arXiv subjects

Chao Zhang

Publications and source records attributed to Chao Zhang.

At least 829 records · Page 46Linked to original sources

arXiv2022

CERES: Pretraining of Graph-Conditioned Transformer for Semi-Structured Session Data

User sessions empower many search and recommendation tasks on a daily basis. Such session data are semi-structured, which encode heterogeneous relations between queries and products, and each item is described by the unstructured text. Despite recent advances in self-supervised learning for text or graphs, there lack of self-supervised learning models that can effectively capture both intra-item semantics and inter-item interactions for semi-structured sessions. To fill this gap, we propose CERES, a graph-based transformer model for semi-structured session data. CERES learns representations that capture both inter- and intra-item semantics with (1) a graph-conditioned masked language pretraining task that jointly learns from item text and item-item relations; and (2) a graph-conditioned transformer architecture that propagates inter-item contexts to item-level representations. We pretrained CERES using ~468 million Amazon sessions and find that CERES outperforms strong pretraining baselines by up to 9% in three session search and entity linking tasks.

cs.IR↗

arXiv2022

Input Length Matters: Improving RNN-T and MWER Training for Long-form Telephony Speech Recognition

End-to-end models have achieved state-of-the-art results on several automatic speech recognition tasks. However, they perform poorly when evaluated on long-form data, e.g., minutes long conversational telephony audio. One reason the model fails on long-form speech is that it has only seen short utterances during training. In this paper we study the effect of training utterance length on the word error rate (WER) for RNN-transducer (RNN-T) model. We compare two widely used training objectives, log loss (or RNN-T loss) and minimum word error rate (MWER) loss. We conduct experiments on telephony datasets in four languages. Our experiments show that for both losses, the WER on long-form speech reduces substantially as the training utterance length increases. The average relative WER gain is 15.7% for log loss and 8.8% for MWER loss. When training on short utterances, MWER loss leads to a lower WER than the log loss. Such difference between the two losses diminishes when the input length increases.

eess.AS↗

arXiv2022

Entanglement swapping and quantum correlations via Elegant Joint Measurements

We use hyper-entanglement to experimentally realize deterministic entanglement swapping based on quantum Elegant Joint Measurements. These are joint projections of two qubits onto highly symmetric, iso-entangled, bases. We report measurement fidelities no smaller than $97.4\%$. We showcase the applications of these measurements by using the entanglement swapping procedure to demonstrate quantum correlations in the form of proof-of-principle violations of both bilocal Bell inequalities and more stringent correlation criteria corresponding to full network nonlocality. Our results are a foray into entangled measurements and nonlocality beyond the paradigmatic Bell state measurement and they show the relevance of more general measurements in entanglement swapping scenarios.

quant-ph↗

arXiv2022

Learning a Structured Latent Space for Unsupervised Point Cloud Completion

Unsupervised point cloud completion aims at estimating the corresponding complete point cloud of a partial point cloud in an unpaired manner. It is a crucial but challenging problem since there is no paired partial-complete supervision that can be exploited directly. In this work, we propose a novel framework, which learns a unified and structured latent space that encoding both partial and complete point clouds. Specifically, we map a series of related partial point clouds into multiple complete shape and occlusion code pairs and fuse the codes to obtain their representations in the unified latent space. To enforce the learning of such a structured latent space, the proposed method adopts a series of constraints including structured ranking regularization, latent code swapping constraint, and distribution supervision on the related partial point clouds. By establishing such a unified and structured latent space, better partial-complete geometry consistency and shape completion accuracy can be achieved. Extensive experiments show that our proposed method consistently outperforms state-of-the-art unsupervised methods on both synthetic ShapeNet and real-world KITTI, ScanNet, and Matterport3D datasets.

cs.CV↗

arXiv2022

Abandoning the Bayer-Filter to See in the Dark

Low-light image enhancement - a pervasive but challenging problem, plays a central role in enhancing the visibility of an image captured in a poor illumination environment. Due to the fact that not all photons can pass the Bayer-Filter on the sensor of the color camera, in this work, we first present a De-Bayer-Filter simulator based on deep neural networks to generate a monochrome raw image from the colored raw image. Next, a fully convolutional network is proposed to achieve the low-light image enhancement by fusing colored raw data with synthesized monochrome raw data. Channel-wise attention is also introduced to the fusion process to establish a complementary interaction between features from colored and monochrome raw images. To train the convolutional networks, we propose a dataset with monochrome and color raw pairs named Mono-Colored Raw paired dataset (MCR) collected by using a monochrome camera without Bayer-Filter and a color camera with Bayer-Filter. The proposed pipeline take advantages of the fusion of the virtual monochrome and the color raw images and our extensive experiments indicate that significant improvement can be achieved by leveraging raw sensor data and data-driven learning.

