As artificial intelligence (AI) is increasingly embedded in wireless networks, models are becoming core components that influence signal processing, resource scheduling and network control. However, model anomalies, tampering and malicious functions also introduce new security risks. In this article, we focus on model forensics in AI-native wireless networks. Specifically, we first discuss key problems including model authenticity verification, malicious function identification and accountability tracing, and summarize the main categories of model forensics. We then explain the role of model forensics in AI-native wireless networks and review representative application scenarios. In the case study, we use RF fingerprinting as an example and present two concrete workflows based on watermark authentication and backdoor detection, illustrating how provenance authentication and malicious behavior identification can be implemented in practice. The results show that model forensics can provide important support for anomaly assessment, provenance tracing and trustworthy operation in AI-native wireless networks. Finally, we outline several promising directions for future research in this emerging area.

We study the inverse curve shortening flow in the hyperbolic plane $\h^2$. We classify all solitons with respect to parabolic and conformal vector fields of $\h^2$. In the upper half-plane model of $\h^2$, we prove that parabolic solitons are all graphs on the $y$-axis, whereas conformal solitons are graphs on the $x$-axis. We study the concavity of these solitons and when they approach the coordinate axes.

We classify all ruled surfaces in Euclidean space that are critical points of the Dirichlet energy, obtaining explicit parametrizations of these surfaces.

This paper investigates the nonlinear dynamics of Newton's problem of minimal resistance in radial fields. We move beyond classical translational symmetry to analyze two non-equilibrium scenarios: a scale-invariant free expansion and an incompressible source flow. Our analysis reveals that the scale-invariant model suffers from a symmetry-breaking instability (loss of ellipticity) that necessitates geometric truncation. Conversely, we prove that the incompressible flow acts as a structural regularizer, admitting unique, smooth, and strictly concave solutions. These findings provide new qualitative insights into how physical conservation laws ensure the regularity and symmetry of optimal configurations in high-speed central flows, bridging the gap between variational calculus and the physics of complex systems.

Recently, the technique of counterdiabatic driving, which provides an effective strategy for accelerating adiabatic quantum evolution, has been widely applied in the preparation of many-body quantum states. In this work, we propose a theoretical scheme for the efficient preparation of Dicke states in a system of non-interacting two-level atoms. Our approach leverages the one-axis twisting (OAT) interaction to generate non-classical correlations and combines it with time-dependent external fields to achieve precise control over the dynamics of the system. By employing rapid adiabatic passage (RAP), it demonstrates how the system can be steered from an initial coherent spin state to a target Dicke state with high fidelity [S. C. Carrasco, M. H. Goerz, S. A. Malinovskaya, V. Vuletić, W. P. Schleich, and V. S. Malinovsky, Phys. Rev. Lett. \textbf{132}, 153603 (2024)]. To further optimize the preparation process, we introduce counterdiabatic driving (CD), which suppresses non-adiabatic transitions. Numerical simulations confirm that our scheme can achieve high-fidelity Dicke states for a moderate number of particles. Our results provide a scalable and experimentally feasible approach to prepare Dicke states, with potential applications in quantum metrology, quantum communication, and quantum information processing.

This work investigates the mechanism of gravitational baryogenesis (GB) under the formalism of f(R, G_{μν}T^{μν}) gravity, where R denotes the Ricci scalar, G_{μν} is the Einstein tensor and T^{μν} represents the energy--momentum tensor. f(R, G_{μν}T^{μν}) model is considered to evaluate the baryon-to-entropy ratio (BnER), which is subsequently compared against the observational limits. The results obtained exhibit compatibility with the estimated matter imbalance. Moreover, the analysis is extended to generalized GB case, resulting in outcomes that closely match empirical bounds. The findings reveal that the f(R, G_{μν}T^{μν}) formulation yields a viable theoretical setting for explaining the detected matter-antimatter disparity of the universe, highlighting its relevance in early cosmic evolution. To further validate the models, a chi-square (χ^2) analysis of the Hubble parameter, H(z), and distance modulus, μ(z), is performed, confirming their consistency with current cosmological observations. A comparative assessment simultaneously with the ΛCDM paradigm demonstrates a satisfactory level of agreement between the proposed model and cosmological observations from CC and Pantheon+SH0ES datasets.

