Searcharxiv⌕ Search

arXiv subjects

Tian Lan

Publications and source records attributed to Tian Lan.

At least 235 records · Page 13Linked to original sources
arXiv2020

Which Kind Is Better in Open-domain Multi-turn Dialog,Hierarchical or Non-hierarchical Models? An Empirical Study

Currently, open-domain generative dialog systems have attracted considerable attention in academia and industry. Despite the success of single-turn dialog generation, multi-turn dialog generation is still a big challenge. So far, there are two kinds of models for open-domain multi-turn dialog generation: hierarchical and non-hierarchical models. Recently, some works have shown that the hierarchical models are better than non-hierarchical models under their experimental settings; meanwhile, some works also demonstrate the opposite conclusion. Due to the lack of adequate comparisons, it's not clear which kind of models are better in open-domain multi-turn dialog generation. Thus, in this paper, we will measure systematically nearly all representative hierarchical and non-hierarchical models over the same experimental settings to check which kind is better. Through extensive experiments, we have the following three important conclusions: (1) Nearly all hierarchical models are worse than non-hierarchical models in open-domain multi-turn dialog generation, except for the HRAN model. Through further analysis, the excellent performance of HRAN mainly depends on its word-level attention mechanism; (2) The performance of other hierarchical models will also obtain a great improvement if integrating the word-level attention mechanism into these models. The modified hierarchical models even significantly outperform the non-hierarchical models; (3) The reason why the word-level attention mechanism is so powerful for hierarchical models is because it can leverage context information more effectively, especially the fine-grained information. Besides, we have implemented all of the models and already released the codes.

cs.CL↗
arXiv2020

Modeling and Optimization of Latency in Erasure-coded Storage Systems

As consumers are increasingly engaged in social networking and E-commerce activities, businesses grow to rely on Big Data analytics for intelligence, and traditional IT infrastructures continue to migrate to the cloud and edge, these trends cause distributed data storage demand to rise at an unprecedented speed. Erasure coding has seen itself quickly emerged as a promising technique to reduce storage cost while providing similar reliability as replicated systems, widely adopted by companies like Facebook, Microsoft and Google. However, it also brings new challenges in characterizing and optimizing the access latency when erasure codes are used in distributed storage. The aim of this monograph is to provide a review of recent progress (both theoretical and practical) on systems that employ erasure codes for distributed storage. In this monograph, we will first identify the key challenges and taxonomy of the research problems and then give an overview of different approaches that have been developed to quantify and model latency of erasure-coded storage. This includes recent work leveraging MDS-Reservation, Fork-Join, Probabilistic, and Delayed-Relaunch scheduling policies, as well as their applications to characterize access latency (e.g., mean, tail, asymptotic latency) of erasure-coded distributed storage systems. We will also extend the problem to the case when users are streaming videos from erasure-coded distributed storage systems. Next, we bridge the gap between theory and practice, and discuss lessons learned from prototype implementation. In particular, we will discuss exemplary implementations of erasure-coded storage, illuminate key design degrees of freedom and tradeoffs, and summarize remaining challenges in real-world storage systems such as in content delivery and caching. Open problems for future research are discussed at the end of each chapter.

cs.NI↗
arXiv2020

PONE: A Novel Automatic Evaluation Metric for Open-Domain Generative Dialogue Systems

Open-domain generative dialogue systems have attracted considerable attention over the past few years. Currently, how to automatically evaluate them, is still a big challenge problem. As far as we know, there are three kinds of automatic methods to evaluate the open-domain generative dialogue systems: (1) Word-overlap-based metrics; (2) Embedding-based metrics; (3) Learning-based metrics. Due to the lack of systematic comparison, it is not clear which kind of metrics are more effective. In this paper, we will first measure systematically all kinds of automatic evaluation metrics over the same experimental setting to check which kind is best. Through extensive experiments, the learning-based metrics are demonstrated that they are the most effective evaluation metrics for open-domain generative dialogue systems. Moreover, we observe that nearly all learning-based metrics depend on the negative sampling mechanism, which obtains an extremely imbalanced and low-quality dataset to train a score model. In order to address this issue, we propose a novel and feasible learning-based metric that can significantly improve the correlation with human judgments by using augmented POsitive samples and valuable NEgative samples, called PONE. Extensive experiments demonstrate that our proposed evaluation method significantly outperforms the state-of-the-art learning-based evaluation methods, with an average correlation improvement of 13.18%. In addition, we have publicly released the codes of our proposed method and state-of-the-art baselines.

