Statistical Mechanics · Cross-listed from cs.AI
On Distinguishing Capability Elicitation from Capability Creation in Post-Training: A Free-Energy Perspective
Yuhao Li and Shengchao Liu
Debates about large
language model post-training often treat supervised fine-tuning (SFT) as imitation and reinforcement learning (RL) as discovery. But this distinction is too coarse. What matters is whether a training procedure increases the probability of behaviors the pretrained model could already produce, or whether it changes what the model can practically reach. We argue that post-training research should distinguish between capability elicitation and capability creation. We make this distinction operational by introducing the notion of accessible support: the set of behaviors that a model can practically produce under finite budgets. Post-training that reweights behaviors within this support is capability elicitation; whereas changing the support itself corresponds to capability creation. We develop this argument through a free-energy view of post-training. SFT and RL can both be seen as reweighting a pretrained reference distribution, only with different external signals. Demonstration signals define low-energy behavior for SFT, and reward signals define low-energy behavior for RL. When the update remains close to the base model, the main effect is local reweighting, not capability creation. Within this framework, the central question is no longer whether post-training is framed as SFT or RL, but whether it reweights behaviors already within reach, or instead expands the model's reachable behavioral space through search, interaction, tool use, or the incorporation of new information.
Read the full abstract ▾
Disordered Systems and Neural Networks
Stochastic Dynamics of Domain Wall on a Racetrack: Impact of Line-Edge Roughness
Anton V. Hlushchenko, Oksana L. Andrieieva, Mykhailo I. Bratchenko et al. (+3 more)
We investigate the impact of line-edge roughness on current-driven domain wall dynamics in
ferromagnetic racetracks. Modeling the edge disorder as a spatially correlated Ornstein-Uhlenbeck process, we demonstrate that even minimal experimentally relevant roughness induces pronounced
stochastic pinning of domain walls. Notably, this stochasticity of the current-driven motion arises purely from spatial disorder, even in the absence of thermal fluctuations. The probability of a domain wall to reach a given position exhibits a robust sigmoidal dependence on the applied current, reflecting an effective distribution of depinning thresholds. At the same time, the underlying dynamics is highly nontrivial: the mean velocity exhibits a nonlinear dependence on both time and current, while the mean-square displacement exhibits a ballistic regime at short times followed by saturation due to trapping at pinning sites. These results demonstrate that line-edge roughness provides a controllable source of stochasticity and enables p-bit-like functionality in racetrack systems, offering a pathway toward hardware implementations of probabilistic and neuromorphic computing.
Read the full abstract ▾
All authors Anton V. Hlushchenko, Oksana L. Andrieieva, Mykhailo I. Bratchenko, Andriy M. Styervoyedov, Kostyantyn I. Polozhiy, Aleksei V. Chechkin
Statistical Mechanics · Cross-listed from quant-ph
Mesoscopic Regimes of Temporal Entanglement in Ergodic Quantum Systems
Sergio Cerezo-Roquebrún, Jan Thorben Schneider, Stefano Carignano et al. (+4 more)
We study temporal correlations in interacting
quantum systems through the influence functional of a half-infinite
quantum Ising chain. Using Rényi
entropies and temporal mutual information, we confirm that integrable dynamics is captured by the quasiparticle picture. In contrast, generic ergodic
Hamiltonian dynamics exhibits pronounced deviations from random-
circuit universality, and its generalization including a
symmetry accounting for energy conservation. Instead, we find a long mesoscopic regime suggestive of a slow spectral reorganization of the influence functional. Our results reveal a rich temporal structure in generic
Hamiltonian dynamics and point to limitations of existing random-
circuit paradigms at experimentally and numerically relevant timescales.
Read the full abstract ▾
All authors Sergio Cerezo-Roquebrún, Jan Thorben Schneider, Stefano Carignano, Aleix Bou-Comas, Mari Carmen Bañuls, Esperanza López, Luca Tagliacozzo
Statistical Mechanics · Cross-listed from quant-ph
Quantum trajectory simulation of two-dimensional non-equilibrium steady states with a trapped ion quantum processor
Anna Dalmasso, Arash Jafarizadeh, Julian Boesl et al. (+8 more)
Digital
quantum computers offer a promising route for studying complex many-body systems that are otherwise inaccessible by their classical counterparts. Capabilities including mid-
circuit measurements and feedback allow for simulating the dynamics of interacting open
quantum systems. Using the Quantinuum System Model H1 trapped-ion
quantum computer, we experimentally realise
quantum trajectories for a two-dimensional system of (interacting) particles-hard-core
bosons or
fermions-undergoing
stochastic driving at a source and drain at opposite corners of a square
lattice. We study the non-
equilibrium steady state with persistent current resulting from the this in/out flow of particles. The particle statistics, presence of interactions, and introduction of a
magnetic field produce measurable effects on the steady state. Our findings highlight the rich physics in this corner driven two-dimensional setup and showcases both the power and current limitations of
quantum computers as a platform to study it.
