Enhanced Differentiable Architecture Search Based on Asymptotic Regularization

In differentiable search architecture search methods,a more efficient search space design can significantly improve the performance of the searched architecture,thus requiring people to carefully define the search space with different complexity according to various operations.Meanwhile rationalizing the search strategies to explore the well-defined search space will further improve the speed and efficiency of architecture search.With this in mind,we propose a faster and more efficient differentiable architecture search method,AllegroNAS.Firstly,we introduce a more efficient search space enriched by the introduction of two redefined convolution modules.Secondly,we utilize a more efficient architectural parameter regularization method,mitigating the overfitting problem during the search process and reducing the error brought about by gradient approximation.Meanwhile,we introduce a natural exponential cosine annealing method to make the learning rate of the neural network training process more suitable for the search procedure.Moreover,group convolution and data augmentation are employed to reduce the computational cost.Finally,through extensive experiments on several public datasets,we demonstrate that our method can more swiftly search for better-performing neural network architectures in a more efficient search space,thus validating the effectiveness of our approach.

Calculus of Generalized Quasi-differentiable Functions I: Some Results on the Space of Pairs of Convex-Set Collections

A Riesz space K1 whose elements are pairs of convex-set collections is presented for the study on the calculus of generalized quasi-differentiable functions. The space K1 is constructed by introducing a well-defined equivalence relation among pairs of collections of convex sets. Some important properties on the norm and operations in K1 are given.

Sub-Differential Characterizations of Non-Smooth Lower Semi-Continuous Pseudo-Convex Functions on Real Banach Spaces

In this paper, we characterize lower semi-continuous pseudo-convex functions f : X → R ∪ {+ ∞} on convex subset of real Banach spaces K ⊂ X with respect to the pseudo-monotonicity of its Clarke-Rockafellar Sub-differential. We extend the results on the characterizations of non-smooth convex functions f : X → R ∪ {+ ∞} on convex subset of real Banach spaces K ⊂ X with respect to the monotonicity of its sub-differentials to the lower semi-continuous pseudo-convex functions on real Banach spaces.

Differentially private SGD with random features

In the realm of large-scale machine learning,it is crucial to explore methods for reducing computational complexity and memory demands while maintaining generalization performance.Additionally,since the collected data may contain some sensitive information,it is also of great significance to study privacy-preserving machine learning algorithms.This paper focuses on the performance of the differentially private stochastic gradient descent(SGD)algorithm based on random features.To begin,the algorithm maps the original data into a lowdimensional space,thereby avoiding the traditional kernel method for large-scale data storage requirement.Subsequently,the algorithm iteratively optimizes parameters using the stochastic gradient descent approach.Lastly,the output perturbation mechanism is employed to introduce random noise,ensuring algorithmic privacy.We prove that the proposed algorithm satisfies the differential privacy while achieving fast convergence rates under some mild conditions.

DIFFERENTIABILITY OF CONVEX FUNCTIONS AND ASPLUND SPACES

Characterizations of differentiability are obtained for continuous convex functions defined on nonempty open convex sets of Banach spaces as a generalization and application of a mumber of mathematicians several years effort, and a characteristic theorem is given for Banach spaces which are (weak) Asplund spaces.

Differentiability of Convex Functions in Locally Convex Spaces（Ⅰ）──Continuous Convex Functions and Gateaux Differentiability

ＤｉｆｆｅｒｅｎｔｉａｂｉｌｉｔｙｏｆＣｏｎｖｅｘＦｕｎｃｔｉｏｎｓｉｎＬｏｃａｌｌｙＣｏｎｖｅｘＳｐａｃｅｓ（Ⅰ）──ＣｏｎｔｉｎｕｏｕｓＣｏｎｖｅｘＦｕｎｃｔｉｏｎｓａｎｄＧａｔｅａｕｘＤｉｆｆｅｒｅｎｔｉａｂｉｌｉｔｙＷＵＣｏｎｇｘｉｎＣＨＥＮ...

ON THE PRODUCT OF GTEAUX DIFFERENTIABILITY LOCALLY CONVEX SPACES

A locally convex space is said to be a Gateaux differentiability space (GDS) provided every continuous convex function defined on a nonempty convex open subset D of the space is densely Gateaux differentiable in .D.This paper shows that the product of a GDS and a family of separable Prechet spaces is a GDS,and that the product of a GDS and an arbitrary locally convex space endowed with the weak topology is a GDS.

