arXiv subjects

Haitao Liu

Publications and source records attributed to Haitao Liu.

At least 37 records · Page 2Linked to original sources

arXiv2021

Scalable Multi-Task Gaussian Processes with Neural Embedding of Coregionalization

Multi-task regression attempts to exploit the task similarity in order to achieve knowledge transfer across related tasks for performance improvement. The application of Gaussian process (GP) in this scenario yields the non-parametric yet informative Bayesian multi-task regression paradigm. Multi-task GP (MTGP) provides not only the prediction mean but also the associated prediction variance to quantify uncertainty, thus gaining popularity in various scenarios. The linear model of coregionalization (LMC) is a well-known MTGP paradigm which exploits the dependency of tasks through linear combination of several independent and diverse GPs. The LMC however suffers from high model complexity and limited model capability when handling complicated multi-task cases. To this end, we develop the neural embedding of coregionalization that transforms the latent GPs into a high-dimensional latent space to induce rich yet diverse behaviors. Furthermore, we use advanced variational inference as well as sparse approximation to devise a tight and compact evidence lower bound (ELBO) for higher quality of scalable model inference. Extensive numerical experiments have been conducted to verify the higher prediction quality and better generalization of our model, named NSVLMC, on various real-world multi-task datasets and the cross-fluid modeling of unsteady fluidized bed.

stat.ML↗

arXiv2021

Deep Reinforcement Learning-based Task Offloading in Satellite-Terrestrial Edge Computing Networks

In remote regions (e.g., mountain and desert), cellular networks are usually sparsely deployed or unavailable. With the appearance of new applications (e.g., industrial automation and environment monitoring) in remote regions, resource-constrained terminals become unable to meet the latency requirements. Meanwhile, offloading tasks to urban terrestrial cloud (TC) via satellite link will lead to high delay. To tackle above issues, Satellite Edge Computing architecture is proposed, i.e., users can offload computing tasks to visible satellites for executing. However, existing works are usually limited to offload tasks in pure satellite networks, and make offloading decisions based on the predefined models of users. Besides, the runtime consumption of existing algorithms is rather high. In this paper, we study the task offloading problem in satellite-terrestrial edge computing networks, where tasks can be executed by satellite or urban TC. The proposed Deep Reinforcement learning-based Task Offloading (DRTO) algorithm can accelerate learning process by adjusting the number of candidate locations. In addition, offloading location and bandwidth allocation only depend on the current channel states. Simulation results show that DRTO achieves near-optimal offloading cost performance with much less runtime consumption, which is more suitable for satellite-terrestrial network with fast fading channel.

cs.NI↗

arXiv2020

Statistical patterns of word frequency suggesting the probabilistic nature of human languages

Traditional linguistic theories have largely regard language as a formal system composed of rigid rules. However, their failures in processing real language, the recent successes in statistical natural language processing, and the findings of many psychological experiments have suggested that language may be more a probabilistic system than a formal system, and thus cannot be faithfully modeled with the either/or rules of formal linguistic theory. The present study, based on authentic language data, confirmed that those important linguistic issues, such as linguistic universal, diachronic drift, and language variations can be translated into probability and frequency patterns in parole. These findings suggest that human language may well be probabilistic systems by nature, and that statistical may well make inherent properties of human languages.

cs.CL↗

arXiv2020

Modulating Scalable Gaussian Processes for Expressive Statistical Learning

For a learning task, Gaussian process (GP) is interested in learning the statistical relationship between inputs and outputs, since it offers not only the prediction mean but also the associated variability. The vanilla GP however struggles to learn complicated distribution with the property of, e.g., heteroscedastic noise, multi-modality and non-stationarity, from massive data due to the Gaussian marginal and the cubic complexity. To this end, this article studies new scalable GP paradigms including the non-stationary heteroscedastic GP, the mixture of GPs and the latent GP, which introduce additional latent variables to modulate the outputs or inputs in order to learn richer, non-Gaussian statistical representation. We further resort to different variational inference strategies to arrive at analytical or tighter evidence lower bounds (ELBOs) of the marginal likelihood for efficient and effective model training. Extensive numerical experiments against state-of-the-art GP and neural network (NN) counterparts on various tasks verify the superiority of these scalable modulated GPs, especially the scalable latent GP, for learning diverse data distributions.

