arXiv subjects

Hao Shen

Publications and source records attributed to Hao Shen.

At least 91 records · Page 5Linked to original sources

arXiv2022

Hand and Arm Gesture-based Human-Robot Interaction: A Review

The study of Human-Robot Interaction (HRI) aims to create close and friendly communication between humans and robots. In the human-center HRI, an essential aspect of implementing a successful and effective HRI is building a natural and intuitive interaction, including verbal and nonverbal. As a prevalent nonverbally communication approach, hand and arm gesture communication happen ubiquitously in our daily life. A considerable amount of work on gesture-based HRI is scattered in various research domains. However, a systematic understanding of the works on gesture-based HRI is still lacking. This paper intends to provide a comprehensive review of gesture-based HRI and focus on the advanced finding in this area. Following the stimulus-organism-response framework, this review consists of: (i) Generation of human gesture(stimulus). (ii) Robot recognition of human gesture(organism). (iii) Robot reaction to human gesture(response). Besides, this review summarizes the research status of each element in the framework and analyze the advantages and disadvantages of related works. Toward the last part, this paper discusses the current research challenges on gesture-based HRI and provides possible future directions.

cs.HC↗

arXiv2022

Learning Category-Level Generalizable Object Manipulation Policy via Generative Adversarial Self-Imitation Learning from Demonstrations

Generalizable object manipulation skills are critical for intelligent and multi-functional robots to work in real-world complex scenes. Despite the recent progress in reinforcement learning, it is still very challenging to learn a generalizable manipulation policy that can handle a category of geometrically diverse articulated objects. In this work, we tackle this category-level object manipulation policy learning problem via imitation learning in a task-agnostic manner, where we assume no handcrafted dense rewards but only a terminal reward. Given this novel and challenging generalizable policy learning problem, we identify several key issues that can fail the previous imitation learning algorithms and hinder the generalization to unseen instances. We then propose several general but critical techniques, including generative adversarial self-imitation learning from demonstrations, progressive growing of discriminator, and instance-balancing for expert buffer, that accurately pinpoints and tackles these issues and can benefit category-level manipulation policy learning regardless of the tasks. Our experiments on ManiSkill benchmarks demonstrate a remarkable improvement on all tasks and our ablation studies further validate the contribution of each proposed technique.

cs.RO↗

arXiv2022

A stochastic PDE approach to large N problems in quantum field theory: a survey

In this survey we review some recent rigorous results on large N problems in quantum field theory, stochastic quantization and singular stochastic PDEs, and their mean field limit problems. In particular we discuss the O(N) linear sigma model on two and three dimensional torus. The stochastic quantization procedure leads to a coupled system of N interacting $Φ^4$ equations. In d = 2, we show uniform in N bounds for the dynamics and convergence to a mean-field singular SPDE. For large enough mass or small enough coupling, the invariant measures (i.e. the O(N) linear sigma model) converge to the massive Gaussian free field, the unique invariant measure of the mean-field dynamics, in a Wasserstein distance. We also obtain tightness for certain O(N) invariant observables as random fields in suitable Besov spaces as $N\to \infty$, along with exact descriptions of the limiting correlations. In d = 3, the estimates become more involved since the equation is more singular. We discuss in this case how to prove convergence to the massive Gaussian free field. The proofs of these results build on the recent progress of singular SPDE theory and combine many new techniques such as uniform in N estimates and dynamical mean field theory. These are based on joint papers with Scott Smith, Rongchan Zhu and Xiangchan Zhu.

math.PR↗

arXiv2022

Large $N$ limit of the $O(N)$ linear sigma model in 3D

In this paper we study the large N limit of the $O(N)$-invariant linear sigma model, which is a vector-valued generalization of the $Φ^4$ quantum field theory, on the three dimensional torus. We study the problem via its stochastic quantization, which yields a coupled system of N interacting SPDEs. We prove tightness of the invariant measures in the large N limit. For large enough mass or small enough coupling constant, they converge to the (massive) Gaussian free field at a rate of order $1/\sqrt N$ with respect to the Wasserstein distance. We also obtain tightness results for certain $O(N)$ invariant observables. These generalize some of the results in \cite{SSZZ20} from two dimensions to three dimensions. The proof leverages the method recently developed by \cite{GH18} and combines many new techniques such as uniform in $N$ estimates on perturbative objects as well as the solutions.

