlinearity in the corrected data. We discuss possible sources of residual nonlinearity and suggest avenues to investigate optimal BFE correction in Roman survey images.
Read the full abstract ▾
All authors Jennie Paine, Ami Choi, Christopher Merchant, Timothy D. Brandt, Jeffrey Kruk, Joshua E. Schlieder, Edward J. Wollack, Analia Cillis, Jahmour J. Givans, Robert Hill, Christopher Hirata, Dana R. Louie, Gregory Mosby, Anna Porredon, Bernard J. Rauscher, Maxime J. Rizzo, Rujula Yete
High Energy Astrophysical Phenomena
The Virgo Cluster as a Case Study for Ancient Cosmic Ray Halos in Radio and X-rays
Philip F. Hopkins, Emily M. Silich, Jack Sayers et al. (+2 more)
Virgo is the nearest cool core (CC)
galaxy cluster, and is archetypal of the class of X-ray weak CCs. Despite the X-ray continuum implying a large cooling rate ($\gtrsim 10\,{\rm M_{\odot}\,yr^{-1}}$), measurements of cooler gas phases,
dust, and
star formation show almost no cooling outside the M87
nucleus. It is also known that M87$^{\ast}$ is a prodigious source of
cosmic ray (CR)
leptons observed in
radio and $\gamma$-rays, and recent
radio observations have shown that the M87
radio halo is low-frequency and its spectral index steepens rapidly away from the injection region+jets, consistent with an aging, low-energy CR population. We demonstrate that coupling X-ray and
radio to newer Faraday rotation limits on the
magnetic field, this implies the CRs within the CC radius ($\lesssim 40$kpc) must contribute significantly to the total pressure and soft X-ray emission, through a combination of Coulomb heating the gas and inverse-Compton (IC) scattering microwave background
photons to $\sim$keV. We show this can explain the apparent correspondence between jet/
radio power and X-ray cooling luminosity, and why the X-ray inferred cooling rate appears much larger than otherwise observed. This would also explain why the central X-ray inferred
metallicity at $\lesssim 10\,$kpc appears to be suppressed relative to the UV/optical inferred
metallicity (in a wavelength-dependent manner). Compared to more extreme
radio+X-ray sources like Perseus, we find the observations require a softer 'escaping' CR spectrum (escaping or accelerated outside $\gtrsim 1\,$kpc from M87$^{\ast}$).
Read the full abstract ▾
All authors Philip F. Hopkins, Emily M. Silich, Jack Sayers, Sam B. Ponnada, Isabel S. Sands
Machine Learning
Reinforcement Learning on the Discrete Composition Channel of a Crystal Generator: Validated Gains and Reward Hacking
Pawan Prakash, Philipp Höllmer, Addis Fuhr et al. (+4 more)
Inverse materials design is a long-standing goal of computational materials discovery. Generative models for crystalline materials are typically trained to match the distribution of a structure database, while nothing in their training objective points them at specific design goals such as targeted properties. We use group-relative policy optimization (GRPO) to align a generative model based on
stochastic interpolants and discrete flow matching with general black-box reward functions through reinforcement learning. Atom types are generated by a discrete flow and the policy gradient of our generalization of GRPO directly acts on the likelihoods of the atom-type transitions, which differentiates our work from previous reinforcement-learning approaches for
diffusion and flow-based generative models of crystalline materials. We introduce a reward function that raises the yield of metastable, unique and novel structures (mSUN) from 13.4% for the pretrained model to 45.5% for the reinforced model, as evaluated by a community
benchmark. Our reward also improves the performance of a reinforcement learning framework for crystalline materials based on latent denoising
diffusion models. At the same time, we find that directly reinforcing atom-type transition likelihoods enables reward exploitation that has to be prevented with explicit guards. The same analysis also exposes a gap in the community metric. Single-element structures in distinct packings are counted as metastable, unique and novel materials and inflate mSUN without yielding any new compounds. A stability claim is only as good as its reference hull. We report every result split by the number of reference phases behind it and argue that
benchmarks should do the same.
Read the full abstract ▾
All authors Pawan Prakash, Philipp Höllmer, Addis Fuhr, Peter Hirschfeld, P. Ganesh, Stefano Martiniani, Richard Hennig
Networking
Performance Analysis of ISAC Slicing with Sensing-Assisted Communication and Communication-Assisted Sensing
Daniel Campos and Jacek Kibilda
Integrated Sensing and Communication (
ISAC) will allow 6G networks to provide both sensing and communication services through shared infrastructure and spectrum resources. However, network slicing that views these services as distinct slices with distinct objectives and performance requirements will inevitably miss the emergent value of integrating them. By considering Sensing-Assisted Communication (SAC) and Communication-Assisted Sensing (CAS), which create interdependencies between services, we can observe trade-offs in resources dimensioned to support each slice and, ultimately, the
ISAC slicing performance. As such, this paper proposes a performance analysis of an
ISAC slicing framework that accounts for resource allocation and sensor density, as well as SAC and CAS, dimensioned to achieve slice-specific Service-Level Agreement (SLA) satisfaction. Numerical simulations, across different bandwidth allocations, sensor deployments, and demand scenarios, show that
ISAC slicing performance is fundamentally constrained by sensing resolution, shaped by demand, and scaled by sensor deployment. These findings provide insights for the design of future
ISAC slicing mechanisms.