We study gossip algorithms for the fundamental rumor spreading problem, where the goal is to disseminate a rumor from a given source node to all nodes in an arbitrary (and unknown) graph. Gossip algorithms allow each node to call only one neighbor per round and are therefore highly message-efficient, with low per-node communication overhead per round. The state of the art present fast gossip algorithms, however they typically leverage large-sized messages. This undermines the light-weight communication advantage of gossip, since even though only one neighbor is contacted per round, the message size can be linear in $n$, the network size. Hence, a fundamental question is whether one can perform fast gossip using small messages. The main contribution of this paper is to answer the above question in the affirmative and present two gossip algorithms that achieve fast rumor spreading using messages of polylog{n} size. Specifically, we present the following algorithms: 1. An algorithm that runs in $O(c \log n / Φ_c)$ rounds for every $c \geq 1$, and $Φ_c$ is the weak conductance. Our bound in terms of weak conductance is essentially optimal. 2. An algorithm that depends on the network diameter (and is independent of the graph's conductance), which runs in $\tilde{O}(D+\sqrt{n})$ rounds with high probability. Our algorithm can be modified to output a minimum spanning tree (MST) in the same number of rounds, which is essentially round-optimal (even for non-gossip algorithms). Our gossip algorithms use graph sketches [Ahn, Guha, McGregor, SODA 2012] in a novel way to overcome communication bottlenecks and achieve small communication overhead with small message sizes.

We study the fundamental problem of graph exploration in dynamic graphs using mobile agents. We consider $1$-interval connected dynamic graphs, where the topology may change arbitrarily from round to round as long as the graph remains connected, and edges are assigned with the dynamic port labeling at each round. The execution follows a semi-synchronous scheduler, under which an adversary may deactivate an arbitrary subset of agents in each round. For a graph with $n$ nodes and $k$ agents, we show that exploration is impossible if the adversary can deactivate at least $ \left\lceil \frac{k}{n-2} \right\rceil - 1$ agents per round, even when agents are equipped with unbounded memory, have global communication and full visibility. This yields an upper bound, implying that exploration is solvable only when the adversary deactivates at most $\left\lceil \frac{k}{n-2} \right\rceil - 2$ agents per round. We further establish that achieving exploration at this threshold requires agents to have both $1$-hop visibility and $1$-hop communication. Finally, we present the exploration algorithm using $k$ agents when the adversary deactivates at most $ \left\lceil \frac{k}{n-2} \right\rceil - 2$ agents, assuming agents are equipped with $1$-hop visibility and global communication, and matches the adversarial deactivation bound implied by the impossibility results.

We present a complete calculation of the one-loop self-energies for all fields in the linear sigma model coupled to quarks at finite temperature and in the presence of a uniform magnetic field. The analysis consistently incorporates thermal and magnetic effects for both neutral and charged degrees of freedom, providing a unified framework valid for arbitrary values of the temperature and the field strength. The computation is performed using the Matsubara formalism to account for finite temperature effects and the Schwinger proper-time representation for charged propagators in a magnetic background. Special attention is given to loop contributions involving particles with different electric charges, for which the associated Schwinger phases do not cancel. We show that these terms can be systematically evaluated in coordinate space using the Ritus formalism, which provides the appropriate framework for treating external charged states in the presence of a magnetic background, and consistently expressed in momentum space. The resulting expressions exhibit a nontrivial interplay between thermal fluctuations and magnetic effects and allow for a clear separation between vacuum and matter contributions, providing a well-defined structure for the identification of ultraviolet divergences. Our results establish a consistent and systematic framework for the computation of thermomagnetic one-loop corrections in effective models of QCD, capturing the full interplay between thermal and magnetic effects for all dynamical degrees of freedom.

Let $Δ$ be the Dirichlet Laplacian on the interval $(0,π)$, and let $T>0$. We prove a well-posedness results for the structurally damped beam equation $$u_{tt}+Δ^2 u-ρΔu_t=0, x\in (0,π),t>0$$ with various boundary conditions including $$ u(0,t)=u_{xx}(0,t)=0; u(π,t)=f(t),u_{xx}(π,t)=0, $$ and $f\in H_0^2(0,T)$ and appropriate initial conditions. Viewing $f$ as a control, we prove null controllability for all $ρ\leq 2$. For $ρ>2$, we show null controllability for arbitrary $T>0$ holds for almost all $ρ$, but fails for a dense subset of $(2,\infty)$. An analagous result is proven for Neumann control.