Differentiability of Convex Functions in Locally Convex Spaces(Ⅱ)──Frechet

ＤｉｆｆｅｒｅｎｔｉａｂｉｌｉｔｙｏｆＣｏｎｖｅｘＦｕｎｃｔｉｏｎｓｉｎＬｏｃａｌｌｙＣｏｎｖｅｘＳｐａｃｅｓ（ＩＩ）──ＦｒｅｃｈｅｔＤｉｆｆｅｒｅｎｔｉａｂｉｌｉｔｙＷＵＣｏｎｇｘｉｎＷＡＮＧＸｉａｏｍｉｎＣＨＥＮＧＬｉｘｉｎ吴从ａａａｙ，王晓敏...

THE BASIC THEORY OF SOLUTIONS OF FUNCTIONAL DIFFERENTIAL EQUATIONS WITH BOUNDED DELAY IN THE SPACE OF INTEGRABLE FUNCTIONS

在可积函数空间中讨论有界滞量泛函微分方程，建立其解的基本理论，包括解的存在性，唯一性及延展性。

DIFFERENTIAL INVERSE VARIATIONAL INEQUALITIES IN FINITE DIMENSIONAL SPACES

In this article, a new differential inverse variational inequality is introduced and studied in finite dimensional Euclidean spaces. Some results concerned with the linear growth of the solution set for the differential inverse variational inequalities are obtained under different conditions. Some existence theorems of Caratheodory weak solutions for the differential inverse variational inequality are also established under suitable conditions. An application to the time-dependent spatial price equilibrium control problem is also given.

ESSENTIAL NORMS OF PRODUCTS OF WEIGHTED COMPOSITION OPERATORS AND DIFFERENTIATION OPERATORS BETWEEN BANACH SPACES OF ANALYTIC FUNCTIONS

We obtain several estimates of the essential norms of the products of differen- tiation operators and weighted composition operators between weighted Banach spaces of analytic functions with general weights. As applications, we also give estimates of the es- sential norms of weighted composition operators between weighted Banach space of analytic functions and Bloch-type spaces.

A modified method for locating parapharyngeal space neoplasms on magnetic resonance images: implications for differential diagnosis

The parapharyngeal space(PPS) is an inverted pyramid-shaped deep space in the head and neck region, and a variety of tumors, such as salivary gland tumors, neurogenic tumors, nasopharyngeal carcinomas with parapharyngeal invasion, and lymphomas, can be found in this space. The differential diagnosis of PPS tumors remains challenging for radiologists. This study aimed to develop and test a modified method for locating PPS tumors on magnetic resonance(MR) images to improve preoperative differential diagnosis. The new protocol divided the PPS into three compartments: a prestyloid compartment, the carotid sheath, and the areas outside the carotid sheath. PPS tumors were located in these compartments according to the displacements of the tensor veli palatini muscle and the styloid process, with or without blood vessel separations and medial pterygoid invasion. This protocol, as well as a more conventional protocol that is based on displacements of the internal carotid artery(ICA), was used to assess MR images captured from a series of 58 PPS tumors. The consequent distributions of PPS tumor locations determined by both methods were compared. Of all 58 tumors, our new method determined that 57 could be assigned to precise PPS compartments. Nearly all(13/14; 93%) tumors that were located in the pre-styloid compartment were salivary gland tumors. All 15 tumors within the carotid sheath were neurogenic tumors. The vast majority(18/20; 90%) of trans-spatial lesions were malignancies. However, according to the ICA-based method, 28 tumors were located in the pre-styloid compartment, and 24 were located in the post-styloid compartment, leaving 6 tumors that were difficult to locate. Lesions located in both the pre-styloid and the post-styloid compartments comprised various types of tumors. Compared with the conventional ICA-based method, our new method can help radiologists to narrow the differential diagnosis of PPS tumors to specific compartments.

Boundedness for the Singular Integral with Variable Kernel and Fractional Differentiation on Weighted Morrey Spaces

Let T be the singular integral operator with variable kernel, T＊ be the adjoint of T and T# be the pseudo-adjoint of T. Let TIT2 be the product of T1 and T2, T1 o T2 be the pseudo product of T1 and T2. In this paper, we establish the boundedness for commutators of these operators and the fractional differentiation operator D^γ on the weighted Morrey spaces.

PRODUCTS OF WEIGHTED COMPOSITION AND DIFFERENTIATION OPERATORS INTO WEIGHTED ZYGMUND AND BLOCH SPACES

We characterize boundedness and compactness of products of differentiation op- erators and weighted composition operators between weighted Banach spaces of analytic functions and weighted Zygmund spaces or weighted Bloch spaces with general weights.

