arXiv subjects

Qing Li

Publications and source records attributed to Qing Li.

At least 487 records · Page 27Linked to original sources

arXiv2022

Stock Movement Prediction Based on Bi-typed Hybrid-relational Market Knowledge Graph via Dual Attention Networks

Stock Movement Prediction (SMP) aims at predicting listed companies' stock future price trend, which is a challenging task due to the volatile nature of financial markets. Recent financial studies show that the momentum spillover effect plays a significant role in stock fluctuation. However, previous studies typically only learn the simple connection information among related companies, which inevitably fail to model complex relations of listed companies in the real financial market. To address this issue, we first construct a more comprehensive Market Knowledge Graph (MKG) which contains bi-typed entities including listed companies and their associated executives, and hybrid-relations including the explicit relations and implicit relations. Afterward, we propose DanSmp, a novel Dual Attention Networks to learn the momentum spillover signals based upon the constructed MKG for stock prediction. The empirical experiments on our constructed datasets against nine SOTA baselines demonstrate that the proposed DanSmp is capable of improving stock prediction with the constructed MKG.

q-fin.ST↗

arXiv2022

Conceptor Learning for Class Activation Mapping

Class Activation Mapping (CAM) has been widely adopted to generate saliency maps which provides visual explanations for deep neural networks (DNNs). The saliency maps are conventionally generated by fusing the channels of the target feature map using a weighted average scheme. It is a weak model for the inter-channel relation, in the sense that it only models the relation among channels in a contrastive way (i.e., channels that play key roles in the prediction are given higher weights for them to stand out in the fusion). The collaborative relation, which makes the channels work together to provide cross reference, has been ignored. Furthermore, the model has neglected the intra-channel relation thoroughly.In this paper, we address this problem by introducing Conceptor learning into CAM generation. Conceptor leaning has been originally proposed to model the patterns of state changes in recurrent neural networks (RNNs). By relaxing the dependency of Conceptor learning to RNNs, we make Conceptor-CAM not only generalizable to more DNN architectures but also able to learn both the inter- and intra-channel relations for better saliency map generation. Moreover, we have enabled the use of Boolean operations to combine the positive and pseudo-negative evidences, which has made the CAM inference more robust and comprehensive. The effectiveness of Conceptor-CAM has been validated with both formal verifications and experiments on the dataset of the largest scale in literature. The experimental results show that Conceptor-CAM is compatible with and can bring significant improvement to all well recognized CAM-based methods, and has outperformed the state-of-the-art methods by 43.14%~72.79% (88.39%~168.15%) on ILSVRC2012 in Average Increase (Drop), 15.42%~42.55% (47.09%~372.09%) on VOC, and 17.43%~31.32% (47.54%~206.45%) on COCO, respectively.

cs.CV↗

arXiv2022

Improved three-dimensional thermal multiphase lattice Boltzmann model for liquid-vapor phase change

Modeling liquid-vapor phase change using the lattice Boltzmann (LB) method has attracted significant attention in recent years. In this paper, we propose an improved three-dimensional (3D) thermal multiphase LB model for simulating liquid-vapor phase change. The proposed model has the following features. First, it is still within the framework of the thermal LB method using a temperature distribution function and therefore retains the fundamental advantages of the thermal LB method. Second, in the existing thermal LB models for liquid-vapor phase change, the finite-difference computations of the gradient terms $\nabla \cdot u$ and $\nabla T$ usually require special treatment at boundary nodes, while in the proposed thermal LB model these two terms are calculated locally. Moreover, in some of the existing thermal LB models, the error term ${\partial _{t_0}}\left( {Tu} \right)$ is eliminated by adding local correction terms to the collision process in the moment space, which causes these thermal LB models to be limited to the D2Q9 lattice in two dimensions and the D3Q15 or D3Q19 lattice in three dimensions. Conversely, the proposed model does not suffer from such an error term and therefore the thermal LB equation can be constructed on the D3Q7 lattice, which simplifies the model and improves the computational efficiency. Numerical simulations are carried out to validate the accuracy and efficiency of the proposed thermal multiphase LB model for simulating liquid-vapor phase change.