math.PR↗

arXiv2022

Langevin dynamic for the 2D Yang-Mills measure

We define a natural state space and Markov process associated to the stochastic Yang-Mills heat flow in two dimensions. To accomplish this we first introduce a space of distributional connections for which holonomies along sufficiently regular curves (Wilson loop observables) and the action of an associated group of gauge transformations are both well-defined and satisfy good continuity properties. The desired state space is obtained as the corresponding space of orbits under this group action and is shown to be a Polish space when equipped with a natural Hausdorff metric. To construct the Markov process we show that the stochastic Yang-Mills heat flow takes values in our space of connections and use the "DeTurck trick" of introducing a time dependent gauge transformation to show invariance, in law, of the solution under gauge transformations. Our main tool for solving for the Yang-Mills heat flow is the theory of regularity structures and along the way we also develop a "basis-free" framework for applying the theory of regularity structures in the context of vector-valued noise - this provides a conceptual framework for interpreting several previous constructions and we expect this framework to be of independent interest.

math.PR↗

arXiv2022

A stochastic analysis approach to lattice Yang--Mills at strong coupling

We develop a new stochastic analysis approach to the lattice Yang--Mills model at strong coupling in any dimension $d>1$, with t' Hooft scaling $βN$ for the inverse coupling strength. We study their Langevin dynamics, ergodicity, functional inequalities, large $N$ limits, and mass gap. Assuming $|β| < \frac{N-2}{32(d-1)N}$ for the structure group $SO(N)$, or $|β| < \frac{1}{16(d-1)}$ for $SU(N)$, we prove the following results. The invariant measure for the corresponding Langevin dynamic is unique on the entire lattice, and the dynamic is exponentially ergodic under a Wasserstein distance. The finite volume Yang--Mills measures converge to this unique invariant measure in the infinite volume limit, for which Log-Sobolev and Poincaré inequalities hold. These functional inequalities imply that the suitably rescaled Wilson loops for the infinite volume measure has factorized correlations and converges in probability to deterministic limits in the large $N$ limit, and correlations of a large class of observables decay exponentially, namely the infinite volume measure has a strictly positive mass gap. Our method improves earlier results or simplifies the proofs, and provides some new perspectives to the study of lattice Yang--Mills model.

math.PR↗

arXiv2022

3D-SPS: Single-Stage 3D Visual Grounding via Referred Point Progressive Selection

3D visual grounding aims to locate the referred target object in 3D point cloud scenes according to a free-form language description. Previous methods mostly follow a two-stage paradigm, i.e., language-irrelevant detection and cross-modal matching, which is limited by the isolated architecture. In such a paradigm, the detector needs to sample keypoints from raw point clouds due to the inherent properties of 3D point clouds (irregular and large-scale), to generate the corresponding object proposal for each keypoint. However, sparse proposals may leave out the target in detection, while dense proposals may confuse the matching model. Moreover, the language-irrelevant detection stage can only sample a small proportion of keypoints on the target, deteriorating the target prediction. In this paper, we propose a 3D Single-Stage Referred Point Progressive Selection (3D-SPS) method, which progressively selects keypoints with the guidance of language and directly locates the target. Specifically, we propose a Description-aware Keypoint Sampling (DKS) module to coarsely focus on the points of language-relevant objects, which are significant clues for grounding. Besides, we devise a Target-oriented Progressive Mining (TPM) module to finely concentrate on the points of the target, which is enabled by progressive intra-modal relation modeling and inter-modal target mining. 3D-SPS bridges the gap between detection and matching in the 3D visual grounding task, localizing the target at a single stage. Experiments demonstrate that 3D-SPS achieves state-of-the-art performance on both ScanRefer and Nr3D/Sr3D datasets.

cs.CV↗

arXiv2022

Analysis and Optimisation of Bellman Residual Errors with Neural Function Approximation

