Searcharxiv⌕ Search

arXiv subjects

Di Fang

Publications and source records attributed to Di Fang.

At least 19 recordsLinked to original sources

Quantum Gibbs sampling through the detectability lemma

Gibbs state preparation is an important subroutine in quantum computing. In this work we use the detectability lemma to improve Gibbs state preparation. Specifically, we design new Gibbs state preparation methods that do not rely on simulating Lindbladian evolution, thus avoiding the overhead from it. For local Lindbladians consisting of $M$ terms, this approach reduces the cost by a factor of $O(M)$. We also combine the detectability lemma operator and quantum singular value transformation to implement ground state projection operators of frustration-free Hamiltonians, resulting in a quadratic speedup in the spectral gap dependence. Applying this method to Lindbladians for the Gibbs state of local commuting Hamiltonians, we achieve quadratically better dependence on the Lindbladian spectral gap.

quant-ph↗

Semiclassical Scaling Laws of Magnus Quantum Algorithms under Harmonic Confinement

The simulation of semiclassical Schrödinger dynamics is often hindered by rapid oscillations, where classical numerical methods and quantum simulation algorithms can suffer severe precision loss as the semiclassical parameter $\varepsilon \to 0$. In this work, we uncover a surprising robustness of high-order Magnus quantum algorithms in the semiclassical regime. We establish two important results. First, we extend the theory of Magnus superconvergence from free-particle reference dynamics to a general class of Schrödinger operators whose reference Hamiltonian includes harmonic confinement. Second, and more importantly, we reveal a nontrivial semiclassical scaling law: while the expansion remainder term bound suggests an error prefactor of $O(\varepsilon^{-(p+1)})$ for a $p$th-order method, we prove that the loss is in fact order-independent and collapses to $O(\varepsilon^{-1})$ for all $p$. The key analytical mechanism is the exact Egorov transport associated with quadratic flows in semiclassical Weyl quantization, which enables a symbol-level analysis of the interaction-picture Hamiltonian and its Magnus commutators. Our results show that high-order Magnus quantum algorithms remain far more robust in the semiclassical regime than naive series remainder term counting would suggest.

math.NA↗

Spatial Discretization Error for Quantum Chemistry Simulation

Nonrelativistic quantum chemistry is based on the Schrödinger equation in continuous, unbounded space. Computational approaches on classical and quantum computers often truncate the spatial domain and introduce a finite-dimensional basis, reducing the continuum problem to a finite-dimensional one. There is therefore a fundamental question: to what extent does this approximation capture the original continuum problem? In this work, we develop a general Sobolev-Galerkin framework for estimating spatial discretization errors in quantum simulation with unbounded Hamiltonians. We apply this framework to many-body Coulomb systems and establish both periodization and Galerkin error estimates for wave-function dynamics and ground-state energies. For dynamics, the finite-basis projection error of the exact wave function does not capture the full basis-discretization error: an additional dynamical projection error enters our bound at a different rate, and we introduce an inner-product argument to control this contribution. Interestingly, the spatial convergence rates for wave-function dynamics and ground-state energies differ quantitatively. For ground-state energies, we establish a sharp $M^{-1}$ convergence rate at fixed cell size, where $M$ is the number of one-particle basis functions. We also show that a uniform $M^{-1}$ convergence rate is impossible for the corresponding real-time wave-function error with general $H^2$ initial data. The analysis relies on many-body Sobolev estimates with explicit polynomial dependence on system size. Combining these estimates with state-of-the-art quantum algorithms, we obtain, to our knowledge, the first rigorous end-to-end complexity bounds for quantum simulation of molecular and electronic dynamics with all spatial errors explicitly controlled. The number of basis functions is determined directly from the continuum problem, without regularization.

quant-ph↗

$T$-efficient unitary synthesis via block-sparse Hamiltonian simulation

We construct Clifford+$T$ circuits implementing arbitrary $n$-qubit unitaries with worst-case $T$-count $\widetilde{O}(N^{8/7})$ at constant operator-norm error, where $N=2^n$, improving the previous exponent $5/4$. The construction uses $\widetilde{O}(N^{6/7})$ clean ancillas. The main idea is to reduce the data addressed by each query to the block-encoding circuit for a sparse piece while controlling the additional simulation work needed to combine the pieces. We simulate a Hermitian logarithm of the target by flattening its traceless part with a two-layer random similarity transformation, randomly partitioning the transformed generator into block-sparse Hermitian pieces, and combining their evolutions using qDRIFT. For a target operator-norm error $ε\in(0,1/2)$, let $L=\log_2(N/ε)$, the qubit count plus the requested bits of accuracy. By successively correcting the approximation error, we obtain a circuit using $\widetilde{O}(N^{8/7}L^{3/7}+NL)$ $T$ gates and $\widetilde{O}(N)$ clean ancillas whenever $L\le N$. We also prove that, at constant precision, the exponent $8/7$ is optimal within a sequential block-SELECT model.