stat.ML↗

arXiv2020

Deep Latent-Variable Kernel Learning

Deep kernel learning (DKL) leverages the connection between Gaussian process (GP) and neural networks (NN) to build an end-to-end, hybrid model. It combines the capability of NN to learn rich representations under massive data and the non-parametric property of GP to achieve automatic regularization that incorporates a trade-off between model fit and model complexity. However, the deterministic encoder may weaken the model regularization of the following GP part, especially on small datasets, due to the free latent representation. We therefore present a complete deep latent-variable kernel learning (DLVKL) model wherein the latent variables perform stochastic encoding for regularized representation. We further enhance the DLVKL from two aspects: (i) the expressive variational posterior through neural stochastic differential equation (NSDE) to improve the approximation quality, and (ii) the hybrid prior taking knowledge from both the SDE prior and the posterior to arrive at a flexible trade-off. Intensive experiments imply that the DLVKL-NSDE performs similarly to the well calibrated GP on small datasets, and outperforms existing deep GPs on large datasets.

stat.ML↗

arXiv2020

Design and Implementation of a Low-Latency and High-Reliability System Based on Software-Defined Radio (SDR)

Ultra-reliable and low-latency communication (URLLC) is one of the three major service classes supported by the fifth generation (5G) New Radio (NR) technical specifications. In this paper, we introduce a physical layer architecture that can meet the low-latency and high-reliability requirements. The downlink system is designed according to the Third Generation Partner Project (3GPP) specifications based on software defined radio (SDR) system. The URLLC system physical layer downlink is implemented on the open source OpenAirInterface (OAI) platform to evaluate the latency and reliability performance of the scheme. Not only the URLLC system reliability performance is tested based on the simulation platform, but also the delay performance is evaluated by the realization of the over-the-air system. The experimental results show that the designed scheme can approximately meet the reliability and delay performance requirements of the 3GPP specifications.

eess.SP↗

arXiv2020

Large-scale Heteroscedastic Regression via Gaussian Process

Heteroscedastic regression considering the varying noises among observations has many applications in the fields like machine learning and statistics. Here we focus on the heteroscedastic Gaussian process (HGP) regression which integrates the latent function and the noise function together in a unified non-parametric Bayesian framework. Though showing remarkable performance, HGP suffers from the cubic time complexity, which strictly limits its application to big data. To improve the scalability, we first develop a variational sparse inference algorithm, named VSHGP, to handle large-scale datasets. Furthermore, two variants are developed to improve the scalability and capability of VSHGP. The first is stochastic VSHGP (SVSHGP) which derives a factorized evidence lower bound, thus enhancing efficient stochastic variational inference. The second is distributed VSHGP (DVSHGP) which (i) follows the Bayesian committee machine formalism to distribute computations over multiple local VSHGP experts with many inducing points; and (ii) adopts hybrid parameters for experts to guard against over-fitting and capture local variety. The superiority of DVSHGP and SVSHGP as compared to existing scalable heteroscedastic/homoscedastic GPs is then extensively verified on various datasets.

stat.ML↗

arXiv2019

Scalable Gaussian Process Classification with Additive Noise for Various Likelihoods

Gaussian process classification (GPC) provides a flexible and powerful statistical framework describing joint distributions over function space. Conventional GPCs however suffer from (i) poor scalability for big data due to the full kernel matrix, and (ii) intractable inference due to the non-Gaussian likelihoods. Hence, various scalable GPCs have been proposed through (i) the sparse approximation built upon a small inducing set to reduce the time complexity; and (ii) the approximate inference to derive analytical evidence lower bound (ELBO). However, these scalable GPCs equipped with analytical ELBO are limited to specific likelihoods or additional assumptions. In this work, we present a unifying framework which accommodates scalable GPCs using various likelihoods. Analogous to GP regression (GPR), we introduce additive noises to augment the probability space for (i) the GPCs with step, (multinomial) probit and logit likelihoods via the internal variables; and particularly, (ii) the GPC using softmax likelihood via the noise variables themselves. This leads to unified scalable GPCs with analytical ELBO by using variational inference. Empirically, our GPCs showcase better results than state-of-the-art scalable GPCs for extensive binary/multi-class classification tasks with up to two million data points.

stat.ML↗

arXiv2019

Understanding the origin of bandgap problem in transition and post-transition metal oxides

Improving electronic structure calculations for practical and technologically-important materials has been a never-ending pursue. This is especially true for transition and post-transition metal oxides for which the current first-principles approaches still suffer various drawbacks. Here we present a hierarchical-hybrid functional approach built on the use of pseudopotentials. The key is to introduce a discontinuity in the exchange functional between core and valence electrons. It allows for treating the localization errors of sp and d electrons differently, which have been known to be an important source of error for the band gap. Using ZnO as a prototype, we show the approach is successful in simultaneously reproducing the band gap and d-band position. Remarkably, the same approach, without having to change the hybrid mixing parameters from those of Zn, works reasonably well for other binary 3d transition and post-transition metal oxides across board. Our findings point to a new direction of systematically improving the exchange functional in first-principles calculations.

cond-mat.mtrl-sci↗

arXiv2019

Two-dimensional ferromagnetic-ferroelectric multiferroics in violation of the d0 rule