physics.comp-ph↗

arXiv2022

Hexagonal Boron Nitride (hBN) as a Low-loss Dielectric for Superconducting Quantum Circuits and Qubits

Dielectrics with low loss at microwave frequencies are imperative for high-coherence solid-state quantum computing platforms. We study the dielectric loss of hexagonal boron nitride (hBN) thin films in the microwave regime by measuring the quality factor of parallel-plate capacitors (PPCs) made of NbSe$_{2}$-hBN-NbSe$_{2}$ heterostructures integrated into superconducting circuits. The extracted microwave loss tangent of hBN is bounded to be at most in the mid-10$^{-6}$ range in the low temperature, single-photon regime. We integrate hBN PPCs with aluminum Josephson junctions to realize transmon qubits with coherence times reaching 25 $μ$s, consistent with the hBN loss tangent inferred from resonator measurements. The hBN PPC reduces the qubit feature size by approximately two-orders of magnitude compared to conventional all-aluminum coplanar transmons. Our results establish hBN as a promising dielectric for building high-coherence quantum circuits with substantially reduced footprint and, with a high energy participation that helps to reduce unwanted qubit cross-talk.

cond-mat.mes-hall↗

arXiv2021

Tiansuan Constellation: An Open Research Platform

Satellite network is the first step of interstellar voyages. It can provide global Internet connectivity everywhere on earth, where most areas cannot access the Internet by the terrestrial infrastructure due to the geographic accessibility and high cost. The space industry experiences a rise in large low-earth-orbit satellite constellations to achieve universal connectivity. The research community is also urgent to do some leading research to bridge the connectivity divide. Researchers now conduct their work by simulation, which is far from enough. However, experiments on real satellites are blocked by the high threshold of space technology, such as deployment cost and unknown risks. To solve the above dilemma, we are eager to contribute to the universal connectivity and build an open research platform, Tiansuan constellation to support experiments on real satellite networks. We discuss the potential research topics that would benefit from Tiansuan constellation. We provide two case studies that have already deployed in two experimental satellites of Tiansuan constellation.

cs.DC↗

arXiv2021

Towards a Unified Foundation Model: Jointly Pre-Training Transformers on Unpaired Images and Text

In this paper, we explore the possibility of building a unified foundation model that can be adapted to both vision-only and text-only tasks. Starting from BERT and ViT, we design a unified transformer consisting of modality-specific tokenizers, a shared transformer encoder, and task-specific output heads. To efficiently pre-train the proposed model jointly on unpaired images and text, we propose two novel techniques: (i) We employ the separately-trained BERT and ViT models as teachers and apply knowledge distillation to provide additional, accurate supervision signals for the joint training; (ii) We propose a novel gradient masking strategy to balance the parameter updates from the image and text pre-training losses. We evaluate the jointly pre-trained transformer by fine-tuning it on image classification tasks and natural language understanding tasks, respectively. The experiments show that the resultant unified foundation transformer works surprisingly well on both the vision-only and text-only tasks, and the proposed knowledge distillation and gradient masking strategy can effectively lift the performance to approach the level of separately-trained models.

cs.CV↗

arXiv2021

The ergodic and non-ergodic phases in one dimensional clean Jaynes-Cummings-Hubbard system

We study the ergodic and non-ergodic behaviors of a clean Jaynes-Cummings-Hubbard chain for different parameters based on the average level spacings and the generalized fractal dimensions of eigenstates by using exact diagonalization. It can be found that a transition from ergodicity to non-ergodicity phases happens when the atom-photon detuning is large, and the non-ergodic phases maybe exist in the thermodynamic limit. We also find that the non-ergodic phase violates the eigenstate thermalization hypothesis. Finally, we study the many-body multifractality of the ground state and find that the derivative of the generalized fractal dimensions can determine the critical point of the Superfluid-Mott-insulation phase transition in a small range of parameters under different boundary conditions and there is no ergodicity for the ground state.

quant-ph↗

arXiv2021

Octave-spanning microcomb generation in 4H-silicon-carbide-on-insulator photonics platform