Let $\mathcal{P}$ denote the set of all primes, and let $\underlineδ(P)$ denote the relative lower density of a subset $P$ in $\mathcal{P}$. Suppose that $P_1, P_2, P_3, P_4$ are four subsets of primes with $\underlineδ(P_1)+\underlineδ(P_2)>1$ and $ \underlineδ(P_3)+\underlineδ(P_4)>1.$ Then for every sufficiently large even integer $n$, there exist primes $p_i \in P_i$ $(i=1,2,3,4)$ such that $n=p_1+p_2+p_3+p_4$. The condition is the best possible.

Designing effective practice schedules for high-dimensional motor learning tasks remains a challenge, especially when skill states are unobservable and task performance may not reflect the true learning. We propose an automated curriculum design framework that combines a human motor learning model and personalized real-time skill estimation with Stochastic Nonlinear Model Predictive Control in \emph{de-novo} (novel) motor learning paradigms. We validated our framework both through simulations and human-subject studies (N = 36) using a hand exoskeleton. Our proposed approach accelerates skill acquisition by $\sim23\%$, and ${\sim17\%}$ when compared to a random curriculum and a performance heuristics-based curriculum, respectively. These significant gains in learning efficiency highlight the potential of model-based, individualized curricula for motor rehabilitation and complex skill training.

In this work, we consider the interaction of a 3D incompressible fluid with a 2D flexible shell that occupies (a part of) the boundary of the fluid domain. We assume that the shell is perfectly elastic while the fluid is governed by the Navier--Stokes equations. Consequently, damping within the coupled system comes entirely from the parabolic fluid subsystem. Our main result is the construction of a local-in-time unique strong solution to the system of PDEs. Standard techniques from the literature do not apply here. They are restricted to visco-elastic structures, where the corresponding solid phase is parabolic. Our construction relies on a different method built upon a new estimate for the acceleration of the system. In the case of a 2D viscous incompressible fluid interacting with a 1D perfectly elastic shell we can extend the local solution globally in time (until a possible self-intersection of the shell).

Matrix product states (MPS) are a standard tensor-network representation for ground states of one-dimensional quantum many-body systems, and they underpin widely used simulation tools such as DMRG. However, while quantum model checking has been developed mainly for quantum programs and communication protocols (with properties expressed along a time axis), there is still no comparable framework for systematically verifying \emph{spatial} and \emph{size-dependent} properties of physical many-body states, where the key parameter is the system size. This paper takes a step toward bridging the gap. We propose \emph{Linear Chain Logic} (LCL), a spatial logic designed to specify physically meaningful properties of periodic MPS families as the system size grows, such as nontriviality on rings and large-size asymptotic patterns. Our approach builds on a simple but powerful connection: every periodic MPS naturally induces a completely positive map (a quantum operation) on its virtual space, so many quantitative features of the MPS can be analysed through the repeated application of the operation. Using this perspective, we derive an effective procedure to compute the inner products of an MPS at a given size and to support richer LCL specifications, without relying on brute-force state expansion. We then develop approximate model-checking algorithms that combine sound bounding with asymptotic structural analysis, enabling scalable reasoning about large system sizes. Experiments on representative MPS families illustrate that our method can automatically verify nontriviality and detect asymptotic spatial regimes in a way that complements traditional numerical techniques.

Nanoscale fabrication inevitably produces local stochasticity that is commonly treated as a defect, but can instead be harnessed as a material resource for information security. Here we report a chiroptical platform for ternary entropy harvesting based on stochastic Au nanopatterns formed by block copolymer self-assembly. By transducing fabrication-induced stochastic microstates into handedness-dependent optical responses through chiroptical mapping, our platform enables native ternary digitization rather than conventional binary encoding, allowing physically harvested ternary random sequences to be used for key generation. This raises the information density to log2(3) = 1.585 bits per trit, approximately 58.5% higher than the binary limit, enabling more entropy to be harvested from a limited physical footprint. The harvested outputs exhibit near-balanced symbol populations, negligible spatial and inter-sample correlations, Shannon entropy approaching the ternary ideal, and resistance to statistical and machine-learning-based prediction. These results establish self-assembled chiroptical nanostructures as a scalable platform for cryptographic key generation, secure edge devices, and distributed Internet-of-Things platforms.