quant-ph↗

Rethinking Adapter Placement: A Dominant Adaptation Module Perspective

Low-rank adaptation (LoRA) is a widely used parameter-efficient fine-tuning method that places trainable low-rank adapters into frozen pre-trained models. Recent studies show that using fewer LoRA adapters may still maintain or even improve performance, but existing methods still distribute adapters broadly, leaving \emph{where to place a limited number of adapters to maximize performance} largely open. To investigate this, we introduce \textbf{PAGE} (\textbf{P}rojected \textbf{A}dapter \textbf{G}radient \textbf{E}nergy), a gradient-based sensitivity probe that estimates the initial trainable gradient energy available to each candidate LoRA adapter. Surprisingly, we find that PAGE is highly concentrated on a single shallow FFN down-projection across two model families and four downstream tasks. We term this module the \textbf{dominant adaptation module} and show that its layer index is architecture-dependent but task-stable. Motivated by this finding, we propose \textbf{DomLoRA}, a placement method that places a single adapter at the dominant adaptation module. With only \textbf{0.7\%} of vanilla LoRA's trainable parameters, DomLoRA outperforms it on average across downstream tasks, including instruction following, mathematical reasoning, coding, and multi-turn conversation. This method also matches or improves other LoRA variants and reduces training time by up to \textbf{2.74}$\times$ compared with broad placement, supporting the dominant adaptation module perspective as a practical placement guideline.

cs.AI↗

AIR: Analytic Imbalance Rectifier for Continual Learning

Continual learning (CL) agents incrementally learn from sequentially arriving data and adapt to the dynamic, ever-changing nature of real-world environments. However, many existing CL methods suffer performance degradation in evolving, imbalanced data streams due to limited adaptation to changing class frequencies or ineffective use of mixed data from new and previously observed classes. To deal with these challenges, we propose an analytic imbalance rectifier (AIR) algorithm for real-world CL. AIR is an online exemplar-free approach with a frozen backbone as the feature extractor and a closed-form incremental classifier whose weight equals the joint-learning weight for the same class-weighted ridge objective. AIR addresses class imbalance with an analytic reweighting module (ARM) that calculates a reweighting factor for each class in the loss function to equalize total sample weights across classes. Under long-tailed class-incremental learning, AIR leads 28 baselines in aggregate accuracy and exemplar-free methods in aggregate macro F1, gaining 3.21% accuracy and 2.14% macro F1 over the respective strongest exemplar-free baselines. Under the Si-Blurry setting with recurring classes, AIR leads 15 exemplar-based and exemplar-free baselines, gaining 2.32% aggregate accuracy and 1.27% aggregate macro F1 over the strongest baseline. One-sided paired tests support positive mean absolute gains in these four comparisons (Holm-adjusted p<0.006).

cs.LG↗

Trotterization with Many-body Coulomb Interactions: Convergence for General Initial Conditions and State-Dependent Improvements

Efficiently simulating many-body quantum systems with Coulomb interactions is a fundamental question in quantum physics, quantum chemistry, and quantum computing, yet it presents unique challenges: the Hamiltonian is an unbounded operator; its Hilbert space dimension grows exponentially with particle number; and the Coulomb potential is singular, long-ranged, and non-smooth, violating the regularity assumptions of many prior state-of-the-art many-body simulation analyses. In this work, we establish rigorous error bounds for Trotter formulas applied to these systems. Our first main result shows that for general initial conditions in the domain of the Hamiltonian, second-order Trotter achieves a $1/4$ convergence rate with explicit polynomial dependence of the error prefactor on the particle number. The polynomial dependence on system size suggests that the algorithm remains quantumly efficient, even without introducing any regularization of the Coulomb singularity. Such a worst-case rate has been observed in prior work for the hydrogen ground state, demonstrating its physical relevance. We further establish one-step local error lower bounds of order $5/4$ for both formulas, showing that the local exponent is sharp. Our second main result identifies a set of physically meaningful conditions on the initial state under which the convergence rates improve, with the full first-order and second-order rates recovered at sufficiently high angular momentum. Our theoretical findings are consistent with prior numerical observations. Our third main result establishes improved convergence rates for many-body fermionic Coulomb systems with initial data of different Sobolev regularity. In particular, even without imposing any additional regularity beyond the natural domain of the Hamiltonian, both the first- and second-order Trotter formulas converge with rate $1/2$, rather than the general $1/4$ rate.

