arXiv subjects
Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.
By creating some new concepts and methods: checking tree, long unit path, direct contradiction unit pair, indirect contradiction unit pair, additional contradiction unit pair, 2-unit layer and 3-unit layer, redundant units, and destroying parallel pairs , we successfully transform solving a 3SAT problem to solving 2SAT problems in polynomial time. Thus we proved that NP=P.
We investigate four well-known negative translations of classical logic into intuitionistic logic within a substructural setting. We find that in \emph{affine logic} the translation schemes due to Kolmogorov and Gödel both satisfy Troelstra's criteria for a negative translation. On the other hand, the schemes of Glivenko and Gentzen both fail for affine logic, but for different reasons: one can extend affine logic to make Glivenko work and Gentzen fail and {\it vice versa}. By contrast, in the setting of \emph{Łukasiewicz logic}, we can prove a general result asserting that a wide class of formula translations including those of Kolmogorov, Gödel, Gentzen and Glivenko not only satisfy Troelstra's criteria with respect to a natural intuitionistic fragment of Łukasiewicz logic but are all equivalent.
In this paper, we investigate linear stability properties of the 2D isentropic compressible Euler equations linearized around a shear flow given by a monotone profile, close to the Couette flow, with constant density, in the domain $\mathbb{T}\times \mathbb{R}$. We begin by directly investigating the Couette shear flow, where we characterize the linear growth of the compressible part of the fluid while proving time decay for the incompressible part (inviscid damping with slower rates). Then we extend the analysis to monotone shear flows near Couette, where we are able to give an upper bound, superlinear in time, for the compressible part of the fluid. The incompressible part enjoys an inviscid damping property, analogous to the Couette case. In the pure Couette case, we exploit the presence of an additional conservation law (which connects the vorticity and the density on the moving frame) in order to reduce the number of degrees of freedom of the system. The result then follows by using weighted energy estimates. In the general case, unfortunately, this conservation law no longer holds. Therefore we define a suitable weighted energy functional for the whole system, which can be used to estimate the irrotational component of the velocity but does not provide sharp bounds on the solenoidal component. However, even in the absence of the aforementioned additional conservation law, we are still able to show the existence of a functional relation which allows us to recover somehow the vorticity from the density, on the moving frame. By combining the weighted energy estimates with the functional relation we also recover the inviscid damping for the solenoidal component of the velocity.
We construct two Fréchet algebras admitting countably many mutually inequivalent Fréchet algebra topologies. The second example is a modification of the maiden (and first) example of a non-Banach Fréchet algebra with two inequivalent Fréchet algebra topologies, considered by Read to show that the famous Singer-Wermer conjecture does not hold in the Fréchet case. We discuss a couple of avatars of the Singer-Wermer conjecture in Fréchet algebras. In particular, this conjecture (and so, its weaker version, too) does hold by the second example which also admits countably many mutually equivalent Fréchet algebra topologies. However the stronger version of this conjecture (i.e., the Kaplansky's conjecture) does not hold by the second example.
The elastodynamic Peach-Koehler force is computed for a fully-regularized straight dislocation with isotropic core in continuum isotropic elastic elasticity, in compact forms involving partial mass or impulsion functions relative to shear and compressional waves. The force accounts for both dynamic radiation damping and inertia. The expressions are valid indifferently for subsonic or supersonic velocities. Results are compared with the case of a flat-core dislocation of the Peierls-Eshelby type, for a motion of jump from rest to constant velocity. In the steady-state limit, the Lagrangian function relevant to expressing the force in the flat-core case must be replaced by a related but different function for the regularized dislocation. However, by suitably defining the regularizing dislocation width, the steady-state limits of the force for the fully-regularized and flat-core dislocations can be matched exactly.