Read the full abstract ▾
All authors Anna Dalmasso, Arash Jafarizadeh, Julian Boesl, Jared Jeyaretnam, Sheng-Hsuan Lin, Andrew G. Green, Frank Pollmann, Michael Knap, Juan P. Garrahan, Henrik Dreyer, Adam Gammon-Smith
Materials Science
Photovoltaic Possibility of Cu2SiSe3 and Cu2SnS3 Ternary Chalcogenides- Single Junction to Tandem Architecture
Saptarshi Mandal, Surbhi Ramawat, Sumit Kukreti et al. (+1 more)
Cu based ternary chalcogenides are gathering
attention for sustainable energy applications due to their reduced complexity compared to quaternary alternatives. We used drift
diffusion modeling to evaluate the feasibility of photovoltaics employing ternary chalcogenide absorbers based on Cu2SiSe3 and Cu2SnS3. The device metrics are evaluated by analyzing absorber layer thickness intrinsic carrier concentration defect density and energy band
alignment at interfacial junctions. The optimized single junction Cu2SiSe3 based device configuration achieves a power conversion efficiency of 18.13 percent exhibiting a short
circuit current density of 38 mA cm^-2 and an open
circuit voltage of 0.64 V. The Cu2SnS3 based device achieves an efficiency of 15.59 percent with a short
circuit current density of 48.8 mA cm^-2 and an open
circuit voltage of 0.42 V. We examined the impact of the buffer layer on device parameters uncovering further avenues for performance improvement. Additionally we simulated a two terminal tandem
solar cell using Cu2SiSe3 Eg 1.44 eV in the upper
cell to capture
photons from the visible spectrum and Cu2SnS3 Eg 0.91 eV in the lower
cell to absorb from the infrared spectrum. The simulated tandem architecture, featuring a VOC of 1.24 V a JSC of 24.6 mA cm^-2 a fill factor (FF) of 79.2 percent and an efficiency of 24.1 percent markedly surpassed conventional single junction devices demonstrating the viability of Cu2SiSe3-Cu2SnS3 absorber-based tandem
solar cells for next generation high-efficiency
solar technologies.
Read the full abstract ▾
All authors Saptarshi Mandal, Surbhi Ramawat, Sumit Kukreti, Ambesh Dixit
Materials Science · Cross-listed from physics.bio-ph
Indirect Detection of Lactate Through Voltammetry Using Glassy Carbon Microelectrodes
Amish Rohtagia, Elisa Barts, Neharika Ravichandran et al. (+5 more)
Glassy carbon (GC) microelectrodes are increasingly being used for voltametric detection of electroactive neurotransmitters such as dopamine and serotonin. However, non-electroactive molecules including lactate, glutamate, and gamma-aminobutyric acid (GABA) cannot be directly detected using conventional voltammetry without surface functionalization. In this study, lactate oxidase was immobilized within a chitosan
matrix on lithographically patterned GC microelectrodes to enable indirect detection of lactate via enzymatic generation of hydrogen peroxide, an electroactive byproduct. The resulting hydrogen peroxide was detected using fast-scan cyclic voltammetry (FSCV), enabling indirect in vitro detection of lactate at concentrations as low as 10 nM. The functionalized GC microelectrodes were integrated into a four channel array on a 1.6 cm flexible neural probe with potential for in vivo applications. Surface morphology and bonding interactions were characterized using scanning electron microscopy (SEM) and
Fourier transform infrared (FTIR) spectroscopy. FTIR analysis confirmed successful chitosan deposition through characteristic O-H, N-H, amide, and C-O stretching bands. Hydrogen peroxide detection was concentration-dependent, while lactate detection exhibited early saturation consistent with enzyme-limited kinetics. These results demonstrate a mechanically robust GC microelectrode platform for nanomolar-level indirect lactate sensing and provide insight into the reaction-
diffusion coupling governing enzyme-based electrochemical detection.
Read the full abstract ▾
All authors Amish Rohtagia, Elisa Barts, Neharika Ravichandran, Sandra Lara Galindoa, Surabhi Nimbalkara, Mantra Mittal, Samantha Omer, Sam Kassegnea
Statistical Mechanics
Multiscale Structure of Eigenstate Thermalization
Pavel Orlov, Rustem Sharipov and Enej Ilievski
The eigenstate thermalization hypothesis provides a framework for understanding thermalization in isolated
quantum many-body systems by characterizing statistical properties of local observables in energy eigenstates. Here we demonstrate that distributions of
matrix elements in macroscopic systems may depend not only on the macrostate parameters, such as the densities of local conserved charges, but generally also on the properties of ensembles used in sampling eigenstates. To this end, we depart from the conventional analysis of microcanonical windows and consider statistical ensembles with an adjustable scale parameter prescribing the magnitude of charge fluctuations. We specifically consider an integrable field theory that permits efficient numerical sampling of
matrix elements and reliable extrapolation to the thermodynamic limit. Moreover, in this system, we identify a class of states that enables explicit closed-form computation of the suppression rate of
matrix elements. Our findings reveal an underlying multiscale structure of
matrix elements captured by a non-analytic fluctuation-scale dependence of algebraic exponents governing their statistical properties.