Amorphous solids represent the extreme limit of broken translational symmetry, in which the absence of long-range order removes well-defined crystal momenta and invalidates the Bloch description of electronic states. Monolayer amorphous carbon (MAC) has emerged as a unique realization of a strictly two-dimensional (2D) amorphous lattice defined by a structurally contiguous but topologically disordered $sp^2$-bonded random network devoid of any defined long-range crystal symmetry. From atomic-resolution measurements of multifractal wavefunctions, we show that disorder in MAC effectively localizes the low-energy part of the electronic spectrum but retains an extended critical-like state near the band centre ($E\sim 0$). We conjecture that this state is protected from topological disorder by remnant chiral symmetry surviving within the continuous random network, described by a Wess-Zumino-Witten (WZW) topological term. Near criticality, we verify the multifractal scaling relation $η= -Δ_2$, providing quantitative agreement between independently measured spatial correlation decay and multifractal scaling exponents. Our results are confirmed by atomistic tight-binding calculations that closely mirror the multifractal scaling near $E\sim 0$. Our results establish MAC as the first strictly 2D amorphous electronic system to exhibit Anderson criticality driven purely by topological disorder

We present a $Z_1^2$-based framework for estimating the spin frequency and frequency derivative of high-energy pulsars from Poisson-limited photon event lists. The key point is that the width of a coherent detection peak is not, by itself, the statistical uncertainty on the recovered rotational parameters. We develop and compare three computationally efficient estimators: segmented frequency regression, a coherent derivative scan, and a localized two-dimensional coherent fit. For sinusoidal signals, we derive the local form of the Z-squared response as a function of frequency and frequency derivative, and show that expressing the frequency at the midpoint of the observation removes the leading-order covariance between the two parameters. This gives simple uncertainty estimates in terms of the fitted peak amplitude and local widths, without requiring an exhaustive Monte Carlo simulation for each observation. We test these estimates with Monte Carlo simulations over a range of observing spans, signal strengths, grid resolutions, and good-time-interval structures, and show that the predicted uncertainties reproduce the run-to-run scatter of the recovered parameters in the tested regimes. We then apply the framework to AstroSat/LAXPC event lists for the Crab pulsar, Swift J0243.6+6124, and SAX J1808.4-3658. The results provide a practical and statistically motivated route to rotational-parameter estimation for targeted high-energy pulsar searches.

We study discrete dynamics governed by a difference inclusion whose increment is the sum of a selection from a set-valued map and a noise term. For any bounded realization, convergence follows once the inter-iterate diameter is controlled by the variation of a continuous potential. The limit point is then critical for a scaled outer limit of the update map. To certify this diameter criterion, we develop a stratified descent framework: we project iterates onto a suitable stratification and track a potential that decreases up to a summable error. Combining the diameter criterion with a diameter estimate obtained from this framework yields convergence of common first-order optimization methods under step sizes of order $1/k$. The guarantees cover inexact and stochastic subgradient methods, as well as the momentum method, for locally Lipschitz objectives definable in polynomially bounded o-minimal structures. Our arguments are entirely discrete, with no appeal to continuous-time approximations.

In this paper, we find several determinants expressing the Fibonomial coefficients. We also give the generating functions, Vandermonde identity, and continued fractions about Fibonomial coefficients.

This paper presents a predictive deep learning framework for dynamic sub-band allocation in Sub-Band Full Duplex (SBFD) systems, addressing the challenge of balancing uplink (UL) and downlink (DL) performance under highly dynamic traffic conditions. The key contribution lies in integrating a hybrid Bidirectional Long Short-Term Memory (Bi-LSTM) model for traffic forecasting with a Double Deep Q-Network (DDQN) for real-time resource allocation. Using both predicted traffic and current queue states, the proposed system enables proactive scheduling based on traffic demand. Evaluation results show that the prediction model achieves high accuracy in capturing bursty traffic patterns, while the DDQN agent effectively adapts UL/DL split ratios according to traffic variations. The framework improves spectrum utilization, reduces queue buildup, and avoids inefficient static configurations. The proposed approach demonstrates that combining predictive intelligence with reinforcement learning significantly enhances the efficiency and adaptability of SBFD systems, making it a strong candidate for autonomous resource management in future 6G networks.

We revisit the derivation of the orbital magnetization formula for periodic crystals in current density functional theory (CDFT)[1]. Our new derivation computes the linear response of the energy density to a periodic magnetic field in the long-wavelength limit. We unveil a Ward identity which connects the current vertex to the derivative of the Kohn-Sham self-energy. The result of Ref.[1] is confirmed: the orbital magnetization of the interacting solid can be computed exactly (in principle) from the self-consistent eigenfunctions and eigenvalues of the Kohn-Sham equation of CDFT.