eess.IV↗

arXiv2022

PRBoost: Prompt-Based Rule Discovery and Boosting for Interactive Weakly-Supervised Learning

Weakly-supervised learning (WSL) has shown promising results in addressing label scarcity on many NLP tasks, but manually designing a comprehensive, high-quality labeling rule set is tedious and difficult. We study interactive weakly-supervised learning -- the problem of iteratively and automatically discovering novel labeling rules from data to improve the WSL model. Our proposed model, named PRBoost, achieves this goal via iterative prompt-based rule discovery and model boosting. It uses boosting to identify large-error instances and then discovers candidate rules from them by prompting pre-trained LMs with rule templates. The candidate rules are judged by human experts, and the accepted rules are used to generate complementary weak labels and strengthen the current model. Experiments on four tasks show PRBoost outperforms state-of-the-art WSL baselines up to 7.1% and bridges the gaps with fully supervised models. Our Implementation is available at \url{https://github.com/rz-zhang/PRBoost}.

cs.CL↗

arXiv2022

FORCE: A Framework of Rule-Based Conversational Recommender System

The conversational recommender systems (CRSs) have received extensive attention in recent years. However, most of the existing works focus on various deep learning models, which are largely limited by the requirement of large-scale human-annotated datasets. Such methods are not able to deal with the cold-start scenarios in industrial products. To alleviate the problem, we propose FORCE, a Framework Of Rule-based Conversational Recommender system that helps developers to quickly build CRS bots by simple configuration. We conduct experiments on two datasets in different languages and domains to verify its effectiveness and usability.

cs.IR↗

arXiv2022

Electric Double Layer at the Metal-Oxide/Electrolyte Interface

Metal-oxide surfaces act as both Brønsted acids and bases, which allows the exchange of protons with the electrolyte solution and generates either positive or negative proton charges depending on the environmental pH. These interfacial proton charges are then compensated by counter-ions from the electrolyte solution, which leads to the formation of the electric double layer (EDL). Because the EDL plays a crucial role in electrochemistry, geochemistry and colloid science, understanding the structure-property relationship of the EDL in metal-oxide systems from both experimental and theoretical approaches is necessary. This chapter focuses on the physical chemistry of the protonic double layer at the metal-oxide/electrolyte interface. In particular, determinations of the EDL capacitance and the double-layer potential from potentiometric titration experiments, electrochemical methods, surface-sensitive vibrational spectroscopy and X-ray photoelectron spectroscopy are summarized. This is followed by discussions from the atomistic modelling aspect of the EDL, with an emphasis on the density-functional theory-based molecular dynamics simulations. A conclusion and outlook for future works on this topic are also given.

physics.chem-ph↗

arXiv2022

Mathematical modelling of transport phenomena in compressible multicomponent flows

The present article proposes a diffuse interface model for compressible multicomponent flows with transport phenomena of mass, momentum and energy (i.e., mass diffusion, viscous dissipation and heat conduction). The model is reduced from the seven-equation Baer-Nuziato type model with asymptotic analysis in the limit of instantaneous mechanical relaxations. The main difference between the present model and the Kapila's five-equation model consists in that different time scales for pressure and velocity relaxations are assumed, the former being much smaller than the latter. Thanks to this assumption, the velocity disequilibrium is retained to model the mass diffusion process. Aided by the diffusion laws, the final model still formally consists of five equations. The proposed model satisfy two desirable properties : (1) it respects the laws of thermodynamics, (2) it is free of the spurious oscillation problem in the vicinity of the diffused interface zone. The mass diffusion, viscous dissipation and heat conduction processes contribute parabolic partial differential equations that are solved with the Chebyshev method of local iterations. Numerical results show that the proposed model maintains pressure, velocity and temperature equilibrium near the diffused interface. Convergence tests demonstrate that the numerical methods achieve second order in space and time. The proposed model and numerical methods are applied to simulate the laser-driven RM instability problem in inertial confinement fusion, good agreement with experimental results are observed.

math.AP↗

arXiv2022

Negative Chemical Pressure Effect on Superconductivity and Charge Density Wave of Cu0.5Ir1-xZrxTe2