Recent development of Deep Reinforcement Learning (DRL) has demonstrated superior performance of neural networks in solving challenging problems with large or even continuous state spaces. One specific approach is to deploy neural networks to approximate value functions by minimising the Mean Squared Bellman Error (MSBE) function. Despite great successes of DRL, development of reliable and efficient numerical algorithms to minimise the MSBE is still of great scientific interest and practical demand. Such a challenge is partially due to the underlying optimisation problem being highly non-convex or using incomplete gradient information as done in Semi-Gradient algorithms. In this work, we analyse the MSBE from a smooth optimisation perspective and develop an efficient Approximate Newton's algorithm. First, we conduct a critical point analysis of the error function and provide technical insights on optimisation and design choices for neural networks. When the existence of global minima is assumed and the objective fulfils certain conditions, suboptimal local minima can be avoided when using over-parametrised neural networks. We construct a Gauss Newton Residual Gradient algorithm based on the analysis in two variations. The first variation applies to discrete state spaces and exact learning. We confirm theoretical properties of this algorithm such as being locally quadratically convergent to a global minimum numerically. The second employs sampling and can be used in the continuous setting. We demonstrate feasibility and generalisation capabilities of the proposed algorithm empirically using continuous control problems and provide a numerical verification of our critical point analysis. We outline the difficulties of combining Semi-Gradient approaches with Hessian information. To benefit from second-order information complete derivatives of the MSBE must be considered during training.

cs.LG↗

arXiv2022

Learning from Attacks: Attacking Variational Autoencoder for Improving Image Classification

Adversarial attacks are often considered as threats to the robustness of Deep Neural Networks (DNNs). Various defending techniques have been developed to mitigate the potential negative impact of adversarial attacks against task predictions. This work analyzes adversarial attacks from a different perspective. Namely, adversarial examples contain implicit information that is useful to the predictions i.e., image classification, and treat the adversarial attacks against DNNs for data self-expression as extracted abstract representations that are capable of facilitating specific learning tasks. We propose an algorithmic framework that leverages the advantages of the DNNs for data self-expression and task-specific predictions, to improve image classification. The framework jointly learns a DNN for attacking Variational Autoencoder (VAE) networks and a DNN for classification, coined as Attacking VAE for Improve Classification (AVIC). The experiment results show that AVIC can achieve higher accuracy on standard datasets compared to the training with clean examples and the traditional adversarial training.

cs.LG↗

arXiv2021

Unsupervised Domain Adaptation with Dynamics-Aware Rewards in Reinforcement Learning

Unsupervised reinforcement learning aims to acquire skills without prior goal representations, where an agent automatically explores an open-ended environment to represent goals and learn the goal-conditioned policy. However, this procedure is often time-consuming, limiting the rollout in some potentially expensive target environments. The intuitive approach of training in another interaction-rich environment disrupts the reproducibility of trained skills in the target environment due to the dynamics shifts and thus inhibits direct transferring. Assuming free access to a source environment, we propose an unsupervised domain adaptation method to identify and acquire skills across dynamics. Particularly, we introduce a KL regularized objective to encourage emergence of skills, rewarding the agent for both discovering skills and aligning its behaviors respecting dynamics shifts. This suggests that both dynamics (source and target) shape the reward to facilitate the learning of adaptive skills. We also conduct empirical experiments to demonstrate that our method can effectively learn skills that can be smoothly deployed in target.

cs.LG↗

arXiv2021

End-to-End Video Instance Segmentation with Transformers

Video instance segmentation (VIS) is the task that requires simultaneously classifying, segmenting and tracking object instances of interest in video. Recent methods typically develop sophisticated pipelines to tackle this task. Here, we propose a new video instance segmentation framework built upon Transformers, termed VisTR, which views the VIS task as a direct end-to-end parallel sequence decoding/prediction problem. Given a video clip consisting of multiple image frames as input, VisTR outputs the sequence of masks for each instance in the video in order directly. At the core is a new, effective instance sequence matching and segmentation strategy, which supervises and segments instances at the sequence level as a whole. VisTR frames the instance segmentation and tracking in the same perspective of similarity learning, thus considerably simplifying the overall pipeline and is significantly different from existing approaches. Without bells and whistles, VisTR achieves the highest speed among all existing VIS models, and achieves the best result among methods using single model on the YouTube-VIS dataset. For the first time, we demonstrate a much simpler and faster video instance segmentation framework built upon Transformers, achieving competitive accuracy. We hope that VisTR can motivate future research for more video understanding tasks.