We prove the local Hadamard well-posedness of the ``good'' Boussinesq equation formulated on the half-line with nonzero Robin boundary conditions. These boundary data involve the Dirichlet and Neumann boundary values as well as the second spatial derivative of the solution evaluated at the boundary. The nonlinear analysis crucially relies on the linear estimates established through the explicit solution formula obtained for the forced linear counterpart of the problem via Fokas's unified transform. The two pieces of initial data and the two pieces of boundary data belong in appropriate Sobolev spaces. The corresponding solution is established in the natural Hadamard solution space of continuous/continuously differentiable functions from a suitable time interval to the Sobolev spaces associated with the two initial data. Furthermore, in line with the well-posedness theory of the Cauchy problem, in the case of low regularity (namely, below the spatial continuity threshold) the solution space is refined by also including an appropriate spatiotemporal Lebesgue space.

Many real-life applications involve controlling high-dimensional multi-agent systems in real-time. Existing optimal control solvers often suffer from the curse-of-dimensionality and require complete rerunning for each new problem setting. We target nonconvex, nonlinear problems in 100s of dimensions by introducing PI-SONet (Physics-Informed Symplectic Operator Network), a structure-preserving operator learning framework for solving parameterized families of optimal control problems and their Pontraygin Maximum Principle (PMP) systems. PI-SONet combines a latent right-space solver with a conditional symplectic operator to produce tractable Hamiltonian trajectories in a computationally efficient auxiliary space and transform them back to physical space. This decomposition yields a \textit{single} trained operator that approximates the PMP solution map, inherently preserves Hamiltonian structure, and generalizes across unseen problem configurations. Unlike existing methods, which are fundamentally single-instance solvers, PI-SONet achieves sub-second inferences on new problem instances, equating to up to 10,000x speedup over representative baselines. These results suggest that structure-preserving neural operators provide a practical route toward reusable, real-time surrogates for high-dimensional optimal control.

This mixed-methods study examined computational thinking (CT) development among 93 pre-high school students in a five-day AI agent creation workshop using CocoFlow, a no-code platform. Integrating pre-post assessments, behavioral logs, and interviews, we investigated CT development and how initial CT levels shape learning trajectories. Results revealed significant improvements in abstract thinking (effect size d = 0.71) and algorithmic thinking (effect size d = 0.70). Hierarchical regression identified iterative testing engagement as a predictor of self-efficacy gains (beta = 0.20, p = 0.05). Notably, students with moderate initial CT levels demonstrated substantially greater gains than both high-CT and low-CT peers, revealing an Optimal Development Zone effect (eta squared = 0.55). Qualitative analysis showed moderate-CT students exhibited adaptive expertise, while high-CT students risked over-engineering and low-CT students struggled with task decomposition. These findings challenge linear learning assumptions and provide evidence for differentiated scaffolding in CT education.

We report N-polar AlN-based high-electron mobility transistors (HEMTs) with a GaN channel thickness of 5.2 nm on N-polar AlN on sapphire. The threshold voltage is around -2.4 to -3.0 V with saturation currents over 240 mA/mm and on/off ratios as high as 10,000, much higher than previously reported N-polar AlN-based HEMTs. The high on/off ratio is attributed to the use of an abrupt AlN/GaN heterostructure with a dedicated AlN transition layer, together with improved gate leakage. The high frequency properties as well as the on-resistance of ~20 Ohm mm are all limited by the 2000 Ohm/square sheet resistance of the channel layer.

Recent theoretical and experimental advances in quantum ferroelectrics suggest that ferroelectricity can also emerge in non-polar space group, highlighting the limitations of conventional polar space group criteria in identifying ferroelectric materials. Here, we introduce a unified definition based on switchable polarization differences between energetically equivalent states, which naturally encompasses conventional and quantum ferroelectrics. Guided by this principle, we implement a high-throughput screening strategy that systematically identifies both conventional and quantum ferroelectrics among experimentally synthesized materials. In particular, we identify a new type of quantum ferroelectric in which the quantized polarization arises from arbitrary ionic displacements, in contrast to previous quantum ferroelectrics (including both fractional and integer quantum ferroelectrics) where quantized polarization results from fractional or integer ionic displacements. Notably, we find that materials such as Ba3I6 and Cs2PdC2 exhibit low switching barriers and robust insulating behavior, highlighting their experimental viability. Our results reconcile conventional and quantum ferroelectrics, expand the accessible materials landscape, and provide a practical roadmap for discovering next-generation ferroelectrics with advanced switchable functionalities.