Read the full abstract ▾
Strongly Correlated Electrons
Gapped 1/9 Magnetization Plateau in the Anisotropic Kagome Antiferromagnet Y-kapellasite
Dipranjan Chatterjee, Paul A. Goddard, Ewan R. P. Thomas et al. (+11 more)
Fractional magnetization plateaus provide a sensitive probe of many-body
spin states in frustrated
quantum magnets, yet their microscopic origin in kagome antiferromagnets remains unresolved. This is particularly true of the mysterious $1/9$ plateau, which is predicted by theory but infrequently observed in experiment. Here, we investigate this problem in the $S = 1/2$ anisotropic kagome antiferromagnet Y-kapellasite, Y$_3$Cu$_9$(OH)$_{19}$Cl$_8$, using pulsed-field magnetization measurements on single
crystals and high-field $^{35}$Cl NMR. We identify a hierarchy of field-induced fractional features, including $1/3$ and $1/9$ plateaus, as well as a weaker low-field feature. Analysis of the NMR spectra and the
magnetic susceptibility across the $1/9$ plateau demonstrate that it is accompanied by an ordered local
spin configuration, a strong suppression of low-energy
spin fluctuations and activated behavior, consistent with a gapped fractional state. These features differ from those in the only other material YCu$_3$(OH)$_6$Br$_2$[Br$_{1-y}$(OH)$_y$] in which this plateau is observed, implying a surprising robustness of the $1/9$ state to the details of the underlying magnetism.
Read the full abstract ▾
All authors Dipranjan Chatterjee, Paul A. Goddard, Ewan R. P. Thomas, Katharina M. Zoch, Hank C. H. Wu, Benjamin M. Huddart, Cornelius Krellner, Edwin Kermarrec, Mladen Horvatić, Steffen Krämer, Pascal Puphal, John Singleton, Stephen J. Blundell, Fabrice Bert
Materials Science
Anisotropic Defect Diffusion in Layered CsPbBr$_\mathrm{x}$I$_\mathrm{3-x}$ Perovskites
Konrad Wilke, Mike Pols, Titus S. van Erp et al. (+2 more)
Mixed-halide perovskites offer a route to enhance phase stability and modify optoelectronic properties. Here, we use large-scale molecular dynamics simulations with a reactive force field to investigate defects in CsPbBr$_\mathrm{x}$I$_\mathrm{3-x}$ perovskites, focusing on how defect mobility can be controlled and the stability of the material improved by layered ordering of Br and I anions. Our results show that layered halide ordering induces strongly anisotropic defect
diffusion: migration proceeds readily along the layers, whereas
diffusion across them is strongly suppressed. For Cs defects, this anisotropy originates from directional
lattice strain and the associated octahedral tilting, whereas halide migration is governed by an interplay between strain and preferential local halide bonding configurations.
Read the full abstract ▾
All authors Konrad Wilke, Mike Pols, Titus S. van Erp, Geert Brocks, Shuxia Tao
Strongly Correlated Electrons
Light-driven octupolar inverse Faraday effect and multipolar order in Mott insulators
Saikat Banerjee, Tara Steinhöfel, Florian Lange et al. (+2 more)
Hidden multipolar orders in
spin-orbit-coupled Mott
insulators provide a promising setting for correlated
quantum matter, yet their control and detection remain major challenges. Here, we demonstrate that circularly polarized light enables both the control and detection of hidden multipolar order in $4d^2/5d^2$ systems with edge-sharing octahedra. Using a Floquet Schrieffer-Wolff expansion of a driven Hubbard-Kanamori model, we derive a low-energy multipolar
Hamiltonian with two qualitatively new light-driven terms. One is an effective static field that couples linearly to the
magnetic octupole, realizing an octupolar inverse Faraday effect. The other is a bond-dependent anisotropic exchange interaction absent in
equilibrium. These two couplings constitute the key mechanisms of this work: the first provides a direct optical handle on hidden octupolar order, while the second reorganizes the multipolar exchange landscape and introduces strong bond-dependent frustration. Together, these couplings define a nonequilibrium multipolar
Hamiltonian whose phase diagram contains robust antiferro-octupolar, ferro-octupolar, and partially polarized ferro-quadrupolar phases, alongside a putative Ising-octupolar regime and a strongly frustrated,
liquid-like multipolar regime whose finite-size numerical signatures are compatible with Kitaev-type behavior. We further show that the induced multipolar order couples to the
lattice, generating reversible trigonal and orthorhombic or tetragonal distortions that provide
symmetry-resolved structural fingerprints in pump-probe experiments. In this way, our work establishes a general mechanism for the optical generation, control, and detection of hidden multipolar
quantum states.
Read the full abstract ▾
All authors Saikat Banerjee, Tara Steinhöfel, Florian Lange, Matthias Eschrig, Holger Fehske
Materials Science · Cross-listed from physics.optics
Hot biexcitons driven by extreme optical confinement
Xinyi Wang, Kaushik Kudtarkar, Wenjing Wu et al. (+6 more)
A powerful means to understanding condensed matter that possesses a multi-constituent, non-isolated, and complex nature, with a preeminent example being two-dimensional (2D) materials, is studying many-body interactions. However, experimentally observing high-order many-body interactions is a daunting task due to its heavy reliance on the abundance of low-order complexes. Here, we report the observation of four-body hot biexcitons in an energetically unfavorable
bilayer of tungsten disulfide (WS2) through creating extreme optical confinement. Specifically, we integrate a non-radiative bound state in the continuum (BIC) into a photonic
crystal (PhC) defect cavity, forming a quasi-three-dimensional (q-3D) but open confinement for
photons at the driving frequency. The extremely confined
photons in both reciprocal and physical spaces then excite inherently unproductive two-body hot
excitons situated slightly above the indirect bandgap so efficiently that they form overwhelmed higher-order four-body hot biexcitons. Distinctively, these hot biexcitons exhibit substantial valley polarization and coherence at room temperature, which we attribute to the topological nature of BICs and the associated q-3D confinement with an orbital angular momentum. Besides achieving room-temperature biexcitons, the q-3D confinement could be valuable for higher-order interactions, such as triexcitons, and many other many-body phenomena, including Bose-Einstein condensation.
