Multiscale Spatio-Temporal Graph Neural Networks for 3D Skeleton-Based Motion Prediction

We propose a multiscale spatio-temporal graph neural network (MST-GNN) to predict the future 3D skeleton-based human poses in an action-category-agnostic manner. The core of MST-GNN is a multiscale spatio-temporal graph that explicitly models the relations in motions at various spatial and temporal scales. Different from many previous hierarchical structures, our multiscale spatio-temporal graph is built in a data-adaptive fashion, which captures nonphysical, yet motion-based relations. The key module of MST-GNN is a multiscale spatio-temporal graph computational unit (MST-GCU) based on the trainable graph structure. MST-GCU embeds underlying features at individual scales and then fuses features across scales to obtain a comprehensive representation. The overall architecture of MST-GNN follows an encoder-decoder framework, where the encoder consists of a sequence of MST-GCUs to learn the spatial and temporal features of motions, and the decoder uses a graph-based attention gate recurrent unit (GA-GRU) to generate future poses. Extensive experiments are conducted to show that the proposed MST-GNN outperforms state-of-the-art methods in both short and long-term motion prediction on the datasets of Human 3.6M, CMU Mocap and 3DPW, where MST-GNN outperforms previous works by 5.33% and 3.67% of mean angle errors in average for short-term and long-term prediction on Human 3.6M, and by 11.84% and 4.71% of mean angle errors for short-term and long-term prediction on CMU Mocap, and by 1.13% of mean angle errors on 3DPW in average, respectively. We further investigate the learned multiscale graphs for interpretability.

An Efficient Egocentric Regulator for Continuous Targeting Problems of the Underactuated Quadrotor

Flying robots such as the quadrotor could provide an efficient approach for medical treatment or sensor placing of wild animals. In these applications, continuously targeting the moving animal is a crucial requirement. Due to the underactuated characteristics of the quadrotor and the coupled kinematics with the animal, nonlinear optimal tracking approaches, other than smooth feedback control, are required. However, with severe nonlinearities, it would be time-consuming to evaluate control inputs, and real-time tracking may not be achieved with generic optimizers onboard. To tackle this problem, a novel efficient egocentric regulation approach with high computational efficiency is proposed in this paper. Specifically, it directly formulates the optimal tracking problem in an egocentric manner regarding the quadrotors body coordinates. Meanwhile, the nonlinearities of the system are peeled off through a mapping of the feedback states as well as control inputs, between the inertial and body coordinates. In this way, the proposed efficient egocentric regulator only requires solving a quadratic performance objective with linear constraints and then generate control inputs analytically. Comparative simulations and mimic biological experiment are carried out to verify the effectiveness and computational efficiency. Results demonstrate that the proposed control approach presents the highest and stablest computational efficiency than generic optimizers on different platforms. Particularly, on a commonly utilized onboard computer, our method can compute the control action in approximately 0.3 ms, which is on the order of 350 times faster than that of generic nonlinear optimizers, establishing a control frequency around 3000 Hz.

Euler sums of generalized alternating hyperharmonic numbers II

In this paper, we introduce a new type of generalized alternating hyperharmonic numbers $H_n^{(p,r,s_{1},s_{2})}$, and show that Euler sums of the generalized alternating hyperharmonic numbers $H_n^{(p,r,s_{1},s_{2})}$ can be expressed in terms of linear combinations of classical (alternating) Euler sums.

Optical Brewster metasurfaces exhibiting ultra-broadband reflectionless absorption and extreme angular-asymmetry

Impedance mismatch between free space and absorptive materials is a fundamental issue plaguing the pursue of high-efficiency light absorption. In this work, we design and numerically demonstrate a type of non-resonant impedance-matched optical metasurfaces exhibiting ultra-broadband reflectionless absorption based on anomalous Brewster effect, which are donated as optical Brewster metasurfaces here. Interestingly, such Brewster metasurfaces exhibit a unique type of extreme angular-asymmetry: a transition between perfect transparency and perfect absorption appears when the sign of the incident angle is changed. Such a remarkable phenomenon originates in the coexistence of traditional and anomalous Brewster effects. Guidelines of material selection based on an effective-medium description and strategies such as the integration of a metal back-reflector or folded metasurfaces are proposed to improve the absorption performance. Finally, a gradient optical Brewster metasurface exhibiting ultra-broadband and near-omnidirectional reflectionless absorption is demonstrated. Such high-efficiency asymmetric optical metasurfaces may find applications in optoelectrical and thermal devices like photodetectors, thermal emitters and photovoltaics.

Online Multi-Agent Forecasting with Interpretable Collaborative Graph Neural Network

This paper considers predicting future statuses of multiple agents in an online fashion by exploiting dynamic interactions in the system. We propose a novel collaborative prediction unit (CoPU), which aggregates the predictions from multiple collaborative predictors according to a collaborative graph. Each collaborative predictor is trained to predict the status of an agent by considering the impact of another agent. The edge weights of the collaborative graph reflect the importance of each predictor. The collaborative graph is adjusted online by multiplicative update, which can be motivated by minimizing an explicit objective. With this objective, we also conduct regret analysis to indicate that, along with training, our CoPU achieves similar performance with the best individual collaborative predictor in hindsight. This theoretical interpretability distinguishes our method from many other graph networks. To progressively refine predictions, multiple CoPUs are stacked to form a collaborative graph neural network. Extensive experiments are conducted on three tasks: online simulated trajectory prediction, online human motion prediction and online traffic speed prediction, and our methods outperform state-of-the-art works on the three tasks by 28.6%, 17.4% and 21.0% on average, respectively.