Contribution of d-electron to ferroelectricity of type-II multiferroics causes strong magneto-electric coupling and distinguishes them from the conventional type-I multiferroics. However, their therein polarization is too small because the ferroelectricity is merely a derivative from the magnetic order. Here we report a new class of multiferroic materials, monolayer VOX2 (X = Cl, Br, and I), which combine the advantages of type-I and type-II multiferroics. Both ferroelectricity and magnetism arise from the same V cation, where the filled d-orbital is perpendicular to an a priori ferroelectric polarization and thus poses no hindrance to ferroelectricity, indicating a violation of the usual d0 rule. This makes the combination of large polarizations and strong magneto-electric coupling possible. Our findings not only add new ferromagnetic-ferroelectric multiferroics, but also point to a unique mechanism to engineer multiferroics.

cond-mat.mtrl-sci↗

arXiv2019

When Gaussian Process Meets Big Data: A Review of Scalable GPs

The vast quantity of information brought by big data as well as the evolving computer hardware encourages success stories in the machine learning community. In the meanwhile, it poses challenges for the Gaussian process (GP) regression, a well-known non-parametric and interpretable Bayesian model, which suffers from cubic complexity to data size. To improve the scalability while retaining desirable prediction quality, a variety of scalable GPs have been presented. But they have not yet been comprehensively reviewed and analyzed in order to be well understood by both academia and industry. The review of scalable GPs in the GP community is timely and important due to the explosion of data size. To this end, this paper is devoted to the review on state-of-the-art scalable GPs involving two main categories: global approximations which distillate the entire data and local approximations which divide the data for subspace learning. Particularly, for global approximations, we mainly focus on sparse approximations comprising prior approximations which modify the prior but perform exact inference, posterior approximations which retain exact prior but perform approximate inference, and structured sparse approximations which exploit specific structures in kernel matrix; for local approximations, we highlight the mixture/product of experts that conducts model averaging from multiple local experts to boost predictions. To present a complete review, recent advances for improving the scalability and capability of scalable GPs are reviewed. Finally, the extensions and open issues regarding the implementation of scalable GPs in various scenarios are reviewed and discussed to inspire novel ideas for future research avenues.

stat.ML↗

arXiv2018

Anomaly Detection via Graphical Lasso

Anomalies and outliers are common in real-world data, and they can arise from many sources, such as sensor faults. Accordingly, anomaly detection is important both for analyzing the anomalies themselves and for cleaning the data for further analysis of its ambient structure. Nonetheless, a precise definition of anomalies is important for automated detection and herein we approach such problems from the perspective of detecting sparse latent effects embedded in large collections of noisy data. Standard Graphical Lasso-based techniques can identify the conditional dependency structure of a collection of random variables based on their sample covariance matrix. However, classic Graphical Lasso is sensitive to outliers in the sample covariance matrix. In particular, several outliers in a sample covariance matrix can destroy the sparsity of its inverse. Accordingly, we propose a novel optimization problem that is similar in spirit to Robust Principal Component Analysis (RPCA) and splits the sample covariance matrix $M$ into two parts, $M=F+S$, where $F$ is the cleaned sample covariance whose inverse is sparse and computable by Graphical Lasso, and $S$ contains the outliers in $M$. We accomplish this decomposition by adding an additional $ \ell_1$ penalty to classic Graphical Lasso, and name it "Robust Graphical Lasso (Rglasso)". Moreover, we propose an Alternating Direction Method of Multipliers (ADMM) solution to the optimization problem which scales to large numbers of unknowns. We evaluate our algorithm on both real and synthetic datasets, obtaining interpretable results and outperforming the standard robust Minimum Covariance Determinant (MCD) method and Robust Principal Component Analysis (RPCA) regarding both accuracy and speed.

stat.ML↗

arXiv2018

Understanding and Comparing Scalable Gaussian Process Regression for Big Data

As a non-parametric Bayesian model which produces informative predictive distribution, Gaussian process (GP) has been widely used in various fields, like regression, classification and optimization. The cubic complexity of standard GP however leads to poor scalability, which poses challenges in the era of big data. Hence, various scalable GPs have been developed in the literature in order to improve the scalability while retaining desirable prediction accuracy. This paper devotes to investigating the methodological characteristics and performance of representative global and local scalable GPs including sparse approximations and local aggregations from four main perspectives: scalability, capability, controllability and robustness. The numerical experiments on two toy examples and five real-world datasets with up to 250K points offer the following findings. In terms of scalability, most of the scalable GPs own a time complexity that is linear to the training size. In terms of capability, the sparse approximations capture the long-term spatial correlations, the local aggregations capture the local patterns but suffer from over-fitting in some scenarios. In terms of controllability, we could improve the performance of sparse approximations by simply increasing the inducing size. But this is not the case for local aggregations. In terms of robustness, local aggregations are robust to various initializations of hyperparameters due to the local attention mechanism. Finally, we highlight that the proper hybrid of global and local scalable GPs may be a promising way to improve both the model capability and scalability for big data.

