arXiv subjects

Zhao Wang

Publications and source records attributed to Zhao Wang.

At least 181 records · Page 10Linked to original sources

arXiv2020

Machine-learning prediction of infrared spectra of interstellar polycyclic aromatic hydrocarbons

We design and train a neural network (NN) model to efficiently predict the infrared spectra of interstellar polycyclic aromatic hydrocarbons (PAHs) with a computational cost many orders of magnitude lower than what a first-principles calculation would demand. The input to the NN is based on the Morgan fingerprints extracted from the skeletal formulas of the molecules and does not require precise geometrical information such as interatomic distances. The model shows excellent predictive skill for out-of-sample inputs, making it suitable for improving the mixture models currently used for understanding the chemical composition and evolution of the interstellar medium. We also identify the constraints to its applicability caused by the limited diversity of the training data and estimate the prediction errors using a ensemble of NNs trained on subsets of the data. With help from other machine-learning methods like random forests, we dissect the role of different chemical features in this prediction. The power of these topological descriptors is demonstrated by the limited effect of including detailed geometrical information in the form of Coulomb matrix eigenvalues.

astro-ph.GA↗

arXiv2020

Identifying Spurious Correlations for Robust Text Classification

The predictions of text classifiers are often driven by spurious correlations -- e.g., the term `Spielberg' correlates with positively reviewed movies, even though the term itself does not semantically convey a positive sentiment. In this paper, we propose a method to distinguish spurious and genuine correlations in text classification. We treat this as a supervised classification problem, using features derived from treatment effect estimators to distinguish spurious correlations from "genuine" ones. Due to the generic nature of these features and their small dimensionality, we find that the approach works well even with limited training examples, and that it is possible to transport the word classifier to new domains. Experiments on four datasets (sentiment classification and toxicity detection) suggest that using this approach to inform feature selection also leads to more robust classification, as measured by improved worst-case accuracy on the samples affected by spurious correlations.

cs.LG↗

arXiv2020

Are Words Commensurate with Actions? Quantifying Commitment to a Cause from Online Public Messaging

Public entities such as companies and politicians increasingly use online social networks to communicate directly with their constituencies. Often, this public messaging is aimed at aligning the entity with a particular cause or issue, such as the environment or public health. However, as a consumer or voter, it can be difficult to assess an entity's true commitment to a cause based on public messaging. In this paper, we present a text classification approach to categorize a message according to its commitment level toward a cause. We then compare the volume of such messages with external ratings based on entities' actions (e.g., a politician's voting record with respect to the environment or a company's rating from environmental non-profits). We find that by distinguishing between low- and high- level commitment messages, we can more reliably identify truly committed entities. Furthermore, by measuring the discrepancy between classified messages and external ratings, we can identify entities whose public messaging does not align with their actions, thereby providing a methodology to identify potentially "inauthentic" messaging campaigns.

cs.CL↗

arXiv2020

Selective Conduction of Organic Molecules via Free-Standing Graphene

A race is held between ten species of organic gas molecules on a graphene substrate driven by thermal gradients via molecular dynamics. Fast conduction of the molecules is observed with selectivity for aromatic compounds. This selectivity stems from the fact that the planar structure of the aromatic molecule helps keep a shorter distance to the substrate, which is the key to the driving force at the gas-solid interface. The drift velocity monotonically increases with decreasing molecule density, with no ballistic transport observable even for a single molecule. A non-linear regime is discovered for the conduction of benzene molecules under large thermal gradients. At low temperature, molecules formed aggregation and move collectively along specific path in the graphene substrate.

cond-mat.mes-hall↗

arXiv2020

Chirality-Selective Transport of Benzene Molecules on Carbon Nanotubes

Using molecular dynamics simulations, we predict an effect of chirality on the conduction of benzene molecules along the surface of carbon nanotubes (CNTs) subjected to a thermal gradient. The group drift velocity of the molecules is found to be maximal in the case of an armchair CNT, and to decrease with decreasing chiral angle. This chirality effect on thermodiffusion is induced by a variation in the optimized paths of molecules that change with different electronic overlap at the interface. The mechanism for the thermophoretic transport is identified to be coupled with a gradient of adsorbate-substrate interaction energy, which originates from the anharmonic nature of the van der Waals potential.

cond-mat.mes-hall↗

arXiv2020

Formation of interstellar complex polycyclic aromatic hydrocarbons: Insights from molecular dynamics simulations of dehydrogenated benzene

