Searcharxiv⌕ Search

arXiv subjects

Hong Liu

Publications and source records attributed to Hong Liu.

At least 289 records · Page 16Linked to original sources
arXiv2022

A Generative User Simulator with GPT-based Architecture and Goal State Tracking for Reinforced Multi-Domain Dialog Systems

Building user simulators (USs) for reinforcement learning (RL) of task-oriented dialog systems (DSs) has gained more and more attention, which, however, still faces several fundamental challenges. First, it is unclear whether we can leverage pretrained language models to design, for example, GPT-2 based USs, to catch up and interact with the recently advanced GPT-2 based DSs. Second, an important ingredient in a US is that the user goal can be effectively incorporated and tracked; but how to flexibly integrate goal state tracking and develop an end-to-end trainable US for multi-domains has remained to be a challenge. In this work, we propose a generative user simulator (GUS) with GPT-2 based architecture and goal state tracking towards addressing the above two challenges. Extensive experiments are conducted on MultiWOZ2.1. Different DSs are trained via RL with GUS, the classic agenda-based user simulator (ABUS) and other ablation simulators respectively, and are compared for cross-model evaluation, corpus-based evaluation and human evaluation. The GUS achieves superior results in all three evaluation tasks.

cs.CL↗
arXiv2022

A Hybrid System of Sound Event Detection Transformer and Frame-wise Model for DCASE 2022 Task 4

In this paper, we describe in detail our system for DCASE 2022 Task4. The system combines two considerably different models: an end-to-end Sound Event Detection Transformer (SEDT) and a frame-wise model, Metric Learning and Focal Loss CNN (MLFL-CNN). The former is an event-wise model which learns event-level representations and predicts sound event categories and boundaries directly, while the latter is based on the widely adopted frame-classification scheme, under which each frame is classified into event categories and event boundaries are obtained by post-processing such as thresholding and smoothing. For SEDT, self-supervised pre-training using unlabeled data is applied, and semi-supervised learning is adopted by using an online teacher, which is updated from the student model using the Exponential Moving Average (EMA) strategy and generates reliable pseudo labels for weakly-labeled and unlabeled data. For the frame-wise model, the ICT-TOSHIBA system of DCASE 2021 Task 4 is used. Experimental results show that the hybrid system considerably outperforms either individual model and achieves psds1 of 0.420 and psds2 of 0.783 on the validation set without external data. The code is available at https://github.com/965694547/Hybrid-system-of-frame-wise-model-and-SEDT.

cs.SD↗
arXiv2022

Building Markovian Generative Architectures over Pretrained LM Backbones for Efficient Task-Oriented Dialog Systems

Recently, Transformer based pretrained language models (PLMs), such as GPT2 and T5, have been leveraged to build generative task-oriented dialog (TOD) systems. A drawback of existing PLM-based models is their non-Markov architectures across turns, i.e., the whole history is used as the conditioning input at each turn. First, this brings inefficiencies in memory and computation. Furthermore, using the whole history increases model complexity and may hurt the training efficiency, especially when facing small amounts of labeled training data (the low-resource setting). In this paper, motivated by the observation that dialog states could be viewed as Markov states, we propose to build Markovian Generative Architectures (MGA) over PLM backbones for efficient TOD systems. Experiments on MultiWOZ2.1 show that in the rich-resource setting, the proposed Markov models reduce memory and time costs without performance degradation; in the low-resource setting, the training efficiency of the Markov models is more significant.

cs.CL↗
arXiv2022

Jointly Reinforced User Simulator and Task-oriented Dialog System with Simplified Generative Architecture

Recently, there has been progress in supervised funetuning pretrained GPT-2 to build end-to-end task-oriented dialog (TOD) systems. However, online reinforcement learning of a GPT-2 based dialog system (DS), together with a end-to-end user simulator (US), has not ever been explored. Moreover, a drawback with existing GPT-2 based TOD systems is that they mostly employ the whole dialog history as input, which brings inefficiencies in memory and compute. In this paper, we first propose Simplified Generative Architectures (SGA) for DS and US respectively, both based on GPT-2 but using shortened history. Then, we successfully develop Jointly Reinforced US and DS, called SGA-JRUD. Our DS with the proposed SGA, when only supervised trained, achieves state-of-the-art performance on MultiWOZ2.1 and is more compute-efficient in both training and generation. Extensive experiments on MultiWOZ2.1 further show the superiority of SGA-JRUD in both offline and online evaluations.