cs.CL↗
arXiv2020

Real Entropy Can Also Predict Daily Voice Traffic for Wireless Network Users

Voice traffic prediction is significant for network deployment optimization thus to improve the network efficiency. The real entropy based theorectical bound and corresponding prediction models have demonstrated their success in mobility prediction. In this paper, the real entropy based predictability analysis and prediction models are introduced into voice traffic prediction. For this adoption, the traffic quantification methods is proposed and discussed. Based on the real world voice traffic data, the prediction accuracy of N-order Markov models, diffusion based model and MF model are presented, among which, 25-order Markov models performs best and approach close to the maximum predictability. This work demonstrates that, the real entropy can also predict voice traffic well which broaden the understanding on the real entropy based prediction theory.

cs.NI↗
arXiv2019

When to Talk: Chatbot Controls the Timing of Talking during Multi-turn Open-domain Dialogue Generation

Despite the multi-turn open-domain dialogue systems have attracted more and more attention and made great progress, the existing dialogue systems are still very boring. Nearly all the existing dialogue models only provide a response when the user's utterance is accepted. But during daily conversations, humans always decide whether to continue to utter an utterance based on the context. Intuitively, a dialogue model that can control the timing of talking autonomously based on the conversation context can chat with humans more naturally. In this paper, we explore the dialogue system that automatically controls the timing of talking during the conversation. Specifically, we adopt the decision module for the existing dialogue models. Furthermore, modeling conversation context effectively is very important for controlling the timing of talking. So we also adopt the graph neural networks to process the context with the natural graph structure. Extensive experiments on two benchmarks show that controlling the timing of talking can effectively improve the quality of dialogue generation, and the proposed methods significantly improve the accuracy of the timing of talking. In addition, we have publicly released the codes of our proposed model.

cs.CL↗
arXiv2019

Fully superconducting machine for electric aircraft propulsion: study of AC loss for HTS stator

Fully superconducting machines provide the high power density required for future electric aircraft propulsion. However, superconducting windings generate AC losses in AC electrical machine environments. These AC losses are difficult to remove at low temperatures and they add an extra burden to the aircraft cooling system. Due to heavy cooling penalty, AC losses in the HTS stator, is one of the key topics in HTS machine design. In order to evaluate the AC loss of superconducting stator windings in a rotational machine environment, we designed and built a novel axial-flux high temperature superconducting (HTS) machine platform. The AC loss measurement is based on calorimetrically boiling-off liquid nitrogen. Both total AC loss and magnetisation loss in HTS stator are measured in a rotational magnetic field condition. This platform is essential to study ways to minimise AC losses in HTS stator, in order to maximum the efficiency of fully HTS machines.

physics.app-ph↗
arXiv2019

Gapped domain walls between 2+1D topologically ordered states

The 2+1D topological order can be characterized by the mapping-class-group representations for Riemann surfaces of genus-1, genus-2, etc. In this paper, we use those representations to determine the possible gapped boundaries of a 2+1D topological order, as well as the domain walls between two topological orders. We find that mapping-class-group representations for both genus-1 and genus-2 surfaces are needed to determine the gapped domain walls and boundaries. Our systematic theory is based on the fixed-point partition functions for the walls (or the boundaries), which completely characterize the gapped domain walls (or the boundaries). The mapping-class-group representations give rise to conditions that must be satisfied by the fixed-point partition functions, which leads to a systematic theory. Such conditions can be viewed as bulk topological order determining the (non-invertible) gravitational anomaly at the domain wall, and our theory can be viewed as finding all types of the gapped domain wall given a (non-invertible) gravitational anomaly. We also developed a systematic theory of gapped domain walls (boundaries) based on the structure coefficients of condensable algebras.