stat.ML↗

arXiv2018

Zipf's law in 50 languages: its structural pattern, linguistic interpretation, and cognitive motivation

Zipf's law has been found in many human-related fields, including language, where the frequency of a word is persistently found as a power law function of its frequency rank, known as Zipf's law. However, there is much dispute whether it is a universal law or a statistical artifact, and little is known about what mechanisms may have shaped it. To answer these questions, this study conducted a large scale cross language investigation into Zipf's law. The statistical results show that Zipf's laws in 50 languages all share a 3-segment structural pattern, with each segment demonstrating distinctive linguistic properties and the lower segment invariably bending downwards to deviate from theoretical expectation. This finding indicates that this deviation is a fundamental and universal feature of word frequency distributions in natural languages, not the statistical error of low frequency words. A computer simulation based on the dual-process theory yields Zipf's law with the same structural pattern, suggesting that Zipf's law of natural languages are motivated by common cognitive mechanisms. These results show that Zipf's law in languages is motivated by cognitive mechanisms like dual-processing that govern human verbal behaviors.

cs.CL↗

arXiv2018

Generalized Robust Bayesian Committee Machine for Large-scale Gaussian Process Regression

In order to scale standard Gaussian process (GP) regression to large-scale datasets, aggregation models employ factorized training process and then combine predictions from distributed experts. The state-of-the-art aggregation models, however, either provide inconsistent predictions or require time-consuming aggregation process. We first prove the inconsistency of typical aggregations using disjoint or random data partition, and then present a consistent yet efficient aggregation model for large-scale GP. The proposed model inherits the advantages of aggregations, e.g., closed-form inference and aggregation, parallelization and distributed computing. Furthermore, theoretical and empirical analyses reveal that the new aggregation model performs better due to the consistent predictions that converge to the true underlying function when the training size approaches infinity.

stat.ML↗

arXiv2018

Electronic nature of coverage-dependent nanosurface effect by cooperative orbital redistribution

Nanomaterial surface states can effectively modify or even dominate their physical and chemical properties due to large surface-to-volume ratios. Such surface effects are highly dependent on particle size and ligand coverage, yet the underlying electronic-level mechanism still remains unknown. Using TiO2 nanosheet as a model system, we reveal the electronic nature of coverage-dependent nanosurface effects through varying ligand coverage and probing the modified surface bonding and electronic band structures with near-edge X-ray absorption fine structure. We discover experimentally that surface ligands can competitively polarize the 3d orbitals of surface Ti atoms into chemisorption states, which is cooperative with increased ligand coverages. Such coverage-dependent cooperative orbital redistribution accounts for various nanosurface effects on regulating the electronic structure, surface reactivity, optical property, and chemisorption of nanomaterials.

cond-mat.mtrl-sci↗

arXiv2017

Design and Implementation of a LTE-WiFi Aggregation System based on SDR

With explosive growth of the mobile Internet access and the popularization of wireless local area network (WLAN) access points (APs), wireless fidelity (WiFi) offloading is considered as an important supplementary technique to reduce the load of cellular infrastructure and enhance quality of network service. In this paper, we design and implement the framework of LTEWiFi aggregation (LWA) which is in line with the 3rd Generation Partnership Project (3GPP) released standard recently. In the LWA system, the process of data offloading is different from that of WiFi interworking on the cellular core network. WLAN APs directly connect to eNodeBs (eNBs), and the offloading is realized in Packet Data Convergence Protocol layer (PDCP). Thus, this architecture makes full use of existed WLAN APs to improve the performance of indoor cellular network. Besides, benefiting from the flow control between LTE and WiFi, eNB schedule can be more flexible and efficient. Next, it is implemented based on open source OpenAirInterface (OAI) software defined radio (SDR) platform, where a simple and practical reordering method is proposed and carried out. Experimental results demonstrate that our design works stably with the WiFi-offloading and reordering functions.

eess.SP↗

arXiv2016

Existence of Hierarchies and Human's Pursuit of Top Hierarchy Lead to Power Law

The power law is ubiquitous in natural and social phenomena, and is considered as a universal relationship between the frequency and its rank for diverse social systems. However, a general model is still lacking to interpret why these seemingly unrelated systems share great similarity. Through a detailed analysis of natural language texts and simulation experiments based on the proposed 'Hierarchical Selection Model', we found that the existence of hierarchies and human's pursuit of top hierarchy lead to the power law. Further, the power law is a statistical and emergent performance of hierarchies, and it is the universality of hierarchies that contributes to the ubiquity of the power law.

cs.CL↗