Read the full abstract ▾
All authors Xinyi Wang, Kaushik Kudtarkar, Wenjing Wu, Yunjo Jeong, Yuxuan Cosmi Lin, Xiaofeng Qian, Junichiro Kono, Shengxi Huang, Shoufeng Lan
Strongly Correlated Electrons · Cross-listed from quant-ph
Analytical Solution to the Kronig-Penney Model with Harmonic Oscillator Wells: Insights to Tight-Binding
Christopher Moore and Frank Marsiglio
The celebrated Kronig-Penney model traditionally has been formulated with square well potentials representing atomic centres. Here, we use a slightly more realistic potential, the truncated harmonic oscillator, in lieu of square well potentials, and solve the model analytically. We derive the energy dispersion and wave functions for this model. This configuration has some important similarities and differences compared to the usual model. In particular, we write the governing equation in a form suggestive of the tight-binding approximation, as can be done for the usual model. In this way, it is straightforward to derive an expression for the tunneling amplitude used in tight-binding in terms of the harmonic oscillator potential parameters.
Read the full abstract ▾
Statistical Mechanics · Cross-listed from cs.LG
Susceptibilities and Patterning: A Primer on Linear Response in Bayesian Learning
Chris Elliott and Daniel Murfet
These notes introduce the theory of susceptibilities as developed in [arXiv:2504.18274, arXiv:2601.12703] for interpreting neural networks. The susceptibility of an observable $\phi$ to a data perturbation is defined as a derivative of a posterior expectation, which by the fluctuation–dissipation
theorem equals a posterior covariance. Different choices of $\phi$ yield different objects: per-sample losses give the influence
matrix (the
Bayesian influence function of [arXiv:2509.26544]), while component-localized observables give the structural susceptibility
matrix that pairs model components with data patterns. The susceptibility
matrix is (up to a factor of $n\beta$) the Jacobian of the map from data distributions to structural coordinates; its pseudo-inverse provides a linearized solution to the patterning problem of [arXiv:2601.13548]: finding data perturbations that produce a desired structural change. We motivate the theory from its statistical-mechanical foundations, then give a detailed exposition of susceptibilities, their empirical
estimators, and their connection to the geometry of the loss landscape.
Read the full abstract ▾
Superconductivity · Cross-listed from physics.acc-ph
Multilayer model for coatings with arbitrary layers for superconducting radio-frequency applications
Aaron Gobeyn, Wolfgang Ackermann and Herbert De Gersem
We extend the multilayer model of \etal{Kubo} for
superconductor-
insulator-
superconductor (SIS) structures in two ways: first, by generalizing it to arbitrary sequences of layers of arbitrary type, i.e. superconducting, normal conducting, and insulating; and second, by accounting for all contributions, including ohmic losses and dielectric effects. We examine the maximum applicable field for $(\text{SI})^n\text{S}$ structures. We find that the optimum configuration corresponds to the $n=1$ case. However, the thickness of the superconducting coating layers can be reduced to below their penetration depth with minor performance penalty. We discuss the ability to model transitions in SS
bilayers by introducing a set of virtual layers that represent the transition region through interpolated parameters. We find degradation of the maximum applicable field with thicker transition layers, and a larger effective penetration depth of the electromagnetic fields. Furthermore, the surface impedance of the multilayer structure is calculated using the Leontovich boundary condition, yielding a formulation suitable for integration into finite element simulations. Additionally, the Poynting
theorem is used to determine the loss contributions of individual layers.
Read the full abstract ▾
Disordered Systems and Neural Networks · Cross-listed from quant-ph
Reducibility of native weighted graphs on Rydberg Arrays
J. Kombe and J. D. Pritchard
We investigate the classical reducibility of random unit-disk graph (UDG) instances of the maximum independent set (MIS) and maximum weighted independent set (MWIS) problems, which can be natively realised in Rydberg atom
quantum processors. Using state-of-the-art kernelisation techniques, we systematically probe how far classical preprocessing can simplify such native optimisation problems of varying size and connectivity. While many small or sparse instances can be fully reduced, dense graphs often retain finite irreducible kernels even after extensive reductions. Introducing vertex weights tends to increase reducibility, whereas extending the interaction range in the underlying UDG connectivity suppresses the reduction efficiency. By exploring where classical reductions cease to be effective, we aim to delineate the regime of problem instances that remain computationally demanding - those most relevant for testing and benchmarking near-term
quantum optimisation hardware. We find that for the remaining finite kernels,
quantum execution would require non-native
embeddings with substantial resource overheads, suggesting that directly running native instances may be more practical than
embedding a reduced kernel.
Read the full abstract ▾
Statistical Mechanics
Mutual Linearity in Nonequilibrium Langevin Dynamics
Jiming Zheng and Zhiyue Lu
Understanding how nonequilibrium systems respond to perturbations is a central challenge in physics. In this work, we establish mutual linearity in nonequilibrium overdamped Langevin systems. This theory provides a framework for controlling and designing nonequilibrium responses in continuous systems. When a dynamical parameter is locally perturbed at a single position, the stationary densities at any two positions are linearly related. It further leads to mutual linearity among different stationary state-current observables. We also extend the mutual linearity to non-stationary relaxation processes in the Laplace domain. Our theory reveals that mutual linearity in both discrete and continuous systems originates from the same one-dimensional response structure. We further show that mutual linearity is robust under finite-width perturbations. As an application, we demonstrate the mutual linearity and its finite-width robustness in the F$_1$-ATPase rotary motor model.