cs.CV↗

arXiv2021

Investigation of the phase occurrence and H sorption properties in the Y33.33Ni66.67-xAlx (0 <= x <= 33.33) system

The Y33.33Ni66.67-xAlx system has been investigated in the region 0 <= x <= 33.3. The alloys were synthesized by induction melting. Phase occurrence and structural properties were studied by X-Ray powder Diffraction (XRD). The Al solubility in each phase has been investigated by XRD and Electron Probe Micro-Analysis (EPMA). The hydrogenation properties were characterized by pressure-composition isotherm measurements and kinetic curves at 473 K. For x = 0, the binary Y33.33Ni66.67 alloy crystallizes in the cubic superstructure with F4-3m space group and ordered Y vacancies. For 1.67 <= x <= 8.33, the alloys contain mainly Y(Ni, Al)2 and Y(Ni, Al)3 pseudobinary phases; while for 16.67 <= x <= 33.33 they are mainly formed by ternary line compounds Y3Ni6Al2, Y2Ni2Al and YNiAl. Contrary to the binary Y33.33Ni66.67, Y(Ni, Al)2 pseudo-binary compounds crystalize in C15 phase (space group Fd-3m ) with disordered Y vacancies. The solubility limit of Al in the C15 YNi2-yAly phase is y = 0.11 (i.e., x = 3.67). The Y(Ni, Al)3 phase changes from rhombohedral (PuNi3-type, R-3m) to hexagonal (CeNi3-type, P63/mmc) structure for x increasing from 5.00 to 6.67. Upon hydrogenation, the disproportion of C15 Y(Ni, Al)2 and losses of crystallinity of YNi and Y2Ni2Al are the main reasons causing capacity decay of Y33.33Ni66.67-xAlx (0 <= x <= 33.33) alloys upon cycling.

cond-mat.mtrl-sci↗

arXiv2021

First-Generation Inference Accelerator Deployment at Facebook

In this paper, we provide a deep dive into the deployment of inference accelerators at Facebook. Many of our ML workloads have unique characteristics, such as sparse memory accesses, large model sizes, as well as high compute, memory and network bandwidth requirements. We co-designed a high-performance, energy-efficient inference accelerator platform based on these requirements. We describe the inference accelerator platform ecosystem we developed and deployed at Facebook: both hardware, through Open Compute Platform (OCP), and software framework and tooling, through Pytorch/Caffe2/Glow. A characteristic of this ecosystem from the start is its openness to enable a variety of AI accelerators from different vendors. This platform, with six low-power accelerator cards alongside a single-socket host CPU, allows us to serve models of high complexity that cannot be easily or efficiently run on CPUs. We describe various performance optimizations, at both platform and accelerator level, which enables this platform to serve production traffic at Facebook. We also share deployment challenges, lessons learned during performance optimization, as well as provide guidance for future inference hardware co-design.

cs.AR↗

arXiv2021

Identifying Influential Users in Unknown Social Networks for Adaptive Incentive Allocation Under Budget Restriction

In recent years, recommendation systems have been widely applied in many domains. These systems are impotent in affecting users to choose the behavior that the system expects. Meanwhile, providing incentives has been proven to be a more proactive way to affect users' behaviors. Due to the budget limitation, the number of users who can be incentivized is restricted. In this light, we intend to utilize social influence existing among users to enhance the effect of incentivization. Through incentivizing influential users directly, their followers in the social network are possibly incentivized indirectly. However, in many real-world scenarios, the topological structure of the network is usually unknown, which makes identifying influential users difficult. To tackle the aforementioned challenges, in this paper, we propose a novel algorithm for exploring influential users in unknown networks, which can estimate the influential relationships among users based on their historical behaviors and without knowing the topology of the network. Meanwhile, we design an adaptive incentive allocation approach that determines incentive values based on users' preferences and their influence ability. We evaluate the performance of the proposed approaches by conducting experiments on both synthetic and real-world datasets. The experimental results demonstrate the effectiveness of the proposed approaches.