Silicon carbide has recently emerged as a promising photonics material due to its unique properties, including possessing strong second- and third-order nonlinear coefficients and hosting various color centers that can be utilized for a wealth of quantum applications. Here, we report the design and demonstration of octave-spanning microcombs in a 4H-silicon-carbide-on-insulator platform for the first time. Such broadband operation is enabled by optimized nanofabrication achieving >1 million intrinsic quality factors in a 36-$μ$m-radius microring resonator, and careful dispersion engineering by investigating the dispersion properties of different mode families. For example, for the fundamental transverse-electric mode whose dispersion can be tailored by simply varying the microring waveguide width, we realized a microcomb spectrum covering the wavelength range from 1100 nm to 2400 nm with an on-chip power near 120 mW. While the observed comb state is verified to be chaotic and not soliton, attaining such a large bandwidth is a crucial step towards realizing $f$-2$f$ self-referencing. In addition, we have also observed coherent soliton-crystal state for the fundamental transverse-magnetic mode, which exhibits stronger dispersion than the fundamental transverse-electric mode and hence a narrower bandwidth.

physics.optics↗

arXiv2021

No observation of chiral flux current in the topological kagome metal CsV$_{3}$Sb$_{5}$

Compounds with kagome lattice usually host many exotic quantum states, including the quantum spin liquid, non-trivial topological Dirac bands and a strongly renormalized flat band, etc. Recently an interesting vanadium based kagome family $A$V$_{3}$Sb$_{5}$ ($A$ = K, Rb, or Cs) was discovered, and these materials exhibit multiple interesting properties, including unconventional saddle-point driving charge density wave (CDW) state, superconductivity, etc. Furthermore, some experiments show anomalous Hall effect which inspires that there might be some chiral flux current states. Here we report scanning tunneling measurements by using spin-polarized tips. Although we have observed clearly the $2a_0\times2a_0$ CDW and $4a_0$ stripe orders, the well-designed experiments with refined spin-polarized tips do not reveal any trace of the chiral flux current phase in CsV$_3$Sb$_5$ within the limits of experimental accuracy. No observation of the local magnetic moment in our experiments may put an upper bound constraint on the magnitude of magnetic moments induced by the possible chiral loop current which has a time-reversal symmetry breaking along $c$-axis in CsV$_{3}$Sb$_{5}$.

cond-mat.str-el↗

arXiv2021

Incorporating Surrogate Information for Adaptive Subgroup Enrichment Design with Sample Size Re-estimation

Adaptive subgroup enrichment design is an efficient design framework that allows accelerated development for investigational treatments while also having flexibility in population selection within the course of the trial. The adaptive decision at the interim analysis is commonly made based on the conditional probability of trial success. However, one of the critical challenges for such adaptive designs is immature data for interim decisions, particularly in the targeted subgroup with a limited sample size at the first stage of the trial. In this paper, we improve the interim decision making by incorporating information from surrogate endpoints when estimating conditional power at the interim analysis, by predicting the primary treatment effect based on the observed surrogate endpoint and prior knowledge or historical data about the relationship between endpoints. Modified conditional power is developed for both selecting the patient population to be enrolled after the interim analysis and sample size re-estimation. In the simulation study, our proposed design shows a higher chance to make desirable interim decisions and achieves higher overall power, while controlling the overall type I error. This performance is robust over drift of prior knowledge from the true relationship between two endpoints. We also demonstrate the application of our proposed design in two case studies in oncology and vaccine trials.

stat.ME↗

arXiv2021

Deep Keyphrase Completion

Keyphrase provides accurate information of document content that is highly compact, concise, full of meanings, and widely used for discourse comprehension, organization, and text retrieval. Though previous studies have made substantial efforts for automated keyphrase extraction and generation, surprisingly, few studies have been made for \textit{keyphrase completion} (KPC). KPC aims to generate more keyphrases for document (e.g. scientific publication) taking advantage of document content along with a very limited number of known keyphrases, which can be applied to improve text indexing system, etc. In this paper, we propose a novel KPC method with an encoder-decoder framework. We name it \textit{deep keyphrase completion} (DKPC) since it attempts to capture the deep semantic meaning of the document content together with known keyphrases via a deep learning framework. Specifically, the encoder and the decoder in DKPC play different roles to make full use of the known keyphrases. The former considers the keyphrase-guiding factors, which aggregates information of known keyphrases into context. On the contrary, the latter considers the keyphrase-inhibited factor to inhibit semantically repeated keyphrase generation. Extensive experiments on benchmark datasets demonstrate the efficacy of our proposed model.