Read the full abstract ▾
High Energy Physics (Phenomenology)
A Powheg generator for $t$- and $s$-channel single-top production and decay including non-resonant effects
Melissa van Beekveld, Robin van Bijleveld, Silvia Ferrario Ravasio et al. (+2 more)
We present a next-to-leading order + parton shower generator for $t$- and $s$-channel single-top production in
hadron colliders, for a leptonically-decaying top, including off-shell and non-resonant effects. The generator, publicly available in the POWHEG BOX RES framework, describes the hadro-production of a $b$-jet and a leptonically decaying $W$-
boson, accompanied by an additional jet, i.e. $pp\to \ell^{+} \nu_{\ell}\, j_b\, j $ and $pp\to \ell^{-} \bar{\nu}_{\ell}\, j_{\bar{b}}\, j $, where $j_b$ ($j_{\bar{b}}$) is a jet containing a bottom (anti)
quark, and $j$ is a generic jet. We also provide an interface to the Pythia parton shower, suited to matching in the presence of decaying resonances such as the top
quark. Our matching is designed to combine the advantages of a five-flavour calculation with the description of bottom
quark mass effects in final-state jets. We perform a phenomenological study using fiducial cuts and discuss shower effects.
Read the full abstract ▾
All authors Melissa van Beekveld, Robin van Bijleveld, Silvia Ferrario Ravasio, Tomas Jezo, Tommaso Saracco
Astrophysics of Galaxies
Internal Radiative Feedback Triggers Global Collapse and Direct-Collapse Black Hole Formation in Metal-free Atomic-Cooling Haloes
Masaki Kiyuna, Sunmyon Chon, Kazuyuki Omukai et al. (+1 more)
The formation of heavy seed black holes from the collapse of supermassive
stars (SMSs) is a promising pathway to the supermassive black holes (SMBHs). In the conventional direct-collapse scenario, however, SMS formation requires an extremely strong external far-ultraviolet (FUV) field to suppress H$_2$ cooling and fragmentation. Here, we use cosmological radiation-hydrodynamic simulations to show that SMSs can form in metal-free atomic-cooling
haloes even under subcritical FUV backgrounds with $J_{21}=1$-$10^2$. We follow the
evolution for $\sim1$-$2~{\rm Myr}$ after the onset of
star formation, including FUV and ionizing radiation from
stars formed within the
halo.
Star formation proceeds through a characteristic two-stage
evolution. H$_2$ cooling initially induces fragmentation and the formation of ordinary Pop III
stars. Internal FUV radiation from these
stars then photodissociates H$_2$ and suppresses further fragmentation, while the surrounding warm gas reservoir contracts under self-
gravity and undergoes global collapse through atomic cooling. This collapse drives inflow rates of $\sim0.1$-$1~{\rm M_\odot~yr^{-1}}$ and produces rapidly accreting protostars that reach $\sim2$-$6\times10^4~{\rm M_\odot}$ by the simulations end. Since rapid
accretion is still ongoing, these objects are expected to grow to $\sim10^5~{\rm M_\odot}$ SMSs and then form direct-collapse black holes (DCBHs). The global collapse is accompanied by clustered
star formation, producing a compact, top-heavy
stellar system around the SMSs. The estimated DCBH abundance, $n_{\rm DCBH}\sim10^{-3}$-$10^{-2}~{\rm cMpc^{-3}}$, is high enough to provide a plausible progenitor population for Little Red Dots and present-day SMBHs. Our results demonstrate that internal
stellar feedback can trigger, rather than prevent, SMS formation, substantially relaxing the external FUV requirement for heavy-seed formation.
Read the full abstract ▾
All authors Masaki Kiyuna, Sunmyon Chon, Kazuyuki Omukai, Raffaella Schneider
Cosmology and Nongalactic Astrophysics
Anisotropic cosmic birefringence estimates for forthcoming CMB experiments
Matteo Longo Minnolo, Matteo Billi and Giuseppe Puglisi
Cosmic Birefringence (CB) is a unique probe of physics beyond the Standard Model, potentially sourced by an
axion-like field coupled to
photons via a Chern-Simons interaction. Recent Planck analyses have strengthened evidence for a non-zero isotropic rotation angle, yet its anisotropic component remains undetected, despite being the key to distinguishing a homogeneous quintessence-like field from spatially varying
axion-like dark matter. We forecast the sensitivity of forthcoming ground- and space-based CMB experiments to anisotropic CB, applying a patch-based localization technique to simulated CMB-plus-noise maps with white noise levels representative of Simons Observatory SAT- and LiteBIRD-like sensitivities. Local rotation angles are reconstructed on independent sky patches by minimizing the $\chi^2_{EB}$ statistic built from the $EB$ D-
estimator. Under our idealized simulations, the forecast angular power spectrum improves current Planck-based constraints on the CB anisotropy statistical uncertainty by a factor of 26-40 for a satellite-like mission and 11-15 for a realistic ground-based footprint. Translating these into constraints on the scale-invariant amplitude, fitting the reconstructed bandpowers with a scale-invariant amplitude model, we forecast with a 95 % C.L. $A_{\rm CB} < 3.5\times10^{-4}\,{\rm deg}^2$ for the space-like mission and $A_{\rm CB} < 7.1\times10^{-4}\,{\rm deg}^2$ for the ground-based footprint, improving on the current full-sky Planck constraint by a factor of about 20 and 10, respectively. These values represent upper bounds on the achievable gain, and indicate that forthcoming CMB experiments will be able to further shed light on anisotropic CB.