quant-ph↗

Qubit-Efficient Quantum Algorithm for Linear Differential Equations

As quantum hardware rapidly advances toward the early fault-tolerant era, a key challenge is to develop quantum algorithms that are not only theoretically sound but also hardware-friendly on near-term devices. In this work, we propose a quantum algorithm for solving linear ordinary differential equations (ODEs) with a provable runtime guarantee. Our algorithm uses only a single ancilla qubit, and is locality preserving, i.e., when the coefficient matrix of the ODE is $k$-local, the algorithm only needs to implement the time evolution of $(k+1)$-local Hamiltonians. We also discuss the connection between our proposed algorithm and Lindbladian simulation. By applying our algorithm to the interacting Hatano-Nelson model, a widely studied non-Hermitian model with rich phenomenology, and numerically simulating it under realistic noise models, we demonstrate its practical feasibility on near-term quantum devices.

quant-ph↗

Splitting Analysis for Yukawa Potential

Splitting methods are among the most classical and fundamental tools for the simulation of quantum dynamics, and their importance has grown further with the rise of quantum computing. In this work, we analyze the Schrödinger equation with Yukawa potential, a physically relevant and widely used model potential. It may be viewed as a Coulomb interaction with exponential decay at spatial infinity, preserving the Coulomb singularity at the origin while removing the long-range Coulomb tail. We prove that the operator splitting for this unbounded Hamiltonian achieves a global $1/4$-order convergence rate in the time step for many-body Yukawa interactions, with explicit polynomial dependence on the number of particles. The result holds for all initial wavefunctions in $H^2(\mathbb R^{3N})$, the natural domain of the Hamiltonian, and our numerical experiments are consistent with the theoretical estimates. To identify the sharp obstruction behind this rate, we prove a short-time lower bound in the one-body setting of order $t^{5/4}$ for the one-step error, which rules out any uniform global estimate of order better than $1/4$ in general. This agreement with the optimal $1/4$ rate in the Coulomb case is particularly interesting, as Yukawa potential is short-ranged compared to Coulomb potential. For the many-body upper bound, one of the new technical ingredients is the explicit polynomial-in-system-size Sobolev estimates of many-body Yukawa systems. These estimates are crucial for obtaining fully a priori bounds that depend only on the norms of the initial states, rather than on the solution at time $t$. For the one-body lower bound, we leverage a new analysis argument based on Fourier analysis and Kato smoothing.

math.NA↗

Circuit Depth Reduction of One-Ancilla Quantum Differential Equation Solver via Extrapolation

Solving linear differential equations is a fundamental task in scientific computing and an important primitive for quantum computing. A recent one-ancilla quantum differential equation solver provides a hardware-friendly and locality-preserving approach with provable performance guarantees, making it highly suitable for the early fault-tolerant and near-term regimes. Its simple circuit structure comes with a natural trade-off: the maximum single-run circuit depth scales as $O (1/ε)$ in the target accuracy $ε$. In this work, we reduce this depth by combining the solver with classical step-size postprocessing. By running the one-ancilla solver at a logarithmic number of finite time step sizes and using classical post-processing to cancel leading discretization errors, we reduce the maximum single-run circuit depth to $O(\mathrm{polylog}(1/ε))$ without adding quantum ancillae or sacrificing locality. Technically, extending extrapolation ideas beyond Hamiltonian and Lindbladian dynamics requires regularity estimates for observable maps under nonunitary evolution, which we obtain through a holomorphic extension of the adjoint evolution. Numerical experiments on the Hatano-Nelson model (ODE) and the convection-diffusion equation (PDE) demonstrate the effectiveness of the approach.