In 1952, Michael posed a question about functional continuity of commutative Fréchet algebras in his memoir, known as Michael's problem in the literature. We settle this in the affirmative for the non-commutative Fréchet algebras, along with its equivalent form. Indeed, we continue our previous works to directly attack these problems. We show that the existence of a discontinuous character on the non-commutative test case for this problem -- the Fréchet algebra ${\cal U}_{nc}$ of all entire functions on $\ell^\infty$ -- allows us to generate another Fréchet algebra topology on ${\cal U}_{nc}$, inequivalent to the usual Fréchet algebra topology. An important tool is a topological version of the tensor algebra over a Banach space. The elementary, but crucial, idea is to express the test algebra ${\cal U}_{nc}$ as a {\it weighted Fréchet tensor algebra} over the Banach space $\ell^1$. We give some important applications in automatic continuity theory. In particular, we provide affirmative answers to significant, long-standing problems from 1967 (extending Johnson's result for non-commutative, semi-simple Fréchet algebras; Dales' survey, Question 9), 1976 (extending Esterle's result, giving another approach for Michael's problem for more general $(F)$-algebras), and 1978 (a weaker version of an equivalent form of the non-commutative Michael's problem; Dales' survey, p. 143).
Langevin equations are usually understood as integral equations, since continuous and Markov solutions are nowhere differentiable. This is avoidable by the fact that white Gaussian noise can be seen as a generalized derivative of the Wiener process. The simplest LE is thus solved by Wiener paths, which can be computed by a Markov chain. That crucial property persists for functions of these paths, and solutions of the general LE are therefore assumed to have that form. In short time steps these functions must be linear or quadratic. Both cases lead to sampled versions of existing stochastic intrgrals ('Ito' and 'anti-Ito' respectively). In physics they apply for Brownian motion in presence of a temperature gradient: without a feedback to the medium in the linear case, and for a tagged molecule in the quadratic one; in the latter case the Fokker-Planck equation describes a local equilibrium.
We consider the problem of adaptive inference on a regression function at a point under a multivariate nonparametric regression setting. The regression function belongs to a Hölder class and is assumed to be monotone with respect to some or all of the arguments. We derive the minimax rate of convergence for confidence intervals (CIs) that adapt to the underlying smoothness, and provide an adaptive inference procedure that obtains this minimax rate. When the regression function is monotone in $k_{+}$ of its $k$ arguments, the minimax expected length over a Hölder class with exponent $γ'$ of CIs that maintain coverage over a rougher class with exponent $γ$ is of order $n^{-1/(2+k_{+}/γ'+(k-k_{+})/γ)}$. The rate is unchanged if the rougher class restricts the smoothness only at the point, along the nonmonotone arguments. Under monotonicity along a closed convex cone that contains no line, only the dimension of its linear span enters the rate. The procedure differs from that of Cai and Low (2004), intended to yield shorter CIs under practically relevant specifications. It is valid in finite samples under Gaussian errors with a known variance. The proposed method applies to general linear functionals of the regression function.
We interpret the log-Brunn-Minkowski conjecture of Böröczky-Lutwak-Yang-Zhang as a spectral problem in centro-affine differential geometry. In particular, we show that the Hilbert-Brunn-Minkowski operator coincides with the centro-affine Laplacian, thus obtaining a new avenue for tackling the conjecture using insights from affine differential geometry. As every strongly convex hypersurface in $\mathbb{R}^n$ is a centro-affine unit-sphere, it has constant centro-affine Ricci curvature equal to $n-2$, in stark contrast to the standard weighted Ricci curvature of the associated metric-measure space, which will in general be negative. In particular, we may use the classical argument of Lichnerowicz and a centro-affine Bochner formula to give a new proof of the Brunn-Minkowski inequality. For origin-symmetric convex bodies enjoying fairly generous curvature pinching bounds (improving with dimension), we are able to show global uniqueness in the $L^p$- and log-Minkowski problems, as well as the corresponding global $L^p$- and log-Minkowski conjectured inequalities. As a consequence, we resolve the isomorphic version of the log-Minkowski problem: for any origin-symmetric convex body $\bar K$ in $\mathbb{R}^n$, there exists an origin-symmetric convex body $K$ with $\bar K \subset K \subset 8 \bar K$, so that $K$ satisfies the log-Minkowski conjectured inequality, and so that $K$ is uniquely determined by its cone-volume measure $V_K$. If $\bar K$ is not extremely far from a Euclidean ball to begin with, an analogous isometric result, where $8$ is replaced by $1+ε$, is obtained as well.