cs.IR↗

arXiv2021

Benchmark Problems for CEC2021 Competition on Evolutionary Transfer Multiobjectve Optimization

Evolutionary transfer multiobjective optimization (ETMO) has been becoming a hot research topic in the field of evolutionary computation, which is based on the fact that knowledge learning and transfer across the related optimization exercises can improve the efficiency of others. Besides, the potential for transfer optimization is deemed invaluable from the standpoint of human-like problem-solving capabilities where knowledge gather and reuse are instinctive. To promote the research on ETMO, benchmark problems are of great importance to ETMO algorithm analysis, which helps designers or practitioners to understand the merit and demerit better of ETMO algorithms. Therefore, a total number of 40 benchmark functions are proposed in this report, covering diverse types and properties in the case of knowledge transfer, such as various formulation models, various PS geometries and PF shapes, large-scale of variables, dynamically changed environment, and so on. All the benchmark functions have been implemented in JAVA code, which can be downloaded on the following website: https://github.com/songbai-liu/etmo.

cs.NE↗

arXiv2021

Exceptional points in lossy media enable decay-free wave propagation

Waves entering a spatially uniform lossy medium typically undergo exponential decay, arising from either the energy loss of the Beer-Lambert-Bouguer transmission law or the evanescent penetration during reflection. Recently, exceptional point singularities in non-Hermitian systems have been linked to unconventional wave propagation, such as the predicted extremely spatially broad constant-intensity guided modes. Despite such promises, the possibility of decay-free wave propagation in a purely lossy medium has been neither theoretically suggested nor experimentally realized until now. Here we discover and experimentally demonstrate decay-free wave propagation accompanied by a striking uniformly distributed energy loss across arbitrary thicknesses of a homogeneous periodically nanostructured waveguiding medium with exceptional points. Predicted by coupled-mode theory and supported by fully vectorial electromagnetic simulations, hundreds-of-waves deep penetration manifesting spatially constant radiation losses are experimentally observed in photonic slab waveguides. The uniform, decay-free radiative energy loss is measured across the entire structured waveguide region, regardless of its length. While the demonstrated constant-intensity radiation finds an immediate application for generating large, uniform and surface-normal free-space plane waves directly from the photonic chip surface, the uncovered decay-free wave phenomenon is universal and holds true across all domains supporting physical waves, opening new horizons for dispersion-engineered materials empowered by exceptional point physics.

physics.optics↗

arXiv2021

YouRefIt: Embodied Reference Understanding with Language and Gesture

We study the understanding of embodied reference: One agent uses both language and gesture to refer to an object to another agent in a shared physical environment. Of note, this new visual task requires understanding multimodal cues with perspective-taking to identify which object is being referred to. To tackle this problem, we introduce YouRefIt, a new crowd-sourced dataset of embodied reference collected in various physical scenes; the dataset contains 4,195 unique reference clips in 432 indoor scenes. To the best of our knowledge, this is the first embodied reference dataset that allows us to study referring expressions in daily physical scenes to understand referential behavior, human communication, and human-robot interaction. We further devise two benchmarks for image-based and video-based embodied reference understanding. Comprehensive baselines and extensive experiments provide the very first result of machine perception on how the referring expressions and gestures affect the embodied reference understanding. Our results provide essential evidence that gestural cues are as critical as language cues in understanding the embodied reference.