Read the full abstract ▾
Solar and Stellar Astrophysics
Spectroscopic monitoring of long secondary period variables. The case of J-type carbon stars
L. Planquart, A. Jorissen and H. Van Winckel
Long-secondary-period (LSP) variables are a subclass of long-period variables whose origin is a matter of debate. J-type carbon
stars are a subclass of carbon
stars enriched in 13C, and the origin of this chemical peculiarity is unclear. In this paper, we report on the discovery of a LSP in two J-type carbon
stars: VZ Per and UV Cam. Methods. A decade-long spectroscopic monitoring was used to obtain a long-term radial-velocity curve for the two
stars. From a frequency analysis, we isolated a LSP signal. The retrieved periodicity was then compared with changes in spectral profiles and available light-curve variability to further constrain the origin of the modulation. Results. For VZ Per, the LSP derived from the radial-velocity curve (~960 days) is correlated with the light-curve dimming event and a variation in the Balmer absorption line profile. These results translate into changes in the gas and
dust column densities with the phase of the LSP, with the maximum gas column density occurring ahead of the
dust. For UV Cam, the radial velocity monitoring yields a period of ~1178 d with an amplitude of about 1 km/s. Interpreting the RV modulation as a Keplerian signal for both
stars, a low-mass (<0.1Msun) close companion is found. In this scenario, variations in gas and
dust column densities with (orbital) phase trace the material in the expanding comet-like tail trailing the substellar companion. Conclusions. These observations provide time-dependent variations of the gas and
dust in LSP variables that are consistent with the current binary scenario. Furthermore, the detection of LSP among J-type carbon
stars anchors their current evolutionary status in the
dust-producing region of the thermally-pulsing AGB phase.
Read the full abstract ▾
High Energy Physics (Theory)
SU($N$) QCD, four-manifolds and Segre numbers
Elias Furrer and Ranveer Kumar Singh
We study correlation functions of topologically twisted $\mathcal N=2$ supersymmetric QCD with gauge group $\mathrm{SU}(N)$ on a compact four-
manifold $X$. We derive the Coulomb branch integral for all $N$, including new couplings to background fluxes for the
flavor symmetry. For $X$ of simple type with $b_2^+>1$, the correlators become sums over a finite number of strong-coupling singularities on the Coulomb branch, and can be expressed using a number of Coulomb branch couplings. The sum contains the $\mathcal N=1$
vacua, which become the supersymmetric
vacua when the $\mathcal N=2$ theory is broken to $\mathcal N=1$ by an adjoint mass deformation, and whose contribution we attribute to the instanton branch. We focus on asymptotically free $\mathrm{SU}(N)$ QCD with $N_f<2N$ equal mass hypermultiplets, and
conjecture that its instanton contribution to topological correlators agrees with the generating function of virtual Segre numbers of rank $N$, which are intersection numbers on the
moduli space of semistable sheaves on an algebraic surface. We present a precise dictionary for all $N$ between universal series for Segre numbers and Coulomb branch couplings at the $\mathcal N=1$
vacua. The calculation of these couplings requires new higher-rank techniques, involving Siegel modular forms, an exact $\mathbb Z_N$
symmetry, charge conjugation
symmetry, factorisation of the Seiberg–Witten curve, and the blowup formula. For pure $\mathrm{SU}(N)$, our result agrees with earlier work by Mariño and Moore, and the twisted correlators give the $\mathrm{SU}(N)$ Donaldson invariants of $X$. For $\mathrm{SU}(3)$ QCD, we determine closed-form expressions for all 13 Coulomb branch couplings for each $N_f=1,\dots, 5$, and find precise agreement with results of Göttsche and Kool for virtual rank 3 Segre numbers.