quant-ph↗

Logical Resource Estimation for Quantum State Preparation with Compilation

Quantum state preparation is a fundamental primitive in quantum algorithms for encoding classical data into quantum amplitudes. We compare the cost of preparing general $n$-qubit states with real amplitudes using two common paradigms: rotation-based methods, based on controlled rotations, and sampling-based methods, based on a structured representation of the target state. Although these approaches are often theoretically compared using CNOT count and $T$-count, their relative performance in total gate count remains less well understood practically. We compare representative rotation-based and sampling-based methods using $T$-count and total gate count, and analyze how compilation overhead affects their relative performance. We also develop a software package for compiling state preparation circuits, designed as a practical subroutine for more general quantum computations. Numerical experiments on resource states and quantum states related to quantum chemistry, condensed matter physics, and simulation via Magnus expansion over a range of target accuracies $ε$ support the analysis. Our results show that sampling-based methods achieve asymptotically lower $T$-count and retain an overall advantage after accounting for total gate count and compilation overhead.

quant-ph↗

AFL: A Single-Round Analytic Approach for Federated Learning with Pre-trained Models

In this paper, we introduce analytic federated learning (AFL), a new training paradigm that brings analytical (i.e., closed-form) solutions to the federated learning (FL) with pre-trained models. Our AFL draws inspiration from analytic learning -- a gradient-free technique that trains neural networks with analytical solutions in one epoch. In the local client training stage, the AFL facilitates a one-epoch training, eliminating the necessity for multi-epoch updates. In the aggregation stage, we derive an absolute aggregation (AA) law. This AA law allows a single-round aggregation, reducing heavy communication overhead and achieving fast convergence by removing the need for multiple aggregation rounds. More importantly, the AFL exhibits a property that \textit{invariance to data partitioning}, meaning that regardless of how the full dataset is distributed among clients, the aggregated result remains identical. This could spawn various potentials, such as data heterogeneity invariance and client-number invariance. We conduct experiments across various FL settings including extremely non-IID ones, and scenarios with a large number of clients (e.g., $\ge 1000$). In all these settings, our AFL constantly performs competitively while existing FL techniques encounter various obstacles. Our codes are available at https://github.com/ZHUANGHP/Analytic-federated-learning.

cs.LG↗

On the Trotter Error in Many-body Quantum Dynamics with Coulomb Potentials

Efficient simulation of many-body quantum systems is central to advances in physics, chemistry, and quantum computing, with a key question being whether the simulation cost scales polynomially with the system size. In this work, we analyze many-body quantum systems with Coulomb interactions, which are fundamental to electronic and molecular systems. We prove that Trotterization for such unbounded Hamiltonians achieves a $1/4$-order convergence rate, with explicit polynomial dependence on the number of particles. The result holds for all initial wavefunctions in the domain of the Hamiltonian, and the $1/4$-order convergence rate is optimal, as previous work has numerically demonstrated that it can be saturated by a specific initial ground state. The main challenges arise from the many-body structure and the singular nature of the Coulomb potential. Our proof strategy differs from prior state-of-the-art Trotter analyses, addressing both difficulties in a unified framework. Our analysis treats the Coulomb potential as an unbounded operator without modification or regularization, and does not rely on spatial discretization, making it compatible with both first- and second-quantized circuit constructions.

quant-ph↗

Uniform semiclassical observable error bound of Trotter-Suzuki splitting: a simple algebraic proof

Efficient simulation of the semiclassical Schrödinger equation has garnered significant attention in the numerical analysis community. While controlling the error in the unitary evolution or the wavefunction typically requires the time step size to shrink as the semiclassical parameter $h$ decreases, it has been observed -- and proved for first- and second-order Trotterization schemes -- that the error in certain classes of observables admits a time step size independent of $h$. In this work, we explicitly characterize this class of observables and present a new, simple algebraic proof of uniform-in-$h$ error bounds for arbitrarily high-order Trotterization schemes. Our proof relies solely on the algebraic structure of the underlying operators in both the continuous and discrete settings. Unlike previous analyses, it avoids Egorov-type theorems and bypasses heavy semiclassical machinery. To our knowledge, this is the first proof of uniform-in-$h$ observable error bounds for Trotterization in the semiclassical regime that relies only on algebraic structure, without invoking the semiclassical limit.

math.NA↗

REAL: Representation Enhanced Analytic Learning for Exemplar-free Class-incremental Learning