Muchnik's paradox says that enumerable betting strategies are not always reducible to enumerable strategies whose bets are restricted to either even rounds or odd rounds. In other words, there are outcome sequences x where an effectively enumerable strategy succeeds, but no such parity-restricted effectively enumerable strategy does. We characterize the effective Hausdorff dimension of such $x$, showing that it can be as low as 1/2 but not less. We also show that such reals that are random with respect to parity-restricted effectively enumerable strategies with packing dimension as low as $\log\sqrt3$. Finally we exhibit Muchnik's paradox in the case of computable integer-valued strategies.
We determine the wall divisors on irreducible symplectic orbifolds which are deformation equivalent to a special type of examples, called Nikulin orbifolds. The Nikulin orbifolds are obtained as partial resolutions in codimension 2 of a quotient by a symplectic involution of a Hilbert scheme of 2 points on a K3 surface. This builds on the previous article arXiv:2009.04873 in which the theory of wall divisors was generalized to orbifold singularities.
Self-force methods can be applied in calculations of the scatter angle in two-body hyperbolic encounters, working order by order in the mass ratio (assumed small) but with no recourse to a weak-field approximation. This, in turn, can inform ongoing efforts to construct an accurate description of the general-relativistic binary dynamics via an effective-one-body description or other approaches. Existing self-force methods are to a large extent specialised to bound, inspiral orbits. Here we derive the first-order conservative self-force correction to the scattering angle, show its agreement with recent post-Minkowsian results, and develop a technique for (numerical) self-force calculations that can efficiently tackle scatter orbits. In the method, the metric perturbation is reconstructed from a Hertz potential that satisfies (mode-by-mode) a certain inhomogeneous version of the Teukolsky equation. The crucial ingredient in this formulation are certain jump conditions that the (multipole modes of the) Hertz potential must satisfy along the worldline of the small body's orbit. We present a closed-form expression for these jumps, for an arbitrary geodesic orbit in Schwarzschild spacetime. To begin developing the numerical infrastructure, a scalar-field evolution code on a Schwarzschild background (in 1+1D) is developed. Following this, results for the conservative scalar self-force corrections to the scatter angle are calculated. We continue by constructing a Teukolsky evolution code on a Schwarzschild background. This produces numerically unstable solutions due to unphysical homogeneous solutions of the Teukolsky equation at the horizon and null infinity being seeded by numerical error. This can be resolved by a change of variables to a Regge-Wheeler-like field. We then present a full numerical implementation of this method for circular and scatter orbits in Schwarzschild.
We give an account of Mazur's proof that, for an elliptic curve over $\mathbb{Q}$, if it admits a nonconstant mapping from $X(N)$ defined over the complex numbers $\mathbb{C}$, for some $N$, then it also admits a nonconstant mapping from $X_0(M)$ for some (possibly different) $M$ defined over the rational numbers $\mathbb{Q}$. We also briefly discuss two open questions of Khare concerning uniformisations of elliptic curves by noncongruence modular curves. This expository note is based on the author's expository talk in March 2022 during the 2nd Trimester Program on "Modularity and the Generalised Fermat Equation" held online and organised by Bhaskaracharya Pratishthana in Pune, India.
We study multidimensional online generalized assignment under nonstationary arrivals, motivated by resource allocation for large language model agent services. Providers may commit model and tool allowances to agent tasks upon admission, reserving budgets for their internal execution loops. Each request is rejected or accepted with an irrevocable assignment to one compatible server, earning reward and consuming a vector of its nonreplenishing budgets. Arrivals are independent with known, potentially time-varying type distributions. Scalar reductions can discard feasible combinations of requests with complementary resource requirements. To maximize expected reward, we develop Vector-Aware Best-Fit, which combines linear-programming-guided candidate routing with calibrated acceptance while retaining vector feasibility. Relative to the expected clairvoyant offline optimum, its guarantee improves on scalar reduction in the multidimensional case and has the best possible inverse-dimensional order. With an appropriate acceptance parameter, the proposed policy attains the optimal competitive ratio on all-heavy instances, where every feasible assignment consumes more than half of at least one resource budget. We further develop two heuristics that retain its performance guarantee. Computational results show that the heuristics perform well against benchmarks in most tested settings, especially when capacity is scarce relative to demand.