cs.CL↗
arXiv2022

Snowmass Theory Frontier: Effective Field Theory

We summarize recent progress in the development, application, and understanding of effective field theories and highlight promising directions for future research. This Report is prepared as the TF02 "Effective Field Theory" topical group summary for the Theory Frontier as part of the Snowmass 2021 process.

hep-ph↗
arXiv2022

Prototype Design and Efficiency Analysis of a Novel Robot Drive Based on 3K-H-V Topology

Robot actuators directly affect the performance of robots, and robot drives directly affect the performance of robot actuators. With the development of robotics, robots have put higher requirements on robot drives, such as high stiffness, high accuracy, high loading, high efficiency, low backlash, compact size, and hollow structure. In order to meet the demand development of robot actuators, this research base proposes a new robot drive based on 3K-H-V topology using involute and cycloidal gear shapes, planetary cycloidal drive, from the perspective of drive topology and through the design idea of decoupling. In this study, the reduction ratio and the efficiency model of the 3K-H-V topology were analyzed, and a prototype planetary cycloidal actuator was designed. The feasibility of the drive is initially verified by experimentally concluding that the PCA has a hollow structure, compact size, and high torque density (69 kg/Nm).

cs.RO↗
arXiv2022

Black-Box Dissector: Towards Erasing-based Hard-Label Model Stealing Attack

Previous studies have verified that the functionality of black-box models can be stolen with full probability outputs. However, under the more practical hard-label setting, we observe that existing methods suffer from catastrophic performance degradation. We argue this is due to the lack of rich information in the probability prediction and the overfitting caused by hard labels. To this end, we propose a novel hard-label model stealing method termed \emph{black-box dissector}, which consists of two erasing-based modules. One is a CAM-driven erasing strategy that is designed to increase the information capacity hidden in hard labels from the victim model. The other is a random-erasing-based self-knowledge distillation module that utilizes soft labels from the substitute model to mitigate overfitting. Extensive experiments on four widely-used datasets consistently demonstrate that our method outperforms state-of-the-art methods, with an improvement of at most $8.27\%$. We also validate the effectiveness and practical potential of our method on real-world APIs and defense methods. Furthermore, our method promotes other downstream tasks, \emph{i.e.}, transfer adversarial attacks.

cs.CV↗
arXiv2022

A solution to Erdős and Hajnal's odd cycle problem

In 1981, Erdős and Hajnal asked whether the sum of the reciprocals of the odd cycle lengths in a graph with infinite chromatic number is necessarily infinite. Let $\mathcal{C}(G)$ be the set of cycle lengths in a graph $G$ and let $\mathcal{C}_\text{odd}(G)$ be the set of odd numbers in $\mathcal{C}(G)$. We prove that, if $G$ has chromatic number $k$, then $\sum_{\ell\in \mathcal{C}_\text{odd}(G)}1/\ell\geq (1/2-o_k(1))\log k$. This solves Erdős and Hajnal's odd cycle problem, and, furthermore, this bound is asymptotically optimal. In 1984, Erdős asked whether there is some $d$ such that each graph with chromatic number at least $d$ (or perhaps even only average degree at least $d$) has a cycle whose length is a power of 2. We show that an average degree condition is sufficient for this problem, solving it with methods that apply to a wide range of sequences in addition to the powers of 2. Finally, we use our methods to show that, for every $k$, there is some $d$ so that every graph with average degree at least $d$ has a subdivision of the complete graph $K_k$ in which each edge is subdivided the same number of times. This confirms a conjecture of Thomassen from 1984.

math.CO↗
arXiv2022

PetLock:A Genderless and Standard Interface for the Future On-orbit Construction

Modular design is the foundation of on orbit construction technology of large space facilities in the future.Standard interface is the key technology of modular design of the future space robotic systems and space facilities.This paper presents the designed and tested of PetLock,a standard and genderless interface which can transfer mechanical loads,power and data between the future modular space robotic manipulator and spacecraft.PetLock adopts a completely genderless design,including connection face,locking mechanism,data and power interface.The connection surface provides a large translation and rotation misalignment tolerance,due to its 120-degree symmetrical and 3D shape design.The locking mechanism features the three locking pins retraction structure design,which is simple and reliable.POGO pin connectors in the center of the interface provides the power and data transfer capabilities.Due to the advantages of high locking force,large tolerance,high reliability and low cost,PetLock has the very big application potential in future on orbit construction missions.