Read the full abstract ▾
Computer Vision
Streaming Multi-Track Timeline Control for 3D Human Motion Generation
Yangsong Zhang, Anujith Muraleedharan, Rikhat Akizhanov et al. (+3 more)
Text-driven human motion generation has advanced substantially, yet most methods assume instructions are available before synthesis. Interactive applications require responding to new instructions while continuing ongoing actions, such as answering a phone while walking. Existing approaches address streaming generation or simultaneous composition without explicitly combining streaming instruction arrival with independently timed, overlapping actions. We introduce streaming multi-track timeline control and propose TimelineControl to incorporate new instructions alongside ongoing actions. Interval-aware conditioning preserves instruction timing, while causal part-structured representations and part-aware denoising coordinate concurrent actions across body regions. We also construct TimelineMotion, a
dataset with overlapping instruction intervals and body-part annotations. Experiments on TimelineMotion and MTT demonstrate improved semantic
alignment and temporal adherence over evaluated streaming baselines, including models retrained on the same data. Ablations and human evaluations validate our design, complemented by spatial conditioning and
humanoid execution demonstrations. Our code, data and models will become publicly available.
Read the full abstract ▾
All authors Yangsong Zhang, Anujith Muraleedharan, Rikhat Akizhanov, Gül Varol, Fabio Pizzati, Ivan Laptev
Artificial Intelligence
Training Numerical Intelligence via Auto-Diagnosis and Skill Discovery
Peter Chen and Wotao Yin
AI
agents are becoming increasingly capable of generating scientific code, but generating code is not the same as improving the
algorithms behind it. For numerical solvers, execution feedback can expose poor performance, but rarely reveals its underlying cause and how to address it. We introduce Auto-Diagnosis and Skill Discovery (ADSD), a framework that links numerical diagnosis to reusable solver self-improvement. ADSD follows a diagnosis-first paradigm that first explains why a solver performs poorly, then uses this diagnosis to guide the discovery of appropriate numerical methods. The resulting knowledge is packaged into reusable solver skills, turning solver improvement from trial-and-error editing into a structured process of diagnosis, discovery, and implementation. Across four challenging numerical domains–power flow equation, AC optimal power flow control, stiff ordinary differential equations, and heterogeneous
diffusion PDEs–ADSD consistently improves solver accuracy, robustness, and efficiency. On GOC-500 power flow, for example, ADSD reduces mean solver error by nearly $71\times$, with improvements further transferring to unseen grid
topologies and operating regimes.
Read the full abstract ▾
Commutative Algebra
Peskine and Szpiro's Dimension inequality holds over local rings of depth at most one
Lars Winther Christensen and Saeed Nasseh
We prove that the general version—the one that includes the dimension of the tensor product—of Peskine and Szpiro's Dimension Inequality holds over noetherian local rings of depth at most one.
Mathematical Physics
Time-dependent integrable quantum field theories with dynamical boundaries
Pradip Kattel
We present a general method to construct and solve time-dependent integrable
chiral quantum field theories on finite intervals from autonomous scattering and reflection data. Spatial homogeneity forces the spectral labels of right- and left-moving particles to be affine with a common slope, so bulk and boundary couplings sample one spectral trajectory. A complete round trip shifts the spectral labels and gives boundary
quantum Knizhnik–Zamolodchikov equations. The Yang–Baxter and reflection equations provide local consistency, while boundary-qKZ flatness provides global consistency. We reconstruct the local
Hamiltonian by Cayley inversion of the bulk and boundary matching
matrices. We apply the construction to rational $SU(N)$ scattering with fundamental boundary
impurities and to an $O(3)$ model with a
spin-$\tfrac12$
impurity. For the $SU(2)$
impurity, we construct explicit boundary-qKZ solutions as
lattice sums of off-shell boundary Bethe vectors and recover the static open-chain Bethe equations when the drive rate vanishes
Read the full abstract ▾
Quantum Physics
Modular Neural Decoder for Surface Code Memory and Logic on Superconducting Processors
Moritz Lange, Ole Fjeldså, Gustaf Jonasson Johansson et al. (+1 more)
Neural decoders for
quantum error correction have achieved high accuracy on surface code
memory experiments, but practical fault-tolerant computation requires decoders that transfer across devices, noise conditions, and logical operations. We introduce a modular graph-recurrent decoder designed for this setting. Built from reusable local syndrome encoders, it is pretrained in simulation and fine-tuned to hardware using a common pipeline, with only lightweight spatial aggregation adapted between experiments. The decoder performs competitively on conventional
memory benchmarks, outperforming minimum-weight perfect matching (MWPM) at distances 3 and 5 on an IBM Nighthawk-generation processor and closely approaching AlphaQubit on Google's below-threshold hardware data. We then test the framework in two settings beyond
memory decoding on the same IBM device. First, in an $X$-type stability experiment, we continuously tune the effective measurement error rate using a variable pre-readout rotation and train a single decoder jointly across the resulting noise range. It matches or outperforms MWPM across the sweep without requiring a separately calibrated decoder for each setting. Second, we decode a $ZZ$-merge
lattice-surgery experiment between two distance-3 surface-code patches, improving the decoded correlation between the merge outcome and the final $ZZ$ parity relative to MWPM on real hardware. Together, these experiments demonstrate neural decoding that is portable not only across hardware, but also across qualitatively different surface code observables and logical operations, using models trainable on single-GPU academic resources.