Numerical Methods for Solving Space Fractional Partial Differential Equations Using Hadamard Finite-Part Integral Approach

We introduce a novel numerical method for solving two-sided space fractional partial differential equations in two-dimensional case.The approximation of the space fractional Riemann-Liouville derivative is based on the approximation of the Hadamard finite-part integral which has the convergence order O(h^3-a),where h is the space step size and α∈(1,2)is the order of Riemann-Liouville fractional derivative.Based on this scheme,we introduce a shifted finite difference method for solving space fractional partial differential equations.We obtained the error estimates with the convergence orders O(τ+h^3-a+h^β),where τ is the time step size and β>0 is a parameter which measures the smoothness of the fractional derivatives of the solution of the equation.Unlike the numerical methods for solving space fractional partial differential equations constructed using the standard shifted Griinwald-Letnikov formula or higher order Lubich's methods which require the solution of the equation to satisfy the homogeneous Dirichlet boundary condition to get the firstorder convergence,the numerical method for solving the space fractional partial differential equation constructed using the Hadamard finite-part integral approach does not require the solution of the equation to satisfy the Dirichlet homogeneous boundary condition.Numerical results show that the experimentally determined convergence order obtained using the Hadamard finite-part integral approach for solving the space fractional partial differential equation with non-homogeneous Dirichlet boundary conditions is indeed higher than the convergence order obtained using the numerical methods constructed with the standard shifted Griinwald-Letnikov formula or Lubich's higher order approximation schemes.

Formal Similarity of Solutions in the Laplace Space on the Class of Fluid Flow Differential Equation

In this paper, the fluid flow differential equation based on the homogenous reservoirs model is first reviewed. Then a theorem about the formal similarity of solutions in the Laplace space with outer boundary conditions and inner boundary condition is presented and proved. Lastly, a corollary of our theorem is given particularly on inner boundary. The obtained results are very helpful for understanding inherent laws of relevant engineering science and designing practical analysis software.

PARAMETERS DETERMINATION METHOD OF PHASE-SPACE RECONSTRUCTION BASED ON DIFFERENTIAL ENTROPY RATIO AND RBF NEURAL NETWORK

Phase space reconstruction is the first step of recognizing the chaotic time series.On the basis of differential entropy ratio method,the embedding dimension opt m and time delay t are optimal for the state space reconstruction could be determined.But they are not the optimal parameters accepted for prediction.This study proposes an improved method based on the differential entropy ratio and Radial Basis Function(RBF)neural network to estimate the embedding dimension m and the time delay t,which have both optimal characteristics of the state space reconstruction and the prediction.Simulating experiments of Lorenz system and Doffing system show that the original phase space could be reconstructed from the time series effectively,and both the prediction accuracy and prediction length are improved greatly.

LDPC based time-frequency double differential space-time coding for multi-antenna OFDM systems

Differential space-time coding was proposed recently in the literature for multi-antenna systems, where neither the transmitter nor the receiver knows the fading coefficients. Among existing schemes, double differential space-time (DDST) coding is of special interest because it is applicable to continuous fast time-varying channels. However, it is less effective in fre- quency-selective fading channels. This paper’s authors derived a novel time-frequency double differential space-time (TF-DDST) coding scheme for multi-antenna orthogonal frequency division multiplexing (OFDM) systems in a time-varying fre- quency-selective fading environment, where double differential space-time coding is introduced into both time domain and fre- quency domain. Our proposed TF-DDST-OFDM system has a low-complexity non-coherent decoding scheme and is robust for time- and frequency-selective Rayleigh fading. In this paper, we also propose the use of state-of-the-art low-density parity-check (LDPC) code in serial concatenation with our TF-DDST scheme as a channel code. Simulations revealed that the LDPC based TF-DDST OFDM system has low decoding complexity and relatively better performance.

Periodic Solutions in UMD Spaces for Some Neutral Partial Functional Differential Equations

The aim of this work is to study the existence of a periodic solution for some neutral partial functional differential equations. Our approach is based on the R-boundedness of linear operators Lp-multipliers and UMD-spaces.

Boundary Value Problems for Second-order Integro-differential Equations in Banach Spaces

This paper investigates the maximal and minimal solutions of initial value problems for second order nonlinear integro-differential equations of Volterra type on a finite interval in a Banach space by establishing a comparison result and using the monotone iterative technique.