As quantum computers become available through multi-tenant cloud platforms, ensuring privacy against adversaries sharing the same quantum processing unit becomes critical. We introduce and explore \emph{covert quantum computing}, a new concept that ensures an adversary with access to all other quantum computational units (QCUs) of a quantum computer cannot detect computation on the subset that they cannot access. Analogous to covert communication, we employ information theory. However, since here the adversary controls the systems used for detection, we require a richer framework for covertness analysis that accounts for the use of quantum memories and adaptive operations. Thus, we adopt the \emph{quantum-strategy} framework used in quantum game theory and memory channel discrimination. Current quantum computers use planar graph circuit layouts and typically assume nearest-neighbor crosstalk. We derive discrete isoperimetric inequalities to show that, for an $n$-qubit circuit under this model, only $\mathcal{O}(\sqrt{n})$ border qubits provide detection information to the adversary. We then explore this scaling law on IQM's 54-qubit \emph{Emerald} processor and IBM's 156-qubit \emph{ibm\_fez} machine employing the Heron 2 architecture. We implement Ramsey experiments on qubits not used in computation, and detect nearest-neighbor crosstalk, as expected. However, we also observe long-range coupling effects beyond the border qubits, revealing a side channel that the adversary can exploit. We hypothesize that this long-range crosstalk is induced by leakage from the drive and control lines. Beyond weakening covertness, it exposes co-tenants to both adversarial and unintended crosstalk and degrades circuits that span spatially distributed qubits, motivating further work on spatial isolation and crosstalk characterization.

We consider a subtraction Nim with subtraction set {s_1,s_2,s_3={2,4n,4n+2}, where n is a positive integer such that n >= 3. We do not treat the case that n=1 or n=2 in this article. We show that this game satisfies the reverse-mex property of Grundy numbers, i.e., G(x)=mex{G(x+s_1), G(x+s_2), G(x+s_3)}, where the mex is taken over successors rather than predecessors. We modify the rule of this subtraction Nim to allow a one-time pass, that is, a passing move usable at most once during the game, unavailable from terminal positions; once used by either player, it becomes unavailable. In classical Nim, the introduction of a pass move complicates the game, and finding a formula that describes the set of P-positions in traditional three-pile Nim with a pass remains an important open question. In the case of subtraction Nim with a pass, however, the introduction of a pass move does not complicate the game. We prove that this game still satisfies the reverse-mex property of Grundy numbers when a pass move is available.

Firing rate fluctuations in neural populations are observed experimentally over multiple time scales, in single neurons, across trials when elicited by stimuli, and across populations. In this work, we examine how firing rate fluctuations emerge in networks of coupled integrate-and-fire neurons as a function of the initial distribution of voltages in networks with time-varying inputs. We analytically derive an approximation for the evolution of the instantaneous population rate or flux as a function of the initial voltage distribution through a Fokker-Planck system. Unlike earlier mean field approaches based on asynchronous or constant flux steady state solutions to the Fokker-Planck system, the approach considered here is based on the transport solution to the advection equation and assumes that the time-varying inputs are slow, and the neurons are in the excitation-driven regime. The transport mean field system predicts how firing rate fluctuations emerge from a dynamic interaction between time-varying inputs, initial densities, and coupling in populations of neurons.

Superconducting nanowire single-photon detectors (SNSPDs) have demonstrated timing jitter in the few-picosecond regime, yet their timing resolution deteriorates substantially under high-count-rate operation. Existing interpretations mainly attribute this degradation to deterministic waveform distortions, such as multiphoton responses and pulse pile-up, yet the experimentally observed jitter broadening at high count rates cannot be fully accounted for within this picture. Here, we show that stochastic baseline fluctuations arising from finite-memory readout dynamics constitute an intrinsic source of the count-rate-dependent timing jitter in SNSPD systems. For stochastically arriving photons, overlapping recovery responses accumulate in the readout chain and generate statistically fluctuating baselines, which are converted into timing uncertainty through threshold-based timing extraction. We develop a stochastic-process framework that quantitatively connects photon statistics, readout dynamics, and timing jitter. The framework predicts characteristic scaling behaviors, including a nonmonotonic dependence of baseline fluctuations under pulsed excitation with a maximum near half of the repetition frequency. These predictions are quantitatively verified through systematic variations of count rate, circuit time constant, and detector dynamical properties. Our results identify stochastic baseline dynamics as a fundamental mechanism limiting timing resolution in high-count-rate SNSPD operation and provide a general framework for optimizing finite-memory high-speed photon-counting systems.