Sharif Digital Repository

Transferring large models in federated learning (FL) networks is often hindered by clients' limited bandwidth. We propose $ extsf {FedAA}$ , an FL algorithm which achieves fast convergence by exploiting the regularized Anderson acceleration (AA) on the global level. First, we demonstrate that FL can benefit from acceleration methods in numerical analysis. Second, $ extsf {FedAA}$ improves the convergence rate for quadratic losses and improves the empirical performance for smooth and strongly convex objectives, compared to FedAvg, an FL algorithm using gradient descent (GD) local updates. Experimental results demonstrate that employing AA can significantly improve the performance of FedAvg,... 

Micro and nanotechnology can potentially revolutionize drug delivery systems. Novel microfluidic systems have been employed for the cell culture applications and drug delivery by micro and nanocarriers. Cells in the microchannels are under static and dynamic flow perfusion of culture media that provides nutrition and removes waste from the cells. This exerts hydrostatic and hydrodynamic forces on the cells. These forces can considerably affect the functions of the living cells. In this paper, we simulated the flow of air, culture medium, and the particle transport and deposition in the microchannels under different angles of connection inlet. It was found that the shear stress induced by the... 

Thin porous membranes are important components in a microfluidic device, serving as separators, filters, and scaffolds for cell culture. However, the fabrication and the integration of these membranes possess many challenges, which restrict their widespread applications. This paper reports a facile technique to fabricate robust membrane-embedded microfluidic devices. We integrated an electrospun membrane into a polydimethylsiloxane (PDMS) device using the simple plasma-activated bonding technique. To increase the flexibility of the membrane and to address the leakage problem, the electrospun membrane was fabricated with the highest weight ratio of PDMS to polymethylmethacrylate (i.e., 6:1... 

Microfluidic devices have been widely used for biological and cellular studies. Microbioreactors for three-dimensional (3D) multicellular spheroid culture are now considered as the next generation in in vitro diagnostic tools. The feasibility of using 3D cell aggregates to form multicellular spheroids in a microbioreactor with U-shaped barriers has been demonstrated experimentally. A barrier array is an alternative to commonly used microwell traps. The present study investigates oxygen and glucose concentration distributions as key parameters in a U-shaped array microbioreactor using finite element simulation. The effect of spheroid diameter, inlet concentration and flow rate of the medium... 

Dielectrophoresis is a robust approach for manipulating bioparticles in microfluidic devices. In recent years, many groups have developed dielectrophoresis-based microfluidic systems for separation and concentration of various types of bioparticles, where the gradient of the electric field causes dielectrophoresis force acting on the suspended particles. Enhancing the gradient of the electric field with three-dimensional (3D) electrodes can significantly improve the efficiency of the system. Implementing planar electrodes in a 3D arrangement is a simple option to form a 3D-electrode configuration. This paper reports the development of a novel dielectrophoretic microfluidic system for... 

Inverse medium scattering solvers generally reconstruct a single solution without an associated measure of uncertainty. This is true both for the classical iterative solvers and for the emerging deep-learning methods. But ill-posedness and noise can make this single estimate inaccurate or misleading. While deep networks such as conditional normalizing flows can be used to sample posteriors in inverse problems, they often yield low-quality samples and uncertainty estimates. In this paper, we propose U-Flow, a Bayesian U-Net based on conditional normalizing flows, which generates high-quality posterior samples and estimates physically-meaningful uncertainty. We show that the proposed model... 

There is growing interest in applying distributed machine learning to edge computing, forming federated edge learning. Federated edge learning faces non-i.i.d. and heterogeneous data, and the communication between edge workers, possibly through distant locations and with unstable wireless networks, is more costly than their local computational overhead. In this work, we propose ${{sf DONE}}$DONE, a distributed approximate Newton-type algorithm with fast convergence rate for communication-efficient federated edge learning. First, with strongly convex and smooth loss functions, ${{sf DONE}}$DONE approximates the Newton direction in a distributed manner using the classical Richardson iteration... 

A microfluidic system provides an excellent platform for cellular studies. Most importantly, a three-dimensional (3D) cell culture model reconstructs more accurately the in vivo microenvironment of tissue. Accordingly, microfluidic 3D cell culture devices could be ideal candidates for in vitro cell culture platforms. In this paper, two types of 3D cellular aggregates, i.e., toroid and spheroid, are numerically studied. The studies are carried out for microfluidic systems containing U-shaped barrier as well as microwell structure. For the first time, we obtain oxygen and glucose concentration distributions inside a toroid aggregate as well as the shear stress on its surface and compare its... 

Microfluidic cell culture platforms are ideal candidates for modeling the native tumor microenvironment because they can precisely reconstruct in vivo cellular behavior. Moreover, mathematical modeling of tumor growth can pave the way toward description and prediction of growth pattern as well as improving cancer treatment. In this study, a modified mathematical model based on concentration distribution is applied to tumor growth in both conventional static culture and dynamic microfluidic cell culture systems. Apoptosis and necrosis mechanisms are considered as the main inhibitory factors in the model, while tumor growth rate and nutrient consumption rate are modified in both quiescent and... 

Concentration gradient generators (CGGs) help biologists to perform large scale, fast and high-throughput experiments. This paper introduces a design tool called Tree-like Concentration gradient generator Design Tool (TCDT). The performance of this tool is validated both numerically and experimentally. Three CGGs were fabricated using three different fabrication methods and design parameters. The performance of these devices was examined using the measurement of fluorescent and dye intensity. The performance of the design tool for non-linear and multi-drug concentration gradient generations was investigated as well. In addition, a method was developed to investigate the multi-drug... 