Read the full abstract ▾
Mesoscale and Nanoscale Physics
Energy-Resolved Quantum Geometry from Středa Response: Driven-Dissipative Bosonic Lattices and Disordered Systems
Anaïs Defossez, Baptiste Bermond, Lucila Peralta Gavensky et al. (+1 more)
The Středa formula links the
Hall conductivity of an
insulator to the
magnetic-field response of its particle density, providing a local and universal probe of the topological Chern number. Beyond this quantized response, an energy-resolved Středa marker can be defined from the
magnetic response of the density of states, revealing detailed features of the
quantum geometry of Bloch bands. We show that driven-dissipative bosonic
lattices provide direct access to both the integrated and energy-resolved Středa responses. Our scheme uses controlled pumping with uniform strength and random phases across the
lattice, together with uniform loss, to yield a Lorentzian filter of eigenmode occupations. For generic dispersive bands, this enables reconstruction of a coarse-grained energy-resolved Středa response, establishing these platforms as versatile probes of anomalous spectral flow and energy-resolved
quantum geometry. As a striking application, we show that this marker elucidates the fate of topological bands under strong disorder, capturing the
quantum-
geometric structure underlying topological Anderson
insulators.
Read the full abstract ▾
All authors Anaïs Defossez, Baptiste Bermond, Lucila Peralta Gavensky, Nathan Goldman
Strongly Correlated Electrons
Microscopic Magnetism of A(TiO)Cu4(PO4)4 (A = Ba, Pb, Sr): 31P and 63,65Cu NMR Study
Riho Rästa, Ivo Heinmaa, Joosep Link et al. (+5 more)
We report a comprehensive NMR study of the
chiral square-cupola antiferromagnet Pb(TiO)Cu$_4$(PO$_4$)$_4$ and compare its microscopic hyperfine and local-field parameters with the Ba/Sr analogues in the $A$(TiO)Cu$_4$(PO$_4$)$_4$ family. Above $T_{\rm N}\simeq 6.7$ K, the $^{31}$P Knight shift tracks the bulk susceptibility and yields nearly isotropic transferred hyperfine couplings $H_{\rm hf}^{[010]}=6.77(3)$ and $H_{\rm hf}^{[001]}=6.19(3)$ kOe/$\mu_{\rm B}$. Below $T_{\rm N}$, the frequency-swept $^{31}$P spectrum splits into three lines, in contrast to the four-line pattern reported for BaTCPO. The line separation tracks the onset of the static $^{31}$P internal field with a power-law exponent $\beta\simeq 0.23$, consistent with quasi-two-dimensional criticality.
Crystal-rotation $^{31}$P NMR in the ordered state resolves all eight
symmetry-related P sites and their site-dependent anisotropy. In the ordered state, zero-field $^{63,65}$Cu NMR gives a Cu-site internal field $B_{\rm int}=14.50(6)$ T and a quadrupole frequency $\nu_Q=32.72(5)$ MHz, while point-charge electric-field-gradient calculations including Sternheimer corrections yield an on-site Cu hole occupancy $n_d=0.20(4)$, consistent with a ligand-hole-dominated charge-transfer character. Comparing PbTCPO with BaTCPO and SrTCPO, we find that the transferred hyperfine coupling $H_{\rm hf}$ varies across the series, reflecting changes in local Cu-O-P covalency, whereas the ordered-state $^{31}$P internal field in PbTCPO is $69.5$ mT, considerably higher than in BaTCPO ($35.6$ mT) and SrTCPO ($34.6$ mT). This enhancement is not captured by dipolar terms alone and points to the combined effects of transferred contributions and stacking-dependent cancellation.
Read the full abstract ▾
All authors Riho Rästa, Ivo Heinmaa, Joosep Link, Yusuke Kousaka, Tsuyoshi Kimura, Yoshihiko Ihara, Kenta Kimura, Raivo Stern
Materials Science
MatterSim-MT: A multi-task foundation model for in silico materials characterization
Han Yang, Xixian Liu, Chenxi Hu et al. (+25 more)
Accurate property characterization is a major bottleneck in materials design. While first-principles methods and task-specific machine-learning models have driven important progress, they remain fundamentally limited in scalability and generalizability across the vast space of structures and properties relevant to real-world materials design. We present MatterSim-MT, a multi-task foundation model for in silico materials simulation and property characterization. The model is pretrained on over 35 million first-principles-labeled structures covering 89 elements, temperatures up to 5000 K and pressures up to 1000 GPa, and is fine-tuned on various properties including Bader charges,
magnetic moments, Born effective charges, and dielectric
matrices. Out of the box, MatterSim-MT not only serves as a foundation model for predicting material structure, dynamics and thermodynamics, its multi-task architecture also enables a wide range of complex simulations that cannot be captured by potential energy surfaces alone. For example, we demonstrate pressure-dependent LO-TO
phonon splitting in SiC with close agreement with experiment, electric hysteresis in ferroelectric BaTiO3, and the cationic-to-anionic redox transition during delithiation of a Li-rich cathode material. Finally, we show that MatterSim-MT scales well with more data and parameters, can be efficiently fine-tuned to higher levels of theory, and can be efficiently extended to new systems via active learning. Overall, we believe this approach provides a scalable route to accurate in silico materials characterization.