cs.SI↗

arXiv2021

Deep auxiliary learning for visual localization using colorization task

Visual localization is one of the most important components for robotics and autonomous driving. Recently, inspiring results have been shown with CNN-based methods which provide a direct formulation to end-to-end regress 6-DoF absolute pose. Additional information like geometric or semantic constraints is generally introduced to improve performance. Especially, the latter can aggregate high-level semantic information into localization task, but it usually requires enormous manual annotations. To this end, we propose a novel auxiliary learning strategy for camera localization by introducing scene-specific high-level semantics from self-supervised representation learning task. Viewed as a powerful proxy task, image colorization task is chosen as complementary task that outputs pixel-wise color version of grayscale photograph without extra annotations. In our work, feature representations from colorization network are embedded into localization network by design to produce discriminative features for pose regression. Meanwhile an attention mechanism is introduced for the benefit of localization performance. Extensive experiments show that our model significantly improve localization accuracy over state-of-the-arts on both indoor and outdoor datasets.

cs.CV↗

arXiv2021

Scaling limit of a directed polymer among a Poisson field of independent walks

We consider a directed polymer model in dimension $1+1$, where the disorder is given by the occupation field of a Poisson system of independent random walks on $\mathbb Z$. In a suitable continuum and weak disorder limit, we show that the family of quenched partition functions of the directed polymer converges to the Stratonovich solution of a multiplicative stochastic heat equation (SHE) with a Gaussian noise, whose space-time covariance is given by the heat kernel. In contrast to the case with space-time white noise where the solution of the SHE admits a Wiener-Itô chaos expansion, we establish an $L^1$-convergent chaos expansions of iterated integrals generated by Picard iterations. Using this expansion and its discrete counterpart for the polymer partition functions, the convergence of the terms in the expansion is proved via functional analytic arguments and heat kernel estimates. The Poisson random walk system is amenable to careful moment analysis, which is an important input to our arguments.

math.PR↗

arXiv2021

Metapath- and Entity-aware Graph Neural Network for Recommendation

In graph neural networks (GNNs), message passing iteratively aggregates nodes' information from their direct neighbors while neglecting the sequential nature of multi-hop node connections. Such sequential node connections e.g., metapaths, capture critical insights for downstream tasks. Concretely, in recommender systems (RSs), disregarding these insights leads to inadequate distillation of collaborative signals. In this paper, we employ collaborative subgraphs (CSGs) and metapaths to form metapath-aware subgraphs, which explicitly capture sequential semantics in graph structures. We propose meta\textbf{P}ath and \textbf{E}ntity-\textbf{A}ware \textbf{G}raph \textbf{N}eural \textbf{N}etwork (PEAGNN), which trains multilayer GNNs to perform metapath-aware information aggregation on such subgraphs. This aggregated information from different metapaths is then fused using attention mechanism. Finally, PEAGNN gives us the representations for node and subgraph, which can be used to train MLP for predicting score for target user-item pairs. To leverage the local structure of CSGs, we present entity-awareness that acts as a contrastive regularizer on node embedding. Moreover, PEAGNN can be combined with prominent layers such as GAT, GCN and GraphSage. Our empirical evaluation shows that our proposed technique outperforms competitive baselines on several datasets for recommendation tasks. Further analysis demonstrates that PEAGNN also learns meaningful metapath combinations from a given set of metapaths.

cs.LG↗

arXiv2021

Dynamic Texture Recognition via Nuclear Distances on Kernelized Scattering Histogram Spaces

Distance-based dynamic texture recognition is an important research field in multimedia processing with applications ranging from retrieval to segmentation of video data. Based on the conjecture that the most distinctive characteristic of a dynamic texture is the appearance of its individual frames, this work proposes to describe dynamic textures as kernelized spaces of frame-wise feature vectors computed using the Scattering transform. By combining these spaces with a basis-invariant metric, we get a framework that produces competitive results for nearest neighbor classification and state-of-the-art results for nearest class center classification.

cs.CV↗