Read the full abstract ▾
All authors Moritz Lange, Ole Fjeldså, Gustaf Jonasson Johansson, Mats Granath
Number Theory
Exponential generating functions for Apostol-Dedekind and related sums
Yilmaz Simsek
One of the primary objectives of this paper is to construct generating functions for Apostol-Dedekind sums. These exponential generating functions greatly facilitate the discovery of new formulas and relations involving these sums. Other primary objectives of this paper are to define Apostol-Dedekind sums of higher-order and other certain family of finite sums, and to construct their generating functions. These generating functions include the Bernoulli
polynomials and numbers, the Euler
polynomials and numbers, the Array
polynomials, the Stirling numbers, the Carlitz
polynomials with their partial derivative equations, and the Hardy sums etc. Moreover, by using these functions, many novel formulas for the Apostol-Dedekind sums of higher-order are derived.
Read the full abstract ▾
Astrophysics of Galaxies
SCALES. II. The Internal Composition of Giant Star-Forming Clumps at $z = 1 - 4$
Boris S. Kalita, Adélaïde Claeyssens, Luis C. Ho et al. (+11 more)
Star-forming structures of scale $\sim0.5-1\,\mathrm{kpc}$, dubbed '
giant' clumps, are routinely detected in unlensed surveys of $z > 1$
galaxies. Explaining them requires characterizing their internal structure. In this work, part of the
Star-Forming Clumps As Layered Emergent Structures (SCALES) series, we hence ask: What is inside a
giant clump? We analyze 23 lensed
galaxies at $z=1-4$ with JWST/NIRCam using a difference-of-Gaussian decomposition to model (lensed) clumps at $\approx 50-400\,\mathrm{pc}$ scales, and the surrounding extended component (EC) at
giant clump scales. Associating clumps to ECs, we identify 72
giant clumps. We find the constituent clumps above our mass completeness limit ($10^{7.1}\,\mathrm{M_\odot}$) in a
giant clump to be consistent with random sampling of the clump
stellar mass function (CSMF), hence holding no information about their host. This agrees qualitatively with both prevailing scenarios for
giant clumps: (I) long-lived structures with constantly renewed constituents, and (II) blended random clump associations. As most models lack predictions for their internal mass distribution, our observations provide an empirical constraint. Finally, the CSMF follows a power law of slope $1.88\pm0.04$ above a break at $\sim10^{7.5}\,\mathrm{M_\odot}$, close to our completeness limit. The
giant clumps follow a consistent slope ($1.95\pm0.15$), giving the same scale-free mass function slope at (lensed) clump and
giant clump scales. Within
giant clumps, both the EC mass, if made up of undetected clumps, and the mass share of the most massive clump are consistent with the CSMF slope continuing down to $\approx10^{5.5}-10^{6}\,\mathrm{M_\odot}$, while the latter rejects an extension below $10^{5}\,\mathrm{M_\odot}$.
Read the full abstract ▾
All authors Boris S. Kalita, Adélaïde Claeyssens, Luis C. Ho, Miroslava Dessauges-Zavadsky, Natascha M. Förster Schreiber, Emanuele Daddi, John D. Silverman, Naval Kishor Bhadari, Changhao Chen, Fangzhou Jiang, Ke Wang, Jinning Liang, Alvaro Segovia Otero, Jinyi Shangguan
Astrophysics of Galaxies
SCALES. I. Bridging Lensed Clump - Giant Clump Scales in Lensed Galaxies at $z=1-4$
Boris S. Kalita, Adélaïde Claeyssens, Luis C. Ho et al. (+11 more)
Dense
star-forming regions in $z>1$
galaxies, often termed 'clumps', have been studied across two distinct regimes of
telescope resolution, each motivating interpretations in tension with the other. In the SCALES series (
Star-Forming Clumps As Layered Emergent Structures), we aim to connect
star-forming structures across scales, from the $\sim100\,\mathrm{pc}$ clumps in lensed studies to the $\sim\mathrm{kpc}$-scale ”
giant” clumps in unlensed surveys. In this first paper, we present a difference-of-Gaussian (DoG) decomposition that simultaneously models $\approx 50-400\,\mathrm{pc}$ (lensed) clumps and $\sim\mathrm{kpc}$-scale component (the extended component, or EC), which would together constitute the
giant clumps seen at lower resolution. Here the EC is treated as a scale-separated local background rather than a physical object, absorbing all light that is not a compact clump. Applying this to 23 moderately lensed ($\mu \approx 2-9$)
galaxies at $z_{spec} = 1-4$ to evaluate clump properties (deferring ECs and
giant clumps to SCALES-II), we measure a clump
stellar mass function slope of $1.88 \pm 0.04$, consistent with the $\approx2$ expected for hierarchical
star formation, and with that of $\sim\mathrm{kpc}$-scale clumps in unlensed
galaxies. We also show that empirical correlations of clump surface densities and clump-to-host mass ratio with
redshift are exaggerated by
redshift-dependent detection bias, so $\approx 50-400\,\mathrm{pc}$ clumps are largely co-similar across our range ($z = 1-4$). Finally, combining clump mass, size and age estimates with literature gas-dispersion measurements at similar scales, we argue that $\sim100\,\mathrm{pc}$ clumps are themselves composed of
star-cluster-like sub-components at much smaller scales.