Read the full abstract ▾
All authors Han Yang, Xixian Liu, Chenxi Hu, Yichi Zhou, Yu Shi, Chang Liu, Junfu Tan, Jielan Li, Guanzhi Li, Qian Wang, Yu Zhu, Zekun Chen, Shuizhou Chen, Fabian Thiemann, Claudio Zeni, Matthew Horton, Robert Pinsler, Andrew Fowler, Daniel Zügner, Tian Xie, Lixin Sun, Yicheng Chen, Lingyu Kong, Yeqi Bai, Deniz Gunceler, Frank Noé, Hongxia Hao, Ziheng Lu
Materials Science
Interfacial control of hot-carrier extraction and photostability in two-dimensional materials
Claudia Gollner, Mohammad Taghinejad, Chenyi Xia et al. (+8 more)
Two-dimensional transition metal dichalcogenides (TMDCs) are promising materials for next-generation optoelectronic devices, yet their implementation is hindered by limited sample stability and challenges in forming reliable electrical contacts. Here, by utilizing time-domain THz emission spectroscopy we directly probe charge carrier dynamics in monolayer WS2 on gold (Au) and fused silica (SiO2) as a function of interface morphology. For
laser excitation above the band gap of WS2, we independently extract effective transport times for both electrons and holes and find that discontinuous WS2 contacts on rough Au generate larger net photocurrents than uniform, strongly coupled interfaces - a counterintuitive observation attributed to imbalanced electron and hole transfer from WS2 to Au. Crucially, we demonstrate that ultrafast charge extraction and separation suppress recombination-driven energy release and thereby prevent photo-induced degradation under ambient conditions, eliminating the need for encapsulation. These findings redefine interfacial design as a central control parameter for both performance and stability in 2D optoelectronic devices.
Read the full abstract ▾
All authors Claudia Gollner, Mohammad Taghinejad, Chenyi Xia, Zhepeng Zhang, Fang Liu, Francesco Laudani, Annette Foelske, Mark L. Brongersma, Andrew J. Mannix, Tony F. Heinz, Aaron Lindenberg
Other Condensed Matter
Checkerboard Bose Hubbard Ladders using Transmon Arrays
Pranjal Praneel, Thomas G Kiely, Andre G Petukhov et al. (+1 more)
Adding a sublattice bias to the two dimensional Bose Hubbard model greatly enriches the available physics, and introduces knobs which can be used to control and interrogate the
quantum state. We describe the physics of this checkerboard Bose Hubbard model and how it can be explored using transmon arrays. We show that the sublattice bias brings the commensurate superfluid phase into an experimentally accessible regime, and gives new probes. We characterize the superfluid and insulating phases, with careful
attention to finite size effects.
Read the full abstract ▾
All authors Pranjal Praneel, Thomas G Kiely, Andre G Petukhov, Erich J Mueller
Materials Science
Anomalous magnetotransport in a non-collinear correlated kagome ferromagnet MgMn6Sn6
Kakan Deb, Sourav Kanthal, Jyotirmoy Sau et al. (+5 more)
Magnetic kagome metals provide a fertile platform for exploring unusual magnetotransport phenomena arising from the intricate interplay between electronic
topology, electron correlations, and
magnetic order. MgMn6Sn6 is a room-temperature kagome ferromagnet with strong in-plane
magnetic anisotropy. Here, we report a combined study of single-
crystal neutron diffraction (SCND) and magnetotransport properties of MgMn6Sn6, supported by first-principles calculations. Our SCND measurements reveal a non-collinear arrangement of Mn
magnetic moments within the basal plane of the kagome
bilayer. The
Hall conductivity shows a substantial intrinsic contribution of approximately 0.29 e^2/h per kagome layer, which is nearly isotropic with respect to the field orientation. At low temperatures, the anomalous
Hall conductivity develops a pronounced anisotropic extrinsic component, highlighting the directional sensitivity of scattering processes. The significantly large value of the Sommerfeld coefficient, in the absence of f-electrons, underscores enhanced electron correlation. Therefore, the non-collinear kagome ferromagnet MgMn6Sn6 is a promising candidate for studying the effects of electron correlation on magnetotransport properties.
Read the full abstract ▾
All authors Kakan Deb, Sourav Kanthal, Jyotirmoy Sau, Chandra Shekhar, Manoranjan Kumar, Matthias Gutmann, Jhuma Sannigrahi, Nitesh Kumar
Strongly Correlated Electrons
Shuttling of $\mathbb{Z}_4$ parafermions in an electronic ladder model
Botond Osváth, Gergely Barcza and László Oroszlány
Parafermions with non-Abelian statistics have been proposed as a promising platform for
quantum computation, potentially enabling a broader set of topologically protected gates than Majorana
fermions. The experimental and theoretical exploration of these exotic quasiparticles remains challenging, as their stability is linked to strong electron-electron interactions. A key step toward practical applications is the controlled shuttling of parafermionic modes, which is required for implementing
geometric braiding operations. In the present work, we investigate the real-time dynamics of the elementary shuttling process by applying a combination of the density
matrix renormalization group and the time-dependent variational principle approaches. We analyze the transport of $\mathbb{Z}_4$ parafermion edge states and assess the corresponding adiabatic speed limit under experimentally relevant conditions.