Exemplar-free class-incremental learning (EFCIL) aims to mitigate catastrophic forgetting in class-incremental learning (CIL) without available historical training samples as exemplars. Compared with its exemplar-based CIL counterpart that stores exemplars, EFCIL suffers more from forgetting issues. Recently, a new EFCIL branch named Analytic Continual Learning (ACL) introduces a gradient-free paradigm via Recursive Least-Square, achieving a forgetting-resistant classifier training with a frozen backbone during CIL. However, existing ACL suffers from ineffective representations and insufficient utilization of backbone knowledge. In this paper, we propose a representation-enhanced analytic learning (REAL) to address these problems. To enhance the representation, REAL constructs a dual-stream base pretraining followed by representation enhancing distillation process. The dual-stream base pretraining combines self-supervised contrastive learning for general features and supervised learning for class-specific knowledge, followed by the representation enhancing distillation to merge both streams, enhancing representations for subsequent CIL paradigm. To utilize more knowledge from the backbone, REAL presents a feature fusion buffer to multi-layer backbone features, providing informative features for the subsequent classifier training. Our method can be incorporated into existing ACL techniques and provides more competitive performance. Empirical results demonstrate that, REAL achieves state-of-the-art performance on CIFAR-100, ImageNet-100 and ImageNet-1k benchmarks, outperforming exemplar-free methods and rivaling exemplar-based approaches.

cs.LG↗

Discrete Superconvergence Analysis for Quantum Magnus Algorithms of Unbounded Hamiltonian Simulation

Motivated by various applications, unbounded Hamiltonian simulation has recently garnered great attention. Quantum Magnus algorithms, designed to achieve commutator scaling for time-dependent Hamiltonian simulation, have been found to be particularly efficient for such applications. When applied to unbounded Hamiltonian simulation in the interaction picture, they exhibit an unexpected superconvergence phenomenon. However, existing proofs are limited to the spatially continuous setting and do not extend to discrete spatial discretizations. In this work, we provide the first superconvergence estimate in the fully discrete setting with a finite number of spatial discretization points $N$, and show that it holds with an error constant uniform in $N$. The proof is based on the two-parameter symbol class, which, to our knowledge, is applied for the first time in algorithm analysis. The key idea is to establish a semiclassical framework by identifying two parameters through the discretization number and the time step size rescaled by the operator norm, such that the semiclassical uniformity guarantees the uniformity of both. This approach may have broader applications in numerical analysis beyond the specific context of this work.

math.NA↗

Superconvergence of High-order Magnus Quantum Algorithms

The Magnus expansion has long been a celebrated subject in numerical analysis, leading to the development of many useful classical integrators. More recently, it has been discovered to be a powerful tool for designing quantum algorithms for Hamiltonian simulation in quantum computing. In particular, surprising superconvergence behavior has been observed for quantum Magnus algorithms applied to the simulation of the Schrödinger equation, with the first- and second-order methods exhibiting doubled convergence order. In this work, we provide a rigorous proof that such superconvergence extends to general high-order quantum Magnus algorithms. Specifically, we show that a quantum Magnus algorithm of order $p$ achieves the superconvergence of order $2p$ in time when applying to the Schrödinger equation simulation in the interaction picture. Our analysis combines techniques from semiclassical analysis and Weyl calculus, offering a new perspective on the mathematical foundations of quantum algorithms for time-dependent Hamiltonian simulation.

math.NA↗

High-order Magnus Expansion for Hamiltonian Simulation

Efficient simulation of quantum dynamics with time-dependent Hamiltonians is important not only for time-varying systems but also for time-independent Hamiltonians in the interaction picture. Such simulations are more challenging than their time-independent counterparts due to the complexity introduced by time ordering. Existing algorithms that aim to capture commutator-based scaling either exhibit polynomial cost dependence on the Hamiltonian's time derivatives or are limited to low-order accuracy. In this work, we establish the general commutator-scaling error bounds for the truncated Magnus expansion at arbitrary order, where only Hamiltonian terms appear in the nested commutators, with no time derivatives involved. Building on this analysis, we design a high-order quantum algorithm with explicit circuit constructions. The algorithm achieves cost scaling with the commutator structure in the high-precision regime and depends only logarithmically on the Hamiltonian's time variation, making it efficient for general time-dependent settings, including the interaction picture.

quant-ph↗