Read the full abstract ▾
All authors Boris S. Kalita, Adélaïde Claeyssens, Luis C. Ho, Miroslava Dessauges-Zavadsky, Natascha M. Förster Schreiber, Emanuele Daddi, John D. Silverman, Naval Kishor Bhadari, Changhao Chen, Fangzhou Jiang, Ke Wang, Jinning Liang, Alvaro Segovia Otero, Jinyi Shangguan
High Energy Physics (Theory)
Magnetic Higher-Form Symmetries as Dual Fundamental $\infty$-Groupoids
Alonso Perez-Lona
We develop a higher-
geometric framework for
magnetic higher-form
symmetries of pre-
quantum field theories. The
magnetic (topological) charges of a sigma-model with stacky target are encoded by the shape of the target, equivalently of its stack of fields, which is op-dual to the flat
automorphism group controlling electric
symmetries.
Magnetic symmetry backgrounds are obtained by dualizing these charges against a chosen
moduli stack of topological field theories, generalizing Pontryagin duality, and act on the theory by stacking. Cohesive modalities allow for targets with non-finite and non-discrete data. When the dualizing stack is a grouplike $E_\infty$-object, we prove a fibration
theorem expressing the
magnetic symmetries of an electrically gauged theory in terms of those of the ungauged theory and of the gauge group, computed by a homotopy-fixed-point spectral sequence with an explicit first page. For
moduli stacks of higher $U(1)$-gerbes with connection, encoding bulk topological local Lagrangians, we prove a factorization
theorem: the
magnetic symmetry backgrounds of any connected target split, non-canonically, into Pontryagin duals of its integral homology groups, and are necessarily flat, as a consequence of dualizing discrete charges against a target with connections. We illustrate both
theorems with pure $\mathrm{Pin}^\pm(2)$ gauge theory: $O(2)$ carries invertible
magnetic $\mathbb{Z}_2$
symmetries in form degrees $d-3$, $d-4$ and $d-5$ that are absent for $\mathrm{Pin}^-(2)$, the first two distinguished by a single differential of the spectral sequence. Gauging and physical applications are developed in upcoming work.
Read the full abstract ▾
Quantum Physics
Realizing Braided Fusion 2-Categories as (3+1)D Mixed-State Topological Orders
Ryohei Kobayashi, Abhinav Prem and Matthew Yu
We construct large families of mixed-state topological orders in (3+1)D, including intrinsically mixed-state topological orders. We realize these theories by relating the data parameterizing braided fusion 2-categories to the strong higher-form
symmetries and their anomalies that characterize mixed-state topological order. We focus here on bosonic theories, which include theories which admit an emergent
fermion in the spectrum of line operators. Our general framework is exemplified on the
lattice, where we present concrete models and identify how the data of braided fusion 2-categories are physically encoded. We then formalize this data and show that every braided fusion 2-category arises from a general construction (decoherence followed by strong and weak gauging) as the category of topological operators of a (3+1)D mixed state, with intrinsically mixed-state topological orders distinguished by the presence of strong higher-form
symmetries whose anomalies cannot be realized in any (3+1)D gapped ground state. Notably, we show that (3+1)D admits a new mechanism with no (2+1)D counterpart: intrinsically mixed-state order whose topological operators braid non-degenerately, but which carries a gravitational anomaly.
Read the full abstract ▾
High Energy Physics (Phenomenology)
No Einstein Relation, No Problem: How Relativistic Weakly Coupled Heavy Quarks Equilibrate Beyond Leading Logarithm Anyway
Jean F. Du Plessis and Bruno Scheihing-Hitschfeld
Relativistic heavy
quarks equilibrate even when their microscopic drag and momentum
diffusion violate the Einstein relation. We investigate how equilibration proceeds as a function of the coupling and compared to simplified Fokker-Planck models. To do this, we present a detailed calculation of the momentum-transfer kernel in weakly coupled non-Abelian
plasmas through strict $\mathcal{O}(g^4)$, and study the resulting dynamics at leading order in the inverse heavy-
quark mass. Beyond leading logarithm, its non-Gaussian structure plays a crucial role in equilibration. We examine closely the analytic structure of the
evolution kernel, determine the large-order asymptotics of the momentum-transfer cumulants, and show how the kernel's bounded analytic domain fixes asymmetric exponential tails at the level of the momentum transfer probability. Strongly coupled $\mathcal{N}=4$ SYM also features an approximately Gaussian core and exponential tails, but differs in the nature of the kernel's singularities and the velocity dependence of the transport coefficients and tail exponents. These differences have distinct dynamical consequences: for the parameters studied, the persistence of the relativistic population in steeply falling spectra is largely explained by drag alone at very weak coupling, whereas fluctuation-induced survivor bias produces a large relative enhancement at strong coupling. Nevertheless, the shared qualitative structure between weak and strong coupling points the way towards a practical phenomenological description of heavy
quarks anchored in first-principles field theory.
