Stevens Institute of Technology

About: Stevens Institute of Technology is a education organization based out in Hoboken, New Jersey, United States. It is known for research contribution in the topics: Computer science & Cognitive radio. The organization has 5440 authors who have published 12684 publications receiving 296875 citations. The organization is also known as: Stevens & Stevens Tech.

Gas absorption studies in microporous hollow fiber membrane modules

Abstract: A comprehensive experimental investigation of gas-liquid absorption in a shell-and-tube type microporous hydrophobic hollow fiber device in a parallel flow configuration was carried out. Two modes of countercurrent gas-liquid contacting were studied, the wetted mode (absorbent liquid filled pores) and the nonwetted mode (gas-filled pores). The absorbent flowed through the fiber bore in most of the experiments. The systems studied include pure CO[sub 2], pure SO[sub 2], CO[sub 2] - N[sub 2] mixtures and SO[sub 2] - air mixtures. The absorbent was pure water. The absorption process was simulated for each case with a numerical model for species transport with and without chemical reaction. Laminar parabolic velocity profile was used for the tube-side flow, and Happel's free surface model was used to characterize the shell-side flow. The model simulations agreed well with the experimental observations in most cases. SO[sub 2] removals as high as 99% were obtained in small compact contactors. High K[sub L]a and low height of transfer unit (HTU) values were obtained with hollow fiber contactors when compared to those of conventional contactors. The applications of direct interest here are those for acid gas cleanup.

Horizontal Pyramid Matching for Person Re-Identification

Abstract: Despite the remarkable progress in person re-identification (Re-ID), such approaches still suffer from the failure cases where the discriminative body parts are missing. To mitigate this type of failure, we propose a simple yet effective Horizontal Pyramid Matching (HPM) approach to fully exploit various partial information of a given person, so that correct person candidates can be identified even if some key parts are missing. With HPM, we make the following contributions to produce more robust feature representations for the Re-ID task: 1) we learn to classify using partial feature representations at different horizontal pyramid scales, which successfully enhance the discriminative capabilities of various person parts; 2) we exploit average and max pooling strategies to account for person-specific discriminative information in a global-local manner. To validate the effectiveness of our proposed HPM method, extensive experiments are conducted on three popular datasets including Market-1501, DukeMTMCReID and CUHK03. Respectively, we achieve mAP scores of 83.1%, 74.5% and 59.7% on these challenging benchmarks, which are the new state-of-the-arts.

On the Convergence of FedAvg on Non-IID Data

Abstract: Federated learning enables a large amount of edge computing devices to jointly learn a model without data sharing. As a leading algorithm in this setting, Federated Averaging (\texttt{FedAvg}) runs Stochastic Gradient Descent (SGD) in parallel on a small subset of the total devices and averages the sequences only once in a while. Despite its simplicity, it lacks theoretical guarantees under realistic settings. In this paper, we analyze the convergence of \texttt{FedAvg} on non-iid data and establish a convergence rate of $\mathcal{O}(\frac{1}{T})$ for strongly convex and smooth problems, where $T$ is the number of SGDs. Importantly, our bound demonstrates a trade-off between communication-efficiency and convergence rate. As user devices may be disconnected from the server, we relax the assumption of full device participation to partial device participation and study different averaging schemes; low device participation rate can be achieved without severely slowing down the learning. Our results indicate that heterogeneity of data slows down the convergence, which matches empirical observations. Furthermore, we provide a necessary condition for \texttt{FedAvg} on non-iid data: the learning rate $\eta$ must decay, even if full-gradient is used; otherwise, the solution will be $\Omega (\eta)$ away from the optimal.