Small organic molecules are thought to provide building blocks for the formation of complex interstellar polycyclic aromatic hydrocarbons (PAHs). However, the underlying chemical mechanisms remain unclear, particularly concerning the role of interstellar dust. Using molecular dynamics, we simulate the chemical reaction between dehydrogenated benzene molecules in the gas phase or on the surface of an onion-like carbon nanoparticle (NP). The reaction leads to the formation of PAHs of complex structures. The size of the formed molecules is found to roughly increase with increasing temperature up to 800 K, and to be correlated with the level of dehydrogenation. Morphology analysis features the formation of large rings that contain up to 32 carbon atom at high temperature. Density functional theory (DFT) calculations are performed to search the fundamental energetic reaction pathways. The DFT results quantitatively confirm the correlation between the reactivity and the dehydrogenation level, and the formation of stable C-8 rings. Moreover, the nanostructures formed on the NP surface point to a possible layer-by-layer formation mechanism for interstellar fullerene and carbon onions.

astro-ph.GA↗

arXiv2020

Contrastive Cross-site Learning with Redesigned Net for COVID-19 CT Classification

The pandemic of coronavirus disease 2019 (COVID-19) has lead to a global public health crisis spreading hundreds of countries. With the continuous growth of new infections, developing automated tools for COVID-19 identification with CT image is highly desired to assist the clinical diagnosis and reduce the tedious workload of image interpretation. To enlarge the datasets for developing machine learning methods, it is essentially helpful to aggregate the cases from different medical systems for learning robust and generalizable models. This paper proposes a novel joint learning framework to perform accurate COVID-19 identification by effectively learning with heterogeneous datasets with distribution discrepancy. We build a powerful backbone by redesigning the recently proposed COVID-Net in aspects of network architecture and learning strategy to improve the prediction accuracy and learning efficiency. On top of our improved backbone, we further explicitly tackle the cross-site domain shift by conducting separate feature normalization in latent space. Moreover, we propose to use a contrastive training objective to enhance the domain invariance of semantic embeddings for boosting the classification performance on each dataset. We develop and evaluate our method with two public large-scale COVID-19 diagnosis datasets made up of CT images. Extensive experiments show that our approach consistently improves the performances on both datasets, outperforming the original COVID-Net trained on each dataset by 12.16% and 14.23% in AUC respectively, also exceeding existing state-of-the-art multi-site learning methods.

eess.IV↗

arXiv2020

Neural Architecture Search on Acoustic Scene Classification

Convolutional neural networks are widely adopted in Acoustic Scene Classification (ASC) tasks, but they generally carry a heavy computational burden. In this work, we propose a lightweight yet high-performing baseline network inspired by MobileNetV2, which replaces square convolutional kernels with unidirectional ones to extract features alternately in temporal and frequency dimensions. Furthermore, we explore a dynamic architecture space built on the basis of the proposed baseline with the recent Neural Architecture Search (NAS) paradigm, which first trains a supernet that incorporates all candidate networks and then applies a well-known evolutionary algorithm NSGA-II to discover more efficient networks with higher accuracy and lower computational cost. Experimental results demonstrate that our searched network is competent in ASC tasks, which achieves 90.3% F1-score on the DCASE2018 task 5 evaluation set, marking a new state-of-the-art performance while saving 25% of FLOPs compared to our baseline network.

cs.SD↗

arXiv2020

Gallai-Ramsey number for the union of stars

Given a graph $G$ and a positive integer $k$, define the \emph{Gallai-Ramsey number} to be the minimum number of vertices $n$ such that any $k$-edge coloring of the complete graph $K_n$ contains either a rainbow (all different colored) triangle or a monochromatic copy of $G$. In this paper, we obtain the exact value of the Gallai-Ramsey numbers for the union of two stars in many cases and bounds in other cases. This work represents the first class of disconnected graphs to be considered as the desired monochromatic subgraph.

math.CO↗

arXiv2020

Design of a novel monolithic parabolic-mirror ion-trap to precisely align the RF null point with the optical focus

We propose a novel ion trap design with the high collection efficiency parabolic-mirror integrated with the ion trap electrodes. This design has three radio frequency (RF) electrodes and eight direct current(DC) compensation electrodes. By carefully adjusting three RF voltages, the parabolic mirror focus can be made precisely coincident with the RF null point. Thus, the aberration and the ion micromotion can be minimized at the same time. This monolithic design can significantly improve the ion-ion entanglement generation speed by extending the photon collecting solid angle beyond $90\%\cdot4π$. Further analysis of the trapping setup shows that the RF voltage variation method relexes machining accuracy to a broad range. This design is expected to be a robust scheme for trapping ion to speed entanglement network node.

quant-ph↗

arXiv2020

Gallai Ramsey number for double stars

Given a graph $G$ and a positive integer $k$, the \emph{Gallai-Ramsey number} is defined to be the minimum number of vertices $n$ such that any $k$-edge coloring of $K_n$ contains either a rainbow (all different colored) copy of $G$ or a monochromatic copy of $G$. In this paper, we obtain general upper and lower bounds on the Gallai-Ramsey numbers for double stars $S(n,m)$, where $S(n,m)$ is the graph obtained from the union of two stars $K_{1,n}$ and $K_{1,m}$ by adding an edge between their centers. We also provide the sharp result in some cases.