Read the full abstract ▾
High Energy Physics (Theory)
Complex Worldline Instantons in (Anti-)de Sitter Space
Alberto Castellano, Jinwei Chu, Eliseu Kloster et al. (+1 more)
We study complex saddle points of the worldline path integral for massive charged particles in two-dimensional maximally symmetric backgrounds threaded by homogeneous electric fields. Within the worldline formalism, we show that the semiclassical saddles generically require complexifying both the target spacetime and the proper-time parameter. The resulting solutions comprise two slices of a single complex family, distinguished by the direction of proper-time
evolution. The one evolving in imaginary proper time carries an action that typically controls the exponential suppression of non-perturbative effects, while that evolving in real proper time may encode additional complex phases responsible for
quantum interference between saddles. Treating this family as a whole, we give a prescription for constructing and classifying all topologically inequivalent instanton sectors, and we find that the worldline action depends only on their topological data, namely winding numbers around the compact directions and certain isolated punctures. Moreover, in Anti-de Sitter spacetime, the qualitative structure of the solutions is governed by the dimensionless ratio $R |eE|/m$ ($R$ being the
curvature radius), which distinguishes sub-extremal, extremal, and super-extremal regimes. By contrast, no analogous distinction occurs in de Sitter—where the instantons are insensitive to this ratio—or in Minkowski spacetime, which lacks a
curvature scale altogether. We also recast the action as a sum over residues of the relevant (non-conserved) conjugate momentum, making transparent which analytic data control each contribution. Finally, we discuss the subtleties that arise when solutions built in various local patches are lifted to global (Anti-)de Sitter space. These impose further necessary conditions on which saddles can in principle contribute to the one-loop path integral.
Read the full abstract ▾
All authors Alberto Castellano, Jinwei Chu, Eliseu Kloster, Jamie A.P. Law-Smith
Robotics
GOTT: Object-centric Dexterous Manipulation with a Reusable Cross-Embodiment Primitive
Yulin Liu, Lai Wei, Yen-Jen Wang et al. (+4 more)
Foundation models and large-scale human data provide rich sources of manipulation intent, but translating this intent into multi-fingered
robot behavior remains difficult. Dexterous hands still lack a reusable low-level primitive that reliably establishes contact across tasks and embodiments. We propose GOTT, a reach-acquire-move framework built around a single cross-embodiment contact-acquisition primitive. Given a
robot-agnostic object trajectory and a reach specification, GOTT first brings the hand near a task-relevant contact region. The shared closed-loop primitive then establishes stable contact from this approximate initialization, and a pose-conditioned controller tracks the desired object motion. Reach specifications may come from future-aware planning, external models, or human demonstrations, while the primitive and tracking backend remain unchanged. Simulation and real-world experiments show that GOTT is able to establish robust contact across diverse objects, arm-hand platforms, and seen and unseen hand morphologies. It also consistently improves end-to-end task success over open-loop grasp execution.
Read the full abstract ▾
All authors Yulin Liu, Lai Wei, Yen-Jen Wang, Akash Sharma, Pieter Abbeel, Henrik I. Christensen, Haozhi Qi
Quantum Physics
Driven-Dissipative Ground State Preparation: Mixing Time and Randomness
Paul Cazeaux, Marius Junge and Diyi Liu
Dissipative protocols offer a route to ground state preparation that does not require a large initial overlap with the ground state. To move beyond time independent Lindbladians and randomized Lindbladians sampled from a fixed distribution, we study driven-dissipative protocols in which the generator varies over time. Our construction uses a family of randomized Lindbladians whose jump operators are obtained by averaging the
unitary evolution of random
matrices under a fixed target
Hamiltonian. We show that empirical averages of these Lindbladians drive arbitrary initial density
matrices to low energy states. The resulting mixing time bounds depend on the
Hamiltonian only through its
eigenvalue distribution and remain valid even for cold starts.
Read the full abstract ▾
Signal Processing
Two-Dimensional Magnetometer Calibration Using Inertial Sensors for Heading Estimation
Mohammad Sadeghi and Rouhollah Amiri
Magnetometers have numerous applications, one of the most important being heading determination. In practice, magnetometer measurements are affected by sensor manufacturing errors and by hard-iron and soft-iron effects, and therefore calibration is required for reliable heading estimation. Many conventional three-dimensional calibration methods require sufficiently rich multi-axis rotational excitation, which is not feasible in some applications, particularly for in-situ or online calibration. This paper proposes a two-dimensional magnetometer calibration method aided by inertial sensors. The proposed method is simple, closed-form, and non-iterative, and uses a single-axis rotational motion together with measured roll and pitch angles. After calibration, the corrected
magnetic measurements are used for heading determination. Simulation and experimental results demonstrate the effectiveness of the proposed method under limited-rotation conditions.