This study demonstrates the design and synthesis of Cu0.5Ir1-xZrxTe2 system by partial substitution of Ir with Zr acting as a negative chemical pressure. With the doping of Zr, the cell parameters significantly expand, signifying an effective negative chemical pressure. The experimental results found evidence that the charge density wave (CDW)-like order is immediately quenched by subtle Zr substitution for Ir and a classical dome-shape Tc(x) that peaked at 2.80 K can be observed. The optimal Cu0.5Ir0.95Zr0.05Te2 compound is a BCS-type superconductor and exhibits type-II SC. However, high Zr concentration can provoke disorder, inducing the reappearance of CDW order. The present study shows that the Cu0.5Ir1-xZrxTe2 system may provide a new platform for further understanding of multiple electronic orders in transition metal dichalcogenides.

cond-mat.supr-con↗

arXiv2022

Semi-Integrated-Sensing-and-Communication (Semi-ISaC) Networks Assisted by NOMA

This paper investigates non-orthogonal multiple access (NOMA) assisted integrated sensing and communication (ISaC) networks. Compared to the conventional ISaC networks, where the total bandwidth is used for both the radar detection and wireless communications, the proposed Semi-ISaC networks allow that a portion of bandwidth is used for ISaC and the rest of the bandwidth is only utilized for wireless communications. We first derive the analytical expressions of the outage probability for the communication signals, including the signals for the radar target and the communication transmitter. Additionally, we derive the analytical expressions of the ergodic radar estimation information rate (REIR) for the radar echoes. The simulation results show that 1) NOMA ISaC has better spectrum efficiency than the conventional ISaC; and 2) The REIR is enhanced when we enlarge the density of pulses.

cs.IT↗

arXiv2022

Shift-Robust Node Classification via Graph Adversarial Clustering

Graph Neural Networks (GNNs) are de facto node classification models in graph structured data. However, during testing-time, these algorithms assume no data shift, i.e., $\Pr_\text{train}(X,Y) = \Pr_\text{test}(X,Y)$. Domain adaption methods can be adopted for data shift, yet most of them are designed to only encourage similar feature distribution between source and target data. Conditional shift on classes can still affect such adaption. Fortunately, graph yields graph homophily across different data distributions. In response, we propose Shift-Robust Node Classification (SRNC) to address these limitations. We introduce an unsupervised cluster GNN on target graph to group the similar nodes by graph homophily. An adversarial loss with label information on source graph is used upon clustering objective. Then a shift-robust classifier is optimized on training graph and adversarial samples on target graph, which are generated by cluster GNN. We conduct experiments on both open-set shift and representation-shift, which demonstrates the superior accuracy of SRNC on generalizing to test graph with data shift. SRNC is consistently better than previous SoTA domain adaption algorithm on graph that progressively use model predictions on target graph for training.

cs.LG↗

arXiv2022

Optimizing Topological Switching in Confined 2D-Xene Nanoribbons via Finite-Size Effects

In a blueprint for topological electronics, edge state transport in a topological insulator material can be controlled by employing a gate-induced topological quantum phase transition. Here, by studying the width dependence of electronic properties, it is inferred that zigzag-Xene nanoribbons are promising materials for topological electronics with a display of unique physical characteristics associated with the intrinsic band topology and the finite-size effects on gate-induced topological switching. First, due to intertwining with intrinsic band topology-driven energy-zero modes in the pristine case, spin-filtered chiral edge states in zigzag-Xene nanoribbons remain gapless and protected against backward scattering even with finite inter-edge overlapping in ultra-narrow ribbons, i.e., a 2D quantum spin Hall material turns into a 1D topological metal. Second, mainly due to width- and momentum-dependent tunability of the gate-induced inter-edge coupling, the threshold-voltage required for switching between gapless and gapped edge states reduces as the width decreases, without any fundamental lower bound. Third, when the width of zigzag-Xene nanoribbons is smaller than a critical limit, topological switching between edge states can be attained without bulk bandgap closing and reopening. This is primarily due to the quantum confinement effect on the bulk band spectrum which increases the nontrivial bulk bandgap with decrease in width. The existence of such protected gapless edge states and reduction in threshold-voltage accompanied by enhancement in the bulk bandgap overturns the general wisdom of utilizing narrow-gap and wide channel materials for reducing the threshold-voltage in a standard field effect transistor analysis and paves the way toward low-voltage topological devices.