math.CO↗

arXiv2019

Chirality-Dependent Motion Transmission between Aligned Carbon Nanotubes

I demonstrate a directional motion-transmission behavior of aligned carbon nanotubes (CNTs) using atomistic simulations. The network of overlapping $π$ orbitals at the interface act as gear teeth to translate the sliding motion of a CNT into a rotating motion of the adjacent CNT, or \textit{viceversa}. The efficiency of this orthogonal motion transmission is found to strongly depend on the tube chirality, by which the interfacial stacking configuration of the atoms is determined. These results have strong implications on the design of the motion transmission system at the nanoscale.

cond-mat.mtrl-sci↗

arXiv2019

Gear Junctions between Chiral Boron Nitride Nanotubes

A gear effect is demonstrated at parallel and cross junctions between boron nitride nanotubes (BNNTs) via atomistic simulations. The atoms of neighboring BNNTs are meshed together at the junctions like gear teeth through long-range non-covalent interaction, which are shown to be able to transmit motion and power. The sliding motion of a BNNT can be spontaneously translated to rotating motion of an adjoining one or viceversa at a well-defined speed ratio. The transmittable motion and force strongly depend on the helical lattice structure of BNNTs represented by a chiral angle. The motion transmission efficiency of the parallel junctions increases up to a maximum for certain BNNTs depending on displacement rates. It then decreases with increasing chiral angles. For cross junctions, the angular motion transmission ratio increases with decreasing chiral angles of the driven BNNTs, while the translational one exhibits the opposite trend.

cond-mat.mtrl-sci↗

arXiv2019

Fractional matching preclusion number of graphs

The \emph{fractional matching preclusion number} of a graph $G$, denoted by $fmp(G)$, is the minimum number of edges whose deletion results in a graph that has no fractional perfect matchings. In this paper, we first give some sharp upper and lower bounds of fractional matching preclusion number. Next, graphs with large and small fractional matching preclusion number are characterized, respectively. In the end, we investigate some extremal problems on fractional matching preclusion number.

math.CO↗

arXiv2019

Ramsey and Gallai-Ramsey numbers for stars with extra independent edges

Given a graph $G$ and a positive integer $k$, define the \emph{Gallai-Ramsey number} to be the minimum number of vertices $n$ such that any $k$-edge coloring of $K_n$ contains either a rainbow (all different colored) triangle or a monochromatic copy of $G$. In this paper, we obtain general upper and lower bounds on the Gallai-Ramsey numbers for the graph $G = S_t^{r}$ obtained from a star of order $t$ by adding $r$ extra independent edges between leaves of the star so there are $r$ triangles and $t - 2r - 1$ pendent edges in $S_t^{r}$. We also prove some sharp results when $t = 2$.

math.CO↗

arXiv2019

Dramatic effect of a transverse electric field on frictional properties of graphene

We study the influence of transverse electric fields on the interfacial forces between a graphene layer and a carbon nanotube tip by means of atomistic simulations, in which a Gaussian regularized charge-dipole potential is combined with classical force fields. A significant effect of the field-induced electric charge on the normal force is observed. The normal pressure is found to be sensitive to the presence of a transverse electric field, while the friction force remains relatively invariant for the here-used field intensities. The contact can even be turned to have a negative coefficient of friction in a constant-distance scenario when the field strength reaches a critical value, which increases with decreasing tip-surface distance. These results shed light on how the frictional properties of nanomaterials can be controlled via applied electric fields.

cond-mat.mtrl-sci↗

arXiv2019

Ramsey and Gallai-Ramsey number for wheels

Given a graph $G$ and a positive integer $k$, define the \emph{Gallai-Ramsey number} to be the minimum number of vertices $n$ such that any $k$-edge coloring of $K_n$ contains either a rainbow (all different colored) triangle or a monochromatic copy of $G$. Much like graph Ramsey numbers, Gallai-Ramsey numbers have gained a reputation as being very difficult to compute in general. As yet, still only precious few sharp results are known. In this paper, we obtain bounds on the Gallai-Ramsey number for wheels and the exact value for the wheel on $5$ vertices.

math.CO↗

arXiv2019

On the $g$-good-neighbor connectivity of graphs

Connectivity and diagnosability are two important parameters for the fault tolerant of an interconnection network $G$. In 1996, Fàbrega and Fiol proposed the $g$-good-neighbor connectivity of $G$. In this paper, we show that $1\leq κ^g(G)\leq n-2g-2$ for $0\leq g\leq \left\{Δ(G),\left\lfloor \frac{n-3}{2}\right\rfloor\right\}$, and graphs with $κ^g(G)=1,2$ and trees with $κ^g(T_n)=n-t$ for $4\leq t\leq \frac{n+2}{2}$ are characterized, respectively. In the end, we get the three extremal results for the $g$-good-neighbor connectivity.

math.CO↗