Tumor metastasis is responsible for the vast majority of cancer-associated morbidities and mortalities. Recent studies have disclosed the higher metastatic potential of circulating tumor cell (CTC) clusters than single CTCs. Despite long-term study on metastasis, the characterizations of its most potent cellular drivers, i.e., CTC clusters have only recently been investigated. The analysis of CTC clusters offers new intuitions into the mechanism of tumor metastasis and can lead to the development of cancer diagnosis and prognosis, drug screening, detection of gene mutations, and anti-metastatic therapeutics. In recent years, considerable attention has been dedicated to the development of... 

The present work is concerned with the steady incompressible flow through a parallel plate channel with stretching walls under an externally applied magnetic field. The governing continuity and Navier-Stokes equations are reduced to a fourth order nonlinear differential equation by using vorticity definition and similarity solution transformation. The obtained equations are solved by applying the analytical homotopy perturbation method (HPM). The method is called order of magnitude suggested for simplifying series solution to finite expression that is useful in engineering problems. The results are verified by comparing with numerical solutions and demonstrate a good accuracy of the obtained... 

Inertial microfluidics manipulates and separates microparticles based on the finite inertia of the fluid at high flow speed. In inertial microfluidics, modifying the geometry by embedding periodic micro-obstacles into curvilinear channels is emerging as a promising strategy to improve inertial focusing and separation. This work systematically investigated the influence of micro-obstacles on inertial focusing and developed a high-resolution microfluidic device for particle and cell separation. First, we developed numerical modelling to simulate the migration trajectories of particles. Then, we studied the effects of various obstacles on the inertial focusing in the sinusoidal channels. The... 

Let n≥2 be an integer. The graph G(n) is obtained by letting all the elements of {0,…,n−1} to be the vertices and defining distinct vertices x and y to be adjacent if and only if gcd⁡(x+y,n)=1. In this paper, well-coveredness, Cohen–Macaulayness, vertex-decomposability and Gorensteinness of these graphs and their complements are characterized. These characterizations provide large classes of Cohen–Macaulay and non Cohen–Macaulay graphs. © 2021 Elsevier B.V  

Ground reaction force (GRF) characteristics of amputee walking are important for the analysis of clinical gait data, and also to update model reference adaptive impedance (MRAI) controllers. GRF estimation is a better alternative than direct GRF measurement because of the disadvantages of load cells, such as high cost, integration difficulties due to weight and physical dimensions, the possibility of overload, and measurement noise. This paper presents four robust MRAI observer/controller combinations for GRF estimation-based control of a prosthesis and a legged robot model in the presence of parametric uncertainties and unmodeled dynamics, in which the robot model is employed to mimic... 

Ground reaction force (GRF) characteristics of amputee walking are important for the analysis of clinical gait data, and also to update model reference adaptive impedance (MRAI) controllers. GRF estimation is a better alternative than direct GRF measurement because of the disadvantages of load cells, such as high cost, integration difficulties due to weight and physical dimensions, the possibility of overload, and measurement noise. This paper presents four robust MRAI observer/controller combinations for GRF estimation-based control of a prosthesis and a legged robot model in the presence of parametric uncertainties and unmodeled dynamics, in which the robot model is employed to mimic... 

The fresh groundwater lenses (FGLs) of small islands can be highly vulnerable to climate change impacts, including sea-level rise (SLR). Many real cases of atoll or sandy islands involve two-layer hydrogeological conceptualizations. In this paper, the influential factors that affect FGLs in two-layer small islands subject to SLR are investigated. An analytical solution describing FGLs in circular islands, composed of two geological layers, is developed for the simplified case of steady-state and sharp-interface conditions. An application of the developed model is demonstrated to estimate the FGL thickness of some real-world islands by comparison with existing FGL thickness data. Furthermore,... 

A cell spheroid is a three-dimensional (3D) aggregation of cells. Synthetic, in-vitro spheroids provide similar metabolism, proliferation, and species concentration gradients to those found in-vivo. For instance, cancer cell spheroids have been demonstrated to mimic in-vivo tumor microenvironments, and are thus suitable for in-vitro drug screening. The first part of this paper discusses the latest microfluidic designs for spheroid formation and culture, comparing their strategies and efficacy. The most recent microfluidic techniques for spheroid formation utilize emulsion, microwells, U-shaped microstructures, or digital microfluidics. The engineering aspects underpinning spheroid formation... 

The uncertainty and the complexity associated with the domino effect is a barrier to assessing the frequency of such accidents analytically. The use of simulation techniques, such as Monte Carlo, to examine the domino effect instead of analytical techniques has shown great promise. In this paper, a new method to assess the frequency of domino accidents is proposed-FREEDOM II-which is an improvement on the recent algorithm proposed by the authors (FREEDOM). The modifications on FREEDOM were carried out to overcome a limitation of the method and to extend its capabilities. A key shortcoming of the earlier method was its inability to handle multiple failure scenarios. This shortcoming has been... 

In this paper, we present a one-way relaying scheme in which two wireless nodes create an information flow to each other via a single decode-and-forward (DF) relay. We consider an additional secrecy constraint for protection against an honest-but-curious relay. Indeed, while the relay should decode the source message, it should be fully ignorant about the message content. We provide a secure lattice coding strategy based on random difference families (RDF) lattice codes for unidirectional Gaussian relay channels. RDF lattice codes are carved from infinite RDF lattices using a shaping algorithm. By RDF lattice we mean a Construction A lattice with a QC-LDPC code, which is obtained from random...