cs.RO↗
arXiv2022

A Combined Inverse Kinematics Algorithm Using FABRIK with Optimization

Forward and backward reaching inverse kinematics (FABRIK) is a heuristic inverse kinematics solver that is gradually applied to manipulators with the advantages of fast convergence and generating more realistic configurations. However, under the high error constraint, FABRIK exhibits unstable convergence behavior, which is unsatisfactory for the real-time motion planning of manipulators. In this paper, a novel inverse kinematics algorithm that combines FABRIK and the sequential quadratic programming (SQP) algorithm is presented, in which the joint angles deduced by FABRIK will be taken as the initial seed of the SQP algorithm to avoid getting stuck in local minima. The combined algorithm is evaluated with experiments, in which our algorithm can achieve higher success rates and faster solution times than FABRIK under the high error constraint. Furthermore, the combined algorithm can generate continuous trajectories for the UR5 and KUKA LBR IIWA 14 R820 manipulators in path tracking with no pose error and permitted position error of the end-effector.

cs.RO↗
arXiv2022

Identity-Sensitive Knowledge Propagation for Cloth-Changing Person Re-identification

Cloth-changing person re-identification (CC-ReID), which aims to match person identities under clothing changes, is a new rising research topic in recent years. However, typical biometrics-based CC-ReID methods often require cumbersome pose or body part estimators to learn cloth-irrelevant features from human biometric traits, which comes with high computational costs. Besides, the performance is significantly limited due to the resolution degradation of surveillance images. To address the above limitations, we propose an effective Identity-Sensitive Knowledge Propagation framework (DeSKPro) for CC-ReID. Specifically, a Cloth-irrelevant Spatial Attention module is introduced to eliminate the distraction of clothing appearance by acquiring knowledge from the human parsing module. To mitigate the resolution degradation issue and mine identity-sensitive cues from human faces, we propose to restore the missing facial details using prior facial knowledge, which is then propagated to a smaller network. After training, the extra computations for human parsing or face restoration are no longer required. Extensive experiments show that our framework outperforms state-of-the-art methods by a large margin. Our code is available at https://github.com/KimbingNg/DeskPro.

cs.CV↗
arXiv2022

Mission Apollo: Landing Optical Circuit Switching at Datacenter Scale

In this paper, we describe Apollo, to the best of our knowledge, the world's first large-scale production deployment of optical circuit switches (OCSes) for datacenter networking. We will first describe the infrastructure challenges and use cases that motivated optical switching inside datacenters. We then delve into the requirements of OCSes for datacenter applications: balancing cost, port count, switching time, and optical performance, which drive design choices and implementation details of our internally developed 3D MEMS-based OCS. To enable the Apollo optical switching layer, we employ circulators to realize bidirectional links through the OCS, effectively doubling the OCS radix. The OCS and circulator design choices were critical for meeting network bandwidth, scale, and cost targets. We review the critical co-design of WDM transceiver technology for these OCS plus circulator-based bidirectional links and their corresponding physical impairments, delivered over four generations/speeds of optical interconnect. Finally, we conclude with thoughts on future directions in hardware development and associated applications.

cs.NI↗
arXiv2022

Shape of the asymptotic maximum sum-free sets in integer lattice grids

We determine the shape of all sum-free sets in $\{1,2,\ldots,n\}^2$ of size close to the maximum $\frac{3}{5}n^2$, solving a problem of Elsholtz and Rackham. We show that all such asymptotic maximum sum-free sets lie completely in the stripe $\frac{4}{5}n-o(n)\le x+y\le\frac{8}{5}n+ o(n)$. We also determine for any positive integer $p$ the maximum size of a subset $A\subseteq \{1,2,\ldots,n\}^2$ which forbids the triple $(x,y,z)$ satisfying $px+py=z$.