Strain-induced topological phase transition in two-dimensional platinum ditelluride

Topological phase transition is a hot topic in condensed matter physics and computational material science. Here, we investigate the electronic structure and phonon dispersion of the two-dimensional (2D) platinum ditelluride ($PtTe_2$) using the density functional theory. It is found that the $PtTe_2$ monolayer is a trivial insulator with an indirect band gap of 0.347eV. Based on parity analysis, the biaxial tensile strain can drive the topological phase transition. As the strain reaches 19.3%, $PtTe_2$ undergoes a topological phase transition, which changes from a trivial band insulator to a topological insulator with $Z_2=1$. Unlike conventional honeycomb 2D materials with topological phase transition, which gap closes at K points, the strained $PtTe_2$ monolayer becomes gapless at M points under critical biaxial strain. The band inversion leads the switch of the parities near the Fermi level, which gives rise to the topological phase transition. The novel monolayer $PtTe_2$ has a potential application in the field of micro-electronics.

Twenty-three millisecond electron spin coherence of erbium ions in a natural-abundance crystal

Erbium ions doped into crystals have unique properties for quantum information processing, because of their optical transition at 1.5 $mu$m and of the large magnetic moment of their effective spin-1/2 electronic ground state. Most applications of erbium require however long electron spin coherence times, and this has so far been missing. Here, by selecting a host matrix with a low nuclear-spin density (CaWO$_4$) and by quenching the spectral diffusion due to residual paramagnetic impurities at millikelvin temperatures, we obtain an Er$^{3+}$ electron spin coherence time of 23 ms. This is the longest electron spin coherence time measured in a material with a natural abundance of nuclear spins and on a magnetically-sensitive transition. Our results establish Er$^{3+}$:CaWO$_4$ as a leading platform for quantum networks.

Reliable Graph Neural Network Explanations Through Adversarial Training

Graph neural network (GNN) explanations have largely been facilitated through post-hoc introspection. While this has been deemed successful, many post-hoc explanation methods have been shown to fail in capturing a models learned representation. Due to this problem, it is worthwhile to consider how one might train a model so that it is more amenable to post-hoc analysis. Given the success of adversarial training in the computer vision domain to train models with more reliable representations, we propose a similar training paradigm for GNNs and analyze the respective impact on a models explanations. In instances without ground truth labels, we also determine how well an explanation method is utilizing a models learned representation through a new metric and demonstrate adversarial training can help better extract domain-relevant insights in chemistry.

Embedding-based Product Retrieval in Taobao Search

Nowadays, the product search service of e-commerce platforms has become a vital shopping channel in peoples life. The retrieval phase of products determines the search systems quality and gradually attracts researchers attention. Retrieving the most relevant products from a large-scale corpus while preserving personalized user characteristics remains an open question. Recent approaches in this domain have mainly focused on embedding-based retrieval (EBR) systems. However, after a long period of practice on Taobao, we find that the performance of the EBR system is dramatically degraded due to its: (1) low relevance with a given query and (2) discrepancy between the training and inference phases. Therefore, we propose a novel and practical embedding-based product retrieval model, named Multi-Grained Deep Semantic Product Retrieval (MGDSPR). Specifically, we first identify the inconsistency between the training and inference stages, and then use the softmax cross-entropy loss as the training objective, which achieves better performance and faster convergence. Two efficient methods are further proposed to improve retrieval relevance, including smoothing noisy training data and generating relevance-improving hard negative samples without requiring extra knowledge and training procedures. We evaluate MGDSPR on Taobao Product Search with significant metrics gains observed in offline experiments and online A/B tests. MGDSPR has been successfully deployed to the existing multi-channel retrieval system in Taobao Search. We also introduce the online deployment scheme and share practical lessons of our retrieval system to contribute to the community.

Comprehensive Quality Investigations of Wire-feed Additive Manufacturing by Learning of Experimental Data

Wire-feed laser additive manufacturing is an emerging fabrication technique capable of highly automated large-scale volume production that can reduce both material waste and overall cost while improving product lead times. Quality assurance is necessary for implementation into critical structural applications. However, the large number of process variables along with the cost associated with traditional trial and error methods makes this difficult. This study investigates a comprehensive quality framework based on learning from experimental data that will enable improved quality control along with consistent microstructural features of the part. Specifically, a comprehensive experimental data across multiple process variables and output characteristics in terms of overall bead quality, geometric shape (i.g. bead height, width, fusion zone depth, etc.), and microstructural features are collected. The predicted process-geometry-microstructure relations are then captured by virtue of data-driven machine learning models. The properties of printed beads are visualized based on an extensive range of processing space within a 3-dimensional contour map. The insights and impacts of process variables on bead morphology, geometric and microstructural features are comprehensively investigated for quality improvement during manufacturing processes.