Given $\mathbb{F}_{q^{n}}$, a field with $q^n$ elements, where $q $ is a prime power and $n$ is positive integer. For $r_1,r_2,m_1,m_2 \in \mathbb{N}$, $k_1,k_2 \in \mathbb{N}\cup \{0\}$, a rational function $F = \frac{F_1}{F_2}$ in $\mathbb{F}_{q}[x]$ with deg($F_i$) $\leq m_i$; $i=1,2,$ satisfying some conditions, and $a,b \in \mathbb{F}_{q}$, we construct a sufficient condition on $(q,n)$ which guarantees the existence of an $r_1$-primitive, $k_1$-normal element $ε\in \mathbb{F}_{q^n}$ such that $F(ε)$ is $r_2$-primitive, $k_2$-normal with $\operatorname{Tr}_{\mathbb{F}_{q^n}/\mathbb{F}_q}(ε) = a$ and $\operatorname{Tr}_{\mathbb{F}_{q^n}/\mathbb{F}_q}(ε^{-1}) = b$. For $m_1=10, \; m_2=11,\; r_1 = 3, \; r_2 = 2, \; k_1=2,\;k_2 = 1$, we establish bounds on $q$, for various $n$, to determine the existence of such elements in $\mathbb{F}_{q^{n}}$. Furthermore, we identify all such pairs $(q,n)$ excluding 10 possible values of $(q,n)$, in fields of characteristics 13.
This paper studies the closed-loop Volterra method for analyzing and predicting stationary time series. The method represents the dependence of a current observation on past observations through polynomial terms, including powers of individual lags and interactions between different lags. Since the number of coefficients increases rapidly with the memory length and polynomial degree, we use the theory of reproducing kernel Hilbert spaces to estimate these coefficients by regularized least squares through kernel evaluations. We establish conditions on marginal moments, temporal dependence, model complexity and regularization under which the prediction risk converges to the minimum mean squared prediction error as the sample size increases. We also consider data-driven selection of tuning parameters by a rolling-origin procedure based on one-step-ahead prediction errors, and show that the resulting selector asymptotically attains the prediction risk of the best candidate in the considered grid. Simulation studies and a real-data application examine the finite-sample performance of the method and its comparison with linear and nonlinear prediction methods.
We consider the problem of dynamically selecting among competing strategies using observable predictors. To mitigate estimation noise and excessive switching, we regularise the estimated policy with a switching cost, which may exceed actual transaction costs when these are too small to provide sufficient regularisation. Since switching costs make current choices affect future decisions, we formulate the problem using dynamic programming. The optimal policy is estimated using a reinforcement learning algorithm based on iterated regressions, making the procedure simple to implement. We apply the method to monthly rotation among momentum, value, profitability, and defensive U.S. equity factors over a sample of about 60 years. Using a parsimonious set of state variables, the estimated policy outperforms both the myopic dynamic policy and the best fixed factor in hindsight in out-of-sample evaluation. Under weak conditions, we show theoretically that the decision rule estimated from empirical data is asymptotically consistent for the unknown optimal decision rule. Simulation results illustrate the finite-sample relevance of these asymptotic results.
We extend a positive Ricci curvature gluing theorem of Perelman to a range of positive intermediate curvature conditions, ranging from positive scalar curvature up to (and including) positive sectional curvature. As an application of this, we demonstrate that the observer moduli space of metrics with positive intermediate Ricci curvatures can have non-trivial higher homotopy groups. Further applications include deriving a sufficient condition for the existence of a metric with positive intermediate Ricci curvature and totally geodesic boundary.