Read the full abstract ▾
Machine Learning
Retrieval-Centric Deep Learning in Growing Nonparametric Neural Networks
Maximilian Schlegel, Rajai Nasser, Seijin Kobayashi et al. (+6 more)
We investigate a general-purpose layer for deep learning that, instead of compressing arbitrary-size training data into fixed-size weight
matrices, stores a new pair of key-value representations for every data point during training, and retrieves and recombines these representations through an
attention mechanism at inference time - resulting in a growing neural net (NN). While Irie et al. (arXiv:2202.05798) have put forward this perspective from the classic duality expressing any linear layer in a deep NN trained by gradient descent as linear
attention (LA) over the training data points, replacing LA by more powerful
attention functions, as they suggest, turns out to be non-trivial: we show that naively applying learning rules from the LA case to advanced kernels does not lead to principled optimization. Here we fill this gap and develop functional gradient-based learning rules for kernelized
attention layers, based on radial basis function (RBF) and softmax-like kernels - establishing the principled "
retrieval-centric deep learning" (RCDL) paradigm. Empirically, we demonstrate the promising performance and learning-efficiency of RCDL on image classification and synthetic teacher-student learning tasks. Moreover, we show that replacing LA in the dual form of NNs by advanced LA variants, namely MesaNet/DeltaNet, yields a formal connection to recently proposed optimizers for conventional fixed-size NNs, offering a novel perspective on deep learning optimization.
Read the full abstract ▾
All authors Maximilian Schlegel, Rajai Nasser, Seijin Kobayashi, Yanick Schimpf, Oliver Sieberling, Robert Obryk, Kazuki Irie, João Sacramento, Johannes von Oswald
Cryptography and Security
Beware EviLLM: Enabling Vulnerability Injection via Large Language Models
Zeezoo Ryu, Simon Chung, Muhammad Faraz Karim et al. (+4 more)
Advances in large
language models (LLMs) have enabled AI-driven code generation from natural
language specifications, introducing new attack surfaces for injecting vulnerabilities into software. Prior work has studied this problem only in benign settings where vulnerabilities are introduced inadvertently, or under unconventional threat models where the LLM itself is malicious (backdooring) or the user is the attacker (jailbreaking). In this paper, we study a more realistic threat model: a third-party adversary, with capabilities comparable to existing cybercriminals, compromises the AI code generation pipeline to deliberately introduce vulnerabilities. We call this the EviLLM attack. We have implemented two instances of EviLLM, each of which only requires the underlying LLM to be accessed through a compromised account or browser, and can inject vulnerabilities from 13 CWE classes. As we show in our feasibility study, both attack vectors are already used to implement many existing cyberattacks. Our user study shows that 7 out of 8 and 10 out of 13 participants did not notice the vulnerabilities injected by the two instances of EviLLM, and 13 out of 21 participants "rarely" or "never" considered the risk of an attack like EviLLM.
Read the full abstract ▾
All authors Zeezoo Ryu, Simon Chung, Muhammad Faraz Karim, Anna Raymaker, Karan Singh Jodha, Yash Chaturvedi, Sukarno Mertoguno
Combinatorics
Exact and asymptotic enumeration of unrestricted binary phylogenetic networks through automorphism weights
Josep Batle
Let $\cP_{\ell,k}$ be the set of rooted binary phylogenetic networks with $\ell$ labelled leaves, $k$ reticulations and no parallel edges. We write $|\cP_{\ell,k}|=W_k(\ell)+D_k(\ell)$, where the weighted count $W_k(\ell)$ adds the inverse orders of the leaf-fixing
automorphism groups and the defect $D_k(\ell)$ collects the remainder. The weighted count satisfies, for every $k$, a recursion over the source layers of the tree-component structure that involves neither a list of component graphs nor any distinction between symmetric and asymmetric configurations, and its exponential generating function is a Laurent
polynomial in $\sqrt{1-2x}$.
Automorphism groups of networks are $2$-groups, elementary abelian for $k\le5$ but not in general. For $k\le5$ the defect is the weighted count of networks with a distinguished involution, which obeys an extension of the same recursion. An exact symbolic evaluation of the two recursions yields $|\cP_{\ell,k}|$ in closed form for $k\le5$, the case $k=5$ being new; it reproduces the published counts for $k\le4$ and corrects a coefficient in a published generating function for $k=3$. For every $k$, uniformly over explicit ranges of $k$, we prove that the non-tree-child networks are a fraction $2k(k-1)/\ell$ of the tree-child networks to leading order, which gives the third term of the asymptotic expansion of $|\cP_{\ell,k}|$. We also prove that reticulation-visible networks exceed tree-child networks by the fraction $k(k-1)/\ell$, and that a uniformly random network in $\cP_{\ell,k}$ has a nontrivial
automorphism with probability $k(k-1)/(4\ell^3)$ to leading order.
Read the full abstract ▾