cs.CV↗

arXiv2021

DGEM: A New Dual-modal Graph Embedding Method in Recommendation System

In the current deep learning based recommendation system, the embedding method is generally employed to complete the conversion from the high-dimensional sparse feature vector to the low-dimensional dense feature vector. However, as the dimension of the input vector of the embedding layer is too large, the addition of the embedding layer significantly slows down the convergence speed of the entire neural network, which is not acceptable in real-world scenarios. In addition, as the interaction between users and items increases and the relationship between items becomes more complicated, the embedding method proposed for sequence data is no longer suitable for graphic data in the current real environment. Therefore, in this paper, we propose the Dual-modal Graph Embedding Method (DGEM) to solve these problems. DGEM includes two modes, static and dynamic. We first construct the item graph to extract the graph structure and use random walk of unequal probability to capture the high-order proximity between the items. Then we generate the graph embedding vector through the Skip-Gram model, and finally feed the downstream deep neural network for the recommendation task. The experimental results show that DGEM can mine the high-order proximity between items and enhance the expression ability of the recommendation model. Meanwhile it also improves the recommendation performance by utilizing the time dependent relationship between items.

cs.IR↗

arXiv2021

Learning by Fixing: Solving Math Word Problems with Weak Supervision

Previous neural solvers of math word problems (MWPs) are learned with full supervision and fail to generate diverse solutions. In this paper, we address this issue by introducing a \textit{weakly-supervised} paradigm for learning MWPs. Our method only requires the annotations of the final answers and can generate various solutions for a single problem. To boost weakly-supervised learning, we propose a novel \textit{learning-by-fixing} (LBF) framework, which corrects the misperceptions of the neural network via symbolic reasoning. Specifically, for an incorrect solution tree generated by the neural network, the \textit{fixing} mechanism propagates the error from the root node to the leaf nodes and infers the most probable fix that can be executed to get the desired answer. To generate more diverse solutions, \textit{tree regularization} is applied to guide the efficient shrinkage and exploration of the solution space, and a \textit{memory buffer} is designed to track and save the discovered various fixes for each problem. Experimental results on the Math23K dataset show the proposed LBF framework significantly outperforms reinforcement learning baselines in weakly-supervised learning. Furthermore, it achieves comparable top-1 and much better top-3/5 answer accuracies than fully-supervised methods, demonstrating its strength in producing diverse solutions.

cs.AI↗

arXiv2021

Superconductivity at 44.4 K achieved by intercalating EMIM+ into FeSe

Superconductivity with transition temperature Tc above 40 K was observed in protonated FeSe (Hy-FeSe) previously with the ionic liquid of EMIM-BF4 used in the electrochemical process. However, the real superconducting phase is not clear until now, and detailed structural, magnetization and electrical transport measurements are lacking. By using similar protonating technique on FeSe single crystals, we obtained superconducting samples with Tc above 40 K. We show that the obtained superconducting phase is not Hy-FeSe but actually an organic-ion (C6H11N2+ referred to as EMIM+)-intercalated phase (EMIM)xFeSe. By using X-ray diffraction technique, two sets of index peaks corresponding with different c-axis lattice constants are detected in the protonated samples, which belong to the newly formed phase of intercalated (EMIM)xFeSe and the residual FeSe, respectively. The superconductivity of (EMIM)xFeSe with Tc of 44.4 K is confirmed by resistivity and magnetic susceptibility measurements. Temperature dependence of resistivity with different applied magnetic fields reveals that the upper critical field Hc2 is quite high, while the irreversibility field Hirr is suppressed quickly with increasing temperature till about 20 K. This indicates that the resultant compound has a high anisotropy with a large spacing between the FeSe layers.

cond-mat.supr-con↗

arXiv2021

Multiple-criteria Based Active Learning with Fixed-size Determinantal Point Processes

Active learning aims to achieve greater accuracy with less training data by selecting the most useful data samples from which it learns. Single-criterion based methods (i.e., informativeness and representativeness based methods) are simple and efficient; however, they lack adaptability to different real-world scenarios. In this paper, we introduce a multiple-criteria based active learning algorithm, which incorporates three complementary criteria, i.e., informativeness, representativeness and diversity, to make appropriate selections in the active learning rounds under different data types. We consider the selection process as a Determinantal Point Process, which good balance among these criteria. We refine the query selection strategy by both selecting the hardest unlabeled data sample and biasing towards the classifiers that are more suitable for the current data distribution. In addition, we also consider the dependencies and relationships between these data points in data selection by means of centroidbased clustering approaches. Through evaluations on synthetic and real-world datasets, we show that our method performs significantly better and is more stable than other multiple-criteria based AL algorithms.

cs.LG↗