cond-mat.str-el↗
arXiv2019

Generative Dialog Policy for Task-oriented Dialog Systems

There is an increasing demand for task-oriented dialogue systems which can assist users in various activities such as booking tickets and restaurant reservations. In order to complete dialogues effectively, dialogue policy plays a key role in task-oriented dialogue systems. As far as we know, the existing task-oriented dialogue systems obtain the dialogue policy through classification, which can assign either a dialogue act and its corresponding parameters or multiple dialogue acts without their corresponding parameters for a dialogue action. In fact, a good dialogue policy should construct multiple dialogue acts and their corresponding parameters at the same time. However, it's hard for existing classification-based methods to achieve this goal. Thus, to address the issue above, we propose a novel generative dialogue policy learning method. Specifically, the proposed method uses attention mechanism to find relevant segments of given dialogue context and input utterance and then constructs the dialogue policy by a seq2seq way for task-oriented dialogue systems. Extensive experiments on two benchmark datasets show that the proposed model significantly outperforms the state-of-the-art baselines. In addition, we have publicly released our codes.

cs.CL↗
arXiv2019

Matrix formulation for non-Abelian families

We generalize the $K$ matrix formulation to non-trivial non-Abelian families of 2+1D topological orders. Given a topological order $\mathcal C$, any topological order in the same non-Abelian family as $\mathcal C$ can be efficiently described by $\boldsymbol{a}=(a_I)$ where $a_I$ are Abelian anyons in $\mathcal C$, together with a symmetric invertible matrix $K$, $K_{IJ}=k_{IJ}-t_{a_I,a_J}$ where $k_{IJ}$ are integers, $k_{II}$ are even and $t_{a_I,a_J}$ are the mutual statistics between $a_I,a_J$. In particular, when $\mathcal C$ is a root whose rank is the smallest in the family, $K$ becomes an integer matrix. Our results make it possible to generate the data of large numbers of topological orders instantly.

cond-mat.str-el↗
arXiv2019

CHOP: Bypassing Runtime Bounds Checking Through Convex Hull OPtimization

Unsafe memory accesses in programs written using popular programming languages like C/C++ have been among the leading causes for software vulnerability. Prior memory safety checkers such as SoftBound enforce memory spatial safety by checking if every access to array elements are within the corresponding array bounds. However, it often results in high execution time overhead due to the cost of executing the instructions associated with bounds checking. To mitigate this problem, redundant bounds check elimination techniques are needed. In this paper, we propose CHOP, a Convex Hull OPtimization based framework, for bypassing redundant memory bounds checking via profile-guided inferences. In contrast to existing check elimination techniques that are limited by static code analysis, our solution leverages a model-based inference to identify redundant bounds checking based on runtime data from past program executions. For a given function, it rapidly derives and updates a knowledge base containing sufficient conditions for identifying redundant array bounds checking. We evaluate CHOP on real-world applications and benchmark (such as SPEC) and the experimental results show that on average 80.12% of dynamic bounds check instructions can be avoided, resulting in improved performance up to 95.80% over SoftBound.

cs.PL↗
arXiv2019

Large-scale 3D point cloud representations via graph inception networks with applications to autonomous driving

We present a novel graph-neural-network-based system to effectively represent large-scale 3D point clouds with the applications to autonomous driving. Many previous works studied the representations of 3D point clouds based on two approaches, voxelization, which causes discretization errors and learning, which is hard to capture huge variations in large-scale scenarios. In this work, we combine voxelization and learning: we discretize the 3D space into voxels and propose novel graph inception networks to represent 3D points in each voxel. This combination makes the system avoid discretization errors and work for large-scale scenarios. The entire system for large-scale 3D point clouds acts like the blocked discrete cosine transform for 2D images; we thus call it the point cloud neural transform (PCT). We further apply the proposed PCT to represent real-time LiDAR sweeps produced by self-driving cars and the PCT with graph inception networks significantly outperforms its competitors.