Read the full abstract ▾
Other Condensed Matter · Cross-listed from cs.IT
Physics-Inspired Probabilistic Computing for Extremely Large-Scale MIMO Detection in Future 6G Wireless Systems
Andrea Grimaldi, Christian Duffee, Eleonora Raimondo et al. (+8 more)
Extremely large-scale multiple-input multiple-output (XL-
MIMO) architectures are a key enabler of forthcoming 6G wireless communication networks by allowing high data rates through massive spatial multiplexing. Here, we approach these problems with physics-inspired unconventional computing based on Ising machines (IMs). For binary modulation, probabilistic IMs (PIMs) and oscillator-based IMs achieve optimal ML detection with systems up to 2048x2048
antennas with only 100 iterations, matching optimal sphere decoder performance for computationally treatable sizes and outperforming the minimum mean-square error (MMSE) industrial standard. For M-QAM up to 256, a generalized PIM-inspired framework, based on d-dimensional probabilistic variables (p-dits) that directly encode QAM symbols, shows low bit-error-rate across sizes up to 256x256
antennas, outperforming or matching MMSE with reduced algorithmic complexity. Unlike the binary mapping, the p-dit interaction
matrix is independent of the QAM order, enabling adaptive
MIMO modulation. These results show a promising scalable paradigm for XL
MIMO detection in future 6G networks.
Read the full abstract ▾
All authors Andrea Grimaldi, Christian Duffee, Eleonora Raimondo, Edoardo Piccolo, Deborah Volpe, Filip B. Maciejewski, Mario Carpentieri, Massimo Chiappini, Pedram Khalili Amiri, Davide Venturelli, Giovanni Finocchio
Other Condensed Matter
Bulk-mediated reflection of chirality-protected surface spin waves
Vitaliy I. Vasyuchka, Florin Ciubotaru, Andrii V. Chumak et al. (+2 more)
Surface
spin waves of the Damon-Eshbach type exhibit intrinsically nonreciprocal transport properties with a
chiral dynamical field structure that localizes counterpropagating waves at opposite film surfaces. Such chirality has been predicted to suppress direct backscattering in thin films within frequency ranges free of bulk modes. However, how chirality influences reflection in thicker three-dimensional
magnetic media, where a dense spectrum of bulk excitations overlaps with surface waves, remains unclear. Here we demonstrate that, in micrometer-thick yttrium iron garnet films, reflection of the
chiral Damon-Eshbach wave from the boundary of the
magnetic medium is accompanied by excitation of spatially localized thickness-quantized bulk modes, whereas reciprocal backward-volume waves reflect nearly elastically. Brillouin light scattering spectroscopy, infrared thermography, and micromagnetic simulations reveal standing bulk excitations at the reflecting boundary and quantify the associated
magnon energy accumulation and dissipation. These results identify bulk-mode excitations as the physical pathway enabling reversal of chirally localized surface waves in thick films, thereby defining the limits of chirality-based backscattering immunity and providing a general framework for wave transport in nonreciprocal
magnetic media.
Read the full abstract ▾
All authors Vitaliy I. Vasyuchka, Florin Ciubotaru, Andrii V. Chumak, Burkard Hillebrands, Alexander A. Serga
Disordered Systems and Neural Networks
Spectral Dynamics in Deep Networks: Feature Learning, Outlier Escape, and Learning Rate Transfer
Clarissa Lauditi, Cengiz Pehlevan and Blake Bordelon
We study the
evolution of hidden-weight spectra in wide neural networks trained by (
stochastic) gradient descent. We develop a two-level dynamical mean-field theory (DMFT) that jointly tracks bulk and outlier spectral dynamics for spiked ensembles whose spike directions remain statistically dependent on the random bulk. We apply this framework to two settings: (1) infinite-width nonlinear networks in mean-field/$\mu$P scaling and (2) deep linear networks in the proportional high-dimensional limit, where width, input dimension, and sample size diverge with fixed ratios. Our theory predicts how outliers evolve with training time, width, output scale, and initialization variance. In deep linear networks, $\mu$P yields width-consistent outlier dynamics and hyperparameter transfer, including width-stable growth of the leading NTK mode toward the edge of stability (EoS). In contrast, NTK parameterization exhibits strongly width-dependent outlier dynamics, despite converging to a stable large-width limit. We show that this bulk+outlier picture is descriptive of simple tasks with small output channels, but that tasks involving large numbers of outputs (ImageNet classification or GPT
language modeling) are better described by a restructuring of the spectral bulk. We develop a toy model with extensive output channels that recapitulates this phenomenon and show that edge of the spectrum still converges for sufficiently wide networks.
Read the full abstract ▾
Statistical Mechanics
Cluster Dynamics Stay Fast-Until Tricriticality
Minjun Jeon, Alexandros Vasilopoulos, Dong-Hee Kim et al. (+2 more)
Cluster Monte Carlo
algorithms are widely regarded as the most effective route to overcoming critical slowing down in
lattice spin systems. Whether this acceleration persists in the presence of vacancies and multicritical fluctuations, however, remains unresolved. We address this question through a systematic dynamic-scaling study of hybrid cluster-local update schemes in the two-dimensional Blume-Capel model, which exhibits a line of continuous Ising-like transitions terminating at a tricritical point. Along the entire critical line, hybrid dynamics retain the near-optimal efficiency of pure cluster updates despite the presence of annealed vacancies. Strikingly, this acceleration collapses precisely at tricriticality, where the dynamic critical exponent reverts to the local-update value. We trace this breakdown to the correlated percolation of vacancies, whose emergent system-spanning geometry obstructs nonlocal relaxation in the
spin sector. Our results identify a fundamental
geometric limitation of cluster acceleration at tricriticality and establish vacancy percolation as the mechanism controlling dynamic universality in hybrid Monte Carlo dynamics.
