Source author record

Juan Meng

Juan Meng appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Machine Learning math.OC physics.app-ph

Catalog footprint

What is connected

3works

3topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Identifying optimal photovoltaic technologies for underwater applications

Improving solar energy collection in aquatic environments would allow for superior environmental monitoring and remote sensing, but the identification of optimal photovoltaic technologies for such applications is challenging as evaluation requires either field deployment or access to large water tanks. Here, we present a simple bench-top characterization technique that does not require direct access to water and therefore circumvents the need for field testing during initial trials of development. Employing LEDs to simulate underwater solar spectra at various depths, we compare Si and CdTe solar cells, two commercially available technologies, with GaInP cells, a technology with a wide band gap close to ideal for underwater solar harvesting. We use this method to show that while Si cells outperform both CdTe and GaInP under terrestrial AM1.5G solar irradiance, both CdTe and GaInP outperform Si at depths > 2 m, with GaInP cells operating with underwater efficiencies exceeding 51%.

preprint2020arXiv

Difference Attention Based Error Correction LSTM Model for Time Series Prediction

In this paper, we propose a novel model for time series prediction in which difference-attention LSTM model and error-correction LSTM model are respectively employed and combined in a cascade way. While difference-attention LSTM model introduces a difference feature to perform attention in traditional LSTM to focus on the obvious changes in time series. Error-correction LSTM model refines the prediction error of difference-attention LSTM model to further improve the prediction accuracy. Finally, we design a training strategy to jointly train the both models simultaneously. With additional difference features and new principle learning framework, our model can improve the prediction accuracy in time series. Experiments on various time series are conducted to demonstrate the effectiveness of our method.

preprint2017arXiv

Upper bounds for Z$_1$-eigenvalues of generalized Hilbert tensors

In this paper, we introduce the concept of Z$_1$-eigenvalue to infinite dimensional generalized Hilbert tensors (hypermatrix) $\mathcal{H}_λ^{\infty}=(\mathcal{H}_{i_{1}i_{2}\cdots i_{m}})$, $$ \mathcal{H}_{i_{1}i_{2}\cdots i_{m}}=\frac{1}{i_{1}+i_{2}+\cdots i_{m}+λ},\ λ\in \mathbb{R}\setminus\mathbb{Z}^-;\ i_{1},i_{2},\cdots,i_{m}=0,1,2,\cdots,n,\cdots, $$ and proved that its $Z_1$-spectral radius is not larger than $π$ for $λ>\frac{1}{2}$, and is at most $\fracπ{\sin{λπ}}$ for $\frac{1}{2}\geq λ>0$. Besides, the upper bound of $Z_1$-spectral radius of an $m$th-order $n$-dimensional generalized Hilbert tensor $\mathcal{H}_λ^n$ is obtained also, and such a bound only depends on $n$ and $λ$.