eess.SP↗
arXiv2019

A Study Of a Wide-Angle Scanning Phased Array Based On a High-Impedance Surface Ground Plane

This paper presents a two-dimensional infinite dipole array system with a mushroom-like high-impedance surface (HIS) ground plane with wide-angle scanning capability in the E-plane. The unit cell of the proposed antenna array consists of a dipole antenna and a four-by-four HIS ground. The simulation results show that the proposed antenna array can achieve a wide scanning angle of up to 65$^{\circ}$ in the E-plane with an excellent impedance match and a small $S11$. Floquet mode analysis is utilized to analyze the active impedance and the reflection coefficient. Good agreement is obtained between the theoretical results and the simulations. Using numerical and theoretical analyses, we reveal the mechanism of such excellent wide scanning properties. For the range of small scanning angles, these excellent properties result mainly from the special reflection phase of the HIS ground, which can cause the mutual coupling between the elements of the real array to be compensated by the mutual coupling effect between the real array and the mirror array. For the range of large scanning angles, since the surface wave (SW) mode could be resonantly excited by a high-order Floquet mode $\textrm{TM}_{-1,0}$ from the array and since the SW mode could be converted into a leaky wave (LW) mode by the scattering of the array, the radiation field from the LW mode is nearly in phase with the direct radiating field from the array. Therefore, with help from the special reflection phase of the HIS and the designed LW mode of the HIS ground, the antenna array with an HIS ground can achieve wide-angle scanning performance.

physics.app-ph↗
arXiv2018

Fermion decoration construction of symmetry protected trivial orders for fermion systems with any symmetries $G_f$ and in any dimensions

We use higher dimensional bosonization and fermion decoration to construct exactly soluble interacting fermion models to realize fermionic symmetry protected trivial (SPT) orders (which are also known as symmetry protected topological orders) in any dimensions and for generic fermion symmetries $G_f$, which can be a non-trivial $Z_2^f$ extension (where $Z_2^f$ is the fermion-number-parity symmetry). This generalizes the previous results from group superconhomology of Gu and Wen (arXiv:1201.2648), where $G_f$ is assumed to be a trivial $Z_2^f$ extension. We find that the SPT phases from fermion decoration construction can be described in a compact way using higher groups.

cond-mat.str-el↗
arXiv2018

Topological non-linear $σ$-model, higher gauge theory, and a realization of all 3+1D topological orders for boson systems

A discrete non-linear $σ$-model is obtained by triangulate both the space-time $M^{d+1}$ and the target space $K$. If the path integral is given by the sum of all the complex homomorphisms $ϕ: M^{d+1} \to K$, with an partition function that is independent of space-time triangulation, then the corresponding non-linear $σ$-model will be called topological non-linear $σ$-model which is exactly soluble. Those exactly soluble models suggest that phase transitions induced by fluctuations with no topological defects (i.e. fluctuations described by homomorphisms $ϕ$) usually produce a topologically ordered state and are topological phase transitions, while phase transitions induced by fluctuations with all the topological defects give rise to trivial product states and are not topological phase transitions. If $K$ is a space with only non-trivial first homotopy group $G$ which is finite, those topological non-linear $σ$-models can realize all 3+1D bosonic topological orders without emergent fermions, which are described by Dijkgraaf-Witten theory with gauge group $π_1(K)=G$. Here, we show that the 3+1D bosonic topological orders with emergent fermions can be realized by topological non-linear $σ$-models with $π_1(K)=$ finite groups, $π_2(K)=Z_2$, and $π_{n>2}(K)=0$. A subset of those topological non-linear $σ$-models corresponds to 2-gauge theories, which realize and classify bosonic topological orders with emergent fermions that have no emergent Majorana zero modes at triple string intersections. The classification of 3+1D bosonic topological orders may correspond to a classification of unitary fully dualizable fully extended topological quantum field theories in 4-dimensions.

cond-mat.str-el↗
arXiv2018

A classification of 3+1D bosonic topological orders (II): the case when some point-like excitations are fermions