Read the full abstract ▾
All authors Minjun Jeon, Alexandros Vasilopoulos, Dong-Hee Kim, Víctor Martín-Mayor, Nikolaos G. Fytas
Strongly Correlated Electrons
Electronic excitations in the Shastry-Sutherland compound SrCu$_2$(BO$_3$)$_2$
Tariq Leinen, Ola K. Forslund, Eugenio Paris et al. (+13 more)
SrCu2(BO3)2 (SCBO) is a paradigmatic realization of the Shastry-Sutherland model, hosting geometrically frustrated
spin dimers and a variety of
quantum magnetic phases and phenomena. Although its
magnetic properties have been extensively studied, the high-energy electronic excitations that determine the
crystal-field environment and Cu-O hybridization have remained largely unexplored. Here we combine Cu L3-edge resonant inelastic x-ray scattering (RIXS), broadband optical spectroscopy, and electronic-structure calculations to determine the relevant local and interband excitation energy scales in SCBO. RIXS resolves a well-defined
manifold of localized Cu2+ d-d excitations between 1.8 and 2.4 eV, whose energies and polarization dependence are well reproduced by multireference
quantum-chemistry calculations. In contrast, optical spectroscopy identifies charge-transfer excitations with an absorption onset near 1.2-1.6 eV and a broader higher-energy structure around 4.5 eV, which are qualitatively captured by DFT+U calculations. Taken together, these results define the characteristic energy scales of d-d and CT excitations, offering quantitative
benchmarks for computational frameworks and providing essential input for refining superexchange-based
magnetic models of this prototypical frustrated
quantum antiferromagnet.
Read the full abstract ▾
All authors Tariq Leinen, Ola K. Forslund, Eugenio Paris, Nicola Colonna, Marco Caputo, Johan Chang, Gabriel Nagamine, Takashi Tokushima, Conny Såthe, Pascal Puphal, Jeremie Teyssier, Thorsten Schmitt, Nikolay A. Bogdanov, Maria Daghofer, Adrian L. Cavalieri, Flavio Giorgianni
Mesoscale and Nanoscale Physics
Finite-q photon-drag shift current in two-dimensional massive chiral Dirac fermions
Rofii and Eddwi Hasdeo
We investigate the
photon-drag shift current in an isotropic single-valley two-dimensional massive
chiral Dirac model with chirality index $J=1,2,3$ by directly evaluating the full finite-$q$ non-vertical response beyond the perturbative small-$q$ regime. Our central result is that chirality qualitatively reorganizes the sign
topology of the finite-$q$ photocurrent $\mathbf{ j}(\mathbf{ q})$. For $J=1$, the photocurrent remains broadly positive, whereas higher-chirality sectors ($J \ge 2$) generically develop internal zero-current contours and sign reversals within the kinematically allowed region. We further show that the photocurrent is
symmetry-constrained to be purely transverse, $\mathbf{j}(\mathbf{q}) \propto \hat{\mathbf{z}}\times\mathbf{q}$, and vanishes in the strictly vertical-transition limit $q=0$ in centrosymmetric systems. Pauli blocking further shapes the response by selecting the active portion of the resonance contour, while its extinction at large $\Delta$ or $q$ follows from a simple kinematic cutoff. These results establish the isotropic massive
chiral Dirac problem as a
symmetry-controlled
benchmark for chirality-dependent finite-$q$ shift currents.
Read the full abstract ▾
Statistical Mechanics
Tamed Feynman-Kac diffusion processes: Killing-branching intertwine
Piotr Garbaczewski and Mariusz Żaba
Relaxation to
equilibrium of a drifted Brownian motion is quantified by a transition probability density function, whose main (multiplicative) entry is an inferred Feynman-Kac kernel of the Schrödinger
semigroup operator. Although seemingly devoid of a natural probabilistic significance (except for its explicit path integral definition), the pertinent kernel relaxes to
equilibrium as well. The implicit Feynman-Kac potential ${\cal{V}}(x)$, continuous, confining and bounded from below, may take negative values. If positive, ${\cal{V}}(x)$ can be interpreted as the killing rate of the decaying
diffusion process. In case of relaxing F-K kernels the killing effects are tamed (often overcompensated). The taming inavoidably appears in conjunction with the existence of the negativity subdomains of ${\cal{V}}(x)$ in $R$. If locally ${\cal{V}}(x) < 0$, its sign inversion $- {\cal{V}}(x)$ can be interpreted as the branching (cloning, alternatively bifurcation) rate in the course of the other wise free random motion. The arising killed
diffusion processes with branching, we interpret as the possible path-wise background of tamed (relaxing) Feynman-Kac
diffusions. We present acomputer-assisted path-wise arguments, towards a consistency of the killing/branching taming scenario, for a number of nonlinear model systems in one space dimension. Special
attention is paid to Feynman-Kac potential shapes in the double well form, where an analytic access to
eigenvalues and eigenfunctions is scarce. Throughout the paper the dynamics refers to the positive real time. Since the Newton-type equations of motion for admissible classical trajectories have a Euclidean form (due to the sign inverted force term), we give a brief resume of a couple of their explicit solutions, without recourse to the Euclidean time intuitions, and the instanton lore of related
quantum model systems.
Read the full abstract ▾