math.CO↗
arXiv2022

Clique immersion in graphs without fixed bipartite graph

A graph $G$ contains $H$ as an \emph{immersion} if there is an injective mapping $ϕ: V(H)\rightarrow V(G)$ such that for each edge $uv\in E(H)$, there is a path $P_{uv}$ in $G$ joining vertices $ϕ(u)$ and $ϕ(v)$, and all the paths $P_{uv}$, $uv\in E(H)$, are pairwise edge-disjoint. An analogue of Hadwiger's conjecture for the clique immersions by Lescure and Meyniel, and independently by Abu-Khzam and Langston, states that every graph $G$ contains $K_{χ(G)}$ as an immersion. We prove that for any constant $\varepsilon>0$ and integers $s,t\ge2$, there exists $d_0=d_0(\varepsilon,s,t)$ such that every $K_{s,t}$-free graph $G$ with $d(G)\ge d_0$ contains a clique immersion of order $(1-\varepsilon)d(G)$. This implies that the above-mentioned conjecture is asymptotically true for graphs without a fixed complete bipartite graph.

math.CO↗
arXiv2022

Crux and long cycles in graphs

We introduce a notion of the \emph{crux} of a graph $G$, measuring the order of a smallest dense subgraph in $G$. This simple-looking notion leads to some generalisations of known results about cycles, offering an interesting paradigm of `replacing average degree by crux'. In particular, we prove that \emph{every} graph contains a cycle of length linear in its crux. Long proved that every subgraph of a hypercube $Q^m$ (resp. discrete torus $C_3^m$) with average degree $d$ contains a path of length $2^{d/2}$ (resp. $2^{d/4}$), and conjectured that there should be a path of length $2^{d}-1$ (resp. $3^{d/2}-1$). As a corollary of our result, together with isoperimetric inequalities, we close these exponential gaps giving asymptotically optimal bounds on long paths in hypercubes, discrete tori, and more generally Hamming graphs. We also consider random subgraphs of $C_4$-free graphs and hypercubes, proving near optimal bounds on lengths of long cycles.

math.CO↗
arXiv2022

Advancing Semi-Supervised Task Oriented Dialog Systems by JSA Learning of Discrete Latent Variable Models

Developing semi-supervised task-oriented dialog (TOD) systems by leveraging unlabeled dialog data has attracted increasing interests. For semi-supervised learning of latent state TOD models, variational learning is often used, but suffers from the annoying high-variance of the gradients propagated through discrete latent variables and the drawback of indirectly optimizing the target log-likelihood. Recently, an alternative algorithm, called joint stochastic approximation (JSA), has emerged for learning discrete latent variable models with impressive performances. In this paper, we propose to apply JSA to semi-supervised learning of the latent state TOD models, which is referred to as JSA-TOD. To our knowledge, JSA-TOD represents the first work in developing JSA based semi-supervised learning of discrete latent variable conditional models for such long sequential generation problems like in TOD systems. Extensive experiments show that JSA-TOD significantly outperforms its variational learning counterpart. Remarkably, semi-supervised JSA-TOD using 20% labels performs close to the full-supervised baseline on MultiWOZ2.1.

cs.CL↗
arXiv2022

Spatiotemporal Propagation Learning for Network-Wide Flight Delay Prediction

Demystifying the delay propagation mechanisms among multiple airports is fundamental to precise and interpretable delay prediction, which is crucial during decision-making for all aviation industry stakeholders. The principal challenge lies in effectively leveraging the spatiotemporal dependencies and exogenous factors related to the delay propagation. However, previous works only consider limited spatiotemporal patterns with few factors. To promote more comprehensive propagation modeling for delay prediction, we propose SpatioTemporal Propagation Network (STPN), a space-time separable graph convolutional network, which is novel in spatiotemporal dependency capturing. From the aspect of spatial relation modeling, we propose a multi-graph convolution model considering both geographic proximity and airline schedule. From the aspect of temporal dependency capturing, we propose a multi-head self-attentional mechanism that can be learned end-to-end and explicitly reason multiple kinds of temporal dependency of delay time series. We show that the joint spatial and temporal learning models yield a sum of the Kronecker product, which factors the spatiotemporal dependence into the sum of several spatial and temporal adjacency matrices. By this means, STPN allows cross-talk of spatial and temporal factors for modeling delay propagation. Furthermore, a squeeze and excitation module is added to each layer of STPN to boost meaningful spatiotemporal features. To this end, we apply STPN to multi-step ahead arrival and departure delay prediction in large-scale airport networks. To validate the effectiveness of our model, we experiment with two real-world delay datasets, including U.S and China flight delays; and we show that STPN outperforms state-of-the-art methods. In addition, counterfactuals produced by STPN show that it learns explainable delay propagation patterns.

cs.LG↗