cond-mat.mes-hall↗

arXiv2022

Transference number in polymer electrolytes: mind the reference-frame gap

The transport coefficients, in particular the transference number, of electrolyte solutions are important design parameters for electrochemical energy storage devices. Recent observation of negative transference numbers in PEO-LiTFSI under certain conditions has generated much discussion about its molecular origins, by both experimental and theoretical means. However, one overlooked factor in these efforts is the importance of the reference frame (RF). This creates a non-negligible gap when comparing experiment and simulation, because the fluxes in the experimental measurements of transport coefficients and in the linear response theory used in the molecular dynamics simulation are defined in different RFs. In this work, we show that by applying a proper RF transformation, a much improved agreement between experimental and simulation results can be achieved. Moreover, it is revealed that the anion mass and the anion-anion correlation, rather than ion aggregates, play a crucial role for the reported negative transference numbers.

physics.chem-ph↗

arXiv2022

CAMul: Calibrated and Accurate Multi-view Time-Series Forecasting

Probabilistic time-series forecasting enables reliable decision making across many domains. Most forecasting problems have diverse sources of data containing multiple modalities and structures. Leveraging information as well as uncertainty from these data sources for well-calibrated and accurate forecasts is an important challenging problem. Most previous work on multi-modal learning and forecasting simply aggregate intermediate representations from each data view by simple methods of summation or concatenation and do not explicitly model uncertainty for each data-view. We propose a general probabilistic multi-view forecasting framework CAMul, that can learn representations and uncertainty from diverse data sources. It integrates the knowledge and uncertainty from each data view in a dynamic context-specific manner assigning more importance to useful views to model a well-calibrated forecast distribution. We use CAMul for multiple domains with varied sources and modalities and show that CAMul outperforms other state-of-art probabilistic forecasting models by over 25\% in accuracy and calibration.

cs.LG↗

arXiv2022

Quantum simulation of indefinite causal order induced quantum refrigeration

In the classical world, physical events always happen in a fixed causal order. However, it was recently revealed that quantum mechanics allows events to occur with indefinite causal order (ICO). In this study, we use an optical quantum switch to experimentally investigate the application of ICO in thermodynamic tasks. Specifically, we simulate the working system interacting with two identical thermal reservoirs in an ICO, observing the quantum heat extraction even though they are in thermal equilibrium where heat extraction is unaccessible by traditional thermal contact. Using such a process, we simulate an ICO refrigeration cycle and investigate its properties. We also show that by passing through the ICO channel multiple times, one can extract more heat per cycle and thus obtain a higher refrigeration performance. Our results suggest that the causal nonseparability can be a powerful resource for quantum thermodynamic tasks.

quant-ph↗

arXiv2022

A Survey on Programmatic Weak Supervision

Labeling training data has become one of the major roadblocks to using machine learning. Among various weak supervision paradigms, programmatic weak supervision (PWS) has achieved remarkable success in easing the manual labeling bottleneck by programmatically synthesizing training labels from multiple potentially noisy supervision sources. This paper presents a comprehensive survey of recent advances in PWS. In particular, we give a brief introduction of the PWS learning paradigm, and review representative approaches for each component within PWS's learning workflow. In addition, we discuss complementary learning paradigms for tackling limited labeled data scenarios and how these related approaches can be used in conjunction with PWS. Finally, we identify several critical challenges that remain under-explored in the area to hopefully inspire future research directions in the field.

cs.LG↗

arXiv2022

Supervised Multivariate Learning with Simultaneous Feature Auto-grouping and Dimension Reduction

Modern high-dimensional methods often adopt the "bet on sparsity" principle, while in supervised multivariate learning statisticians may face "dense" problems with a large number of nonzero coefficients. This paper proposes a novel clustered reduced-rank learning (CRL) framework that imposes two joint matrix regularizations to automatically group the features in constructing predictive factors. CRL is more interpretable than low-rank modeling and relaxes the stringent sparsity assumption in variable selection. In this paper, new information-theoretical limits are presented to reveal the intrinsic cost of seeking for clusters, as well as the blessing from dimensionality in multivariate learning. Moreover, an efficient optimization algorithm is developed, which performs subspace learning and clustering with guaranteed convergence. The obtained fixed-point estimators, though not necessarily globally optimal, enjoy the desired statistical accuracy beyond the standard likelihood setup under some regularity conditions. Moreover, a new kind of information criterion, as well as its scale-free form, is proposed for cluster and rank selection, and has a rigorous theoretical support without assuming an infinite sample size. Extensive simulations and real-data experiments demonstrate the statistical accuracy and interpretability of the proposed method.

stat.ML↗