74 papers
This paper investigates the Anderson localization of the disordered Hatano-Nelson chain, revealing a critical disorder threshold where the complex eigenvalue spectrum bifurcates and the spectral winding number transitions from 1 to 0, accompanied by a shift from exponentially localized states to two completely delocalized states at weak and critical disorder.
This paper formalizes the loss of diversity in classifier-free guidance as "generative distortion," characterizes its emergence via a high-dimensional phase transition using statistical physics tools, and proposes a novel guidance schedule with a negative-guidance window to mitigate variance shrinkage while preserving class separability.
This paper investigates an open Sachdev-Ye-Kitaev model coupled to a pseudogapped fermionic bath using the Keldysh formalism, revealing a rich dynamical phase diagram where non-Markovian dissipation competes with internal chaos to produce distinct regimes of power-law and exponential relaxation, as well as an intermediate crossover phase.
This paper introduces DOTA, a deep learning model that leverages local density of states to capture orbital interaction patterns, enabling accurate and efficient prediction of surface adsorption energies by aligning scarce experimental data with multi-fidelity quantum chemistry calculations.
This paper demonstrates that violations of random matrix theory predictions for local observables, specifically regarding quantum Fisher information dynamics and fluctuation-dissipation relations, can serve as effective witnesses for detecting ergodicity-breaking transitions in quantum many-body systems across integrability, many-body localization, and quantum many-body scars.
This paper demonstrates that Restricted Boltzmann Machines can effectively model large-scale neural activity from approximately 1,500 to 2,000 simultaneously recorded neurons, capturing complex higher-order statistics and revealing anatomically structured interaction networks that align with visual behavior and global dynamics.
This paper presents a statistical-mechanical theory linking the linear decoding efficiency of continuous variables to the geometric properties of neural manifolds, revealing increasing decoding capacity for object position and size along the monkey visual stream.
Motivated by recent twisted MoTe experiments, this paper develops a disordered interacting edge theory for a fractional topological insulator, revealing distinct conductance phases and an interaction-induced insulating state that demonstrates the insufficiency of two-terminal transport measurements for identifying such systems.
This paper establishes a continuum field theory description for random ensembles of two-dimensional fermionic matchgate tensor networks, demonstrating that their disorder-averaged physics corresponds to a nonlinear sigma-model of symmetry class D with a topological term, thereby linking these discrete networks to the thermal quantum Hall problem and revealing a phase structure that includes localized phases, quantum Hall criticality, and a thermal metal.
This paper establishes a large-deviation principle for the maximal energy of a spin glass with spins by deriving a Parisi-type formula for fractional moments and leveraging convex duality to show that the rate function is asymptotically quadratic near its minimum if and only if an external magnetic field is present.
This paper provides a comprehensive review and formal specification of Predictive Coding Networks (PCNs), highlighting their biological plausibility, computational efficiency through inference learning, and versatility as a probabilistic framework that extends beyond traditional backpropagation-based neural networks.
This paper demonstrates that predictive coding graphs constitute a mathematical superset of feedforward neural networks, thereby strengthening their theoretical foundation in machine learning and highlighting the importance of network topology.
This paper presents and experimentally demonstrates a purely photonic deep neuromorphic network that achieves 100% accuracy on a letter recognition task by utilizing a local optical feedback mechanism with non-volatile phase-change material synapses to enable online, unsupervised Hebbian learning without inefficient optical-electrical-optical conversions.
This paper develops an analytic approach linking the density of complex resonance poles to the distribution of reflection coefficients at complex energies, yielding explicit formulas for the crossover from narrow to broad resonances in both semi-infinite and short one-dimensional disordered samples, and validating these results against numerical simulations of the Anderson tight-binding model.
This study employs neural-network quantum states and Schrieffer-Wolff transformations to demonstrate that isotropic antiferromagnetic Heisenberg perturbations drive the toric code through a quantum phase transition into a fourfold-degenerate Néel phase, characterized by a critical point identified via fidelity and entanglement diagnostics.
This paper establishes that the Petz and Schumacher-Westmoreland recovery schemes achieve the true optimal threshold for quantum error correction and introduces the mutual trace distance as a necessary and sufficient diagnostic to determine this threshold without explicit optimization.
This paper presents a landscape-agnostic study of rare dynamical events in mean-field disordered systems that reveals a rich diversity of instanton structures beyond classical nucleation theory, thereby identifying the point of irreversibility and clarifying the landscape features governing activated relaxation in the RFOT universality class.
This paper establishes a unified theoretical framework based on spinful quasiperiodic systems that achieves exact solvability for all seven fundamental localization phases, including novel models with mobility edges, and proposes feasible experimental schemes for their realization.
This study uses Monte Carlo simulations to demonstrate that quasi-three-dimensional stick percolation systems, despite having a higher percolation threshold than their two-dimensional counterparts, share the same universal finite-size scaling function as 2D continuum and lattice percolation.
This paper investigates the robustness of topological strong zero modes in random Ising-Majorana chains using SZM fidelity as a diagnostic, revealing that while these modes persist throughout the topological phase, the infinite-randomness critical point exhibits distinctive, ensemble-dependent fidelity distributions that suggest an intrinsically stronger topological character and a boundary manifestation of average Kramers-Wannier duality.