In this paper, we classify EF topological orders for 3+1D bosonic systems where some emergent pointlike excitations are fermions. (1) We argue that all 3+1D bosonic topological orders have gappable boundary. (2) All the pointlike excitations in EF topological orders are described by the representations of $G_f=Z_2^f\leftthreetimes_{e_2} G_b$ -- a $Z_2^f$ central extension of a finite group $G_b$ characterized by $e_2\in H^2(G_b,Z_2)$. (3) We find that the EF topological orders are classified by 2+1D anomalous topological orders $\mathcal{A}_b^3$ on their unique canonical boundary. Here $\mathcal{A}_b^3$ is a unitary fusion 2-category with simple objects labeled by $\hat G_b=Z_2^m\leftthreetimes G_b$. $\mathcal{A}_b^3$ also has one invertible fermionic 1-morphism for each object as well as quantum-dimension-$\sqrt 2$ 1-morphisms that connect two objects $g$ and $gm$, where $g\in \hat G_b$ and $m$ is the generator of $Z_2^m$. (4) When $\hat G_b$ is the trivial $Z_2^m$ extension, the EF topological orders are called EF1 topological orders, which is classified by simple data $(G_b,e_2,n_3,ν_4)$. (5) When $\hat G_b$ is a non-trivial $Z_2^m$ extension, the EF topological orders are called EF2 topological orders, where some intersections of three stringlike excitations must carry Majorana zero modes. (6) Every EF2 topological order with $G_f=Z_2^f\leftthreetimes G_b$ can be associated with a EF1 topological order with $G_f=Z_2^f\leftthreetimes \hat G_b$. (7) We find that all EF topological orders correspond to gauged 3+1D fermionic symmetry protected topological (SPT) orders with a finite unitary symmetry group. (8) We further propose that the general classification of 3+1D topological orders with finite unitary symmetries for bosonic and fermionic systems can be obtained by gauging or partially gauging the finite symmetry group of 3+1D SPT phases of bosonic and fermionic systems.

cond-mat.str-el↗
arXiv2018

FastTrack: Minimizing Stalls for CDN-based Over-the-top Video Streaming Systems

Traffic for internet video streaming has been rapidly increasing and is further expected to increase with the higher definition videos and IoT applications, such as 360 degree videos and augmented virtual reality applications. While efficient management of heterogeneous cloud resources to optimize the quality of experience is important, existing work in this problem space often left out important factors. In this paper, we present a model for describing a today's representative system architecture for video streaming applications, typically composed of a centralized origin server and several CDN sites. Our model comprehensively considers the following factors: limited caching spaces at the CDN sites, allocation of CDN for a video request, choice of different ports from the CDN, and the central storage and bandwidth allocation. With the model, we focus on minimizing a performance metric, stall duration tail probability (SDTP), and present a novel, yet efficient, algorithm to solve the formulated optimization problem. The theoretical bounds with respect to the SDTP metric are also analyzed and presented. Our extensive simulation results demonstrate that the proposed algorithms can significantly improve the SDTP metric, compared to the baseline strategies. Small-scale video streaming system implementation in a real cloud environment further validates our results.

cs.NI↗
arXiv2018

RUN:Residual U-Net for Computer-Aided Detection of Pulmonary Nodules without Candidate Selection

The early detection and early diagnosis of lung cancer are crucial to improve the survival rate of lung cancer patients. Pulmonary nodules detection results have a significant impact on the later diagnosis. In this work, we propose a new network named RUN to complete nodule detection in a single step by bypassing the candidate selection. The system introduces the shortcut of the residual network to improve the traditional U-Net, thereby solving the disadvantage of poor results due to its lack of depth. Furthermore, we compare the experimental results with the traditional U-Net. We validate our method in LUng Nodule Analysis 2016 (LUNA16) Nodule Detection Challenge. We acquire a sensitivity of 90.90% at 2 false positives per scan and therefore achieve better performance than the current state-of-the-art approaches.

cs.CV↗