Alert button
Picture for Mingyang Wang

Mingyang Wang

Alert button

Meta-Reinforcement Learning Based on Self-Supervised Task Representation Learning

Apr 29, 2023
Mingyang Wang, Zhenshan Bing, Xiangtong Yao, Shuai Wang, Hang Su, Chenguang Yang, Kai Huang, Alois Knoll

Figure 1 for Meta-Reinforcement Learning Based on Self-Supervised Task Representation Learning
Figure 2 for Meta-Reinforcement Learning Based on Self-Supervised Task Representation Learning
Figure 3 for Meta-Reinforcement Learning Based on Self-Supervised Task Representation Learning
Figure 4 for Meta-Reinforcement Learning Based on Self-Supervised Task Representation Learning

Meta-reinforcement learning enables artificial agents to learn from related training tasks and adapt to new tasks efficiently with minimal interaction data. However, most existing research is still limited to narrow task distributions that are parametric and stationary, and does not consider out-of-distribution tasks during the evaluation, thus, restricting its application. In this paper, we propose MoSS, a context-based Meta-reinforcement learning algorithm based on Self-Supervised task representation learning to address this challenge. We extend meta-RL to broad non-parametric task distributions which have never been explored before, and also achieve state-of-the-art results in non-stationary and out-of-distribution tasks. Specifically, MoSS consists of a task inference module and a policy module. We utilize the Gaussian mixture model for task representation to imitate the parametric and non-parametric task variations. Additionally, our online adaptation strategy enables the agent to react at the first sight of a task change, thus being applicable in non-stationary tasks. MoSS also exhibits strong generalization robustness in out-of-distributions tasks which benefits from the reliable and robust task representation. The policy is built on top of an off-policy RL algorithm and the entire network is trained completely off-policy to ensure high sample efficiency. On MuJoCo and Meta-World benchmarks, MoSS outperforms prior works in terms of asymptotic performance, sample efficiency (3-50x faster), adaptation efficiency, and generalization robustness on broad and diverse task distributions.

Viaarxiv icon

NLNDE at SemEval-2023 Task 12: Adaptive Pretraining and Source Language Selection for Low-Resource Multilingual Sentiment Analysis

Apr 28, 2023
Mingyang Wang, Heike Adel, Lukas Lange, Jannik Strötgen, Hinrich Schütze

Figure 1 for NLNDE at SemEval-2023 Task 12: Adaptive Pretraining and Source Language Selection for Low-Resource Multilingual Sentiment Analysis
Figure 2 for NLNDE at SemEval-2023 Task 12: Adaptive Pretraining and Source Language Selection for Low-Resource Multilingual Sentiment Analysis
Figure 3 for NLNDE at SemEval-2023 Task 12: Adaptive Pretraining and Source Language Selection for Low-Resource Multilingual Sentiment Analysis
Figure 4 for NLNDE at SemEval-2023 Task 12: Adaptive Pretraining and Source Language Selection for Low-Resource Multilingual Sentiment Analysis

This paper describes our system developed for the SemEval-2023 Task 12 "Sentiment Analysis for Low-resource African Languages using Twitter Dataset". Sentiment analysis is one of the most widely studied applications in natural language processing. However, most prior work still focuses on a small number of high-resource languages. Building reliable sentiment analysis systems for low-resource languages remains challenging, due to the limited training data in this task. In this work, we propose to leverage language-adaptive and task-adaptive pretraining on African texts and study transfer learning with source language selection on top of an African language-centric pretrained language model. Our key findings are: (1) Adapting the pretrained model to the target language and task using a small yet relevant corpus improves performance remarkably by more than 10 F1 score points. (2) Selecting source languages with positive transfer gains during training can avoid harmful interference from dissimilar languages, leading to better results in multilingual and cross-lingual settings. In the shared task, our system wins 8 out of 15 tracks and, in particular, performs best in the multilingual evaluation.

Viaarxiv icon

Formation Flight in Dense Environments

Oct 08, 2022
Lun Quan, Longji Yin, Tingrui Zhang, Mingyang Wang, Ruilin Wang, Sheng Zhong, Yanjun Cao, Chao Xu, Fei Gao

Figure 1 for Formation Flight in Dense Environments
Figure 2 for Formation Flight in Dense Environments
Figure 3 for Formation Flight in Dense Environments
Figure 4 for Formation Flight in Dense Environments

Formation flight has a vast potential for aerial robot swarms in various applications. However, existing methods lack the capability to achieve fully autonomous large-scale formation flight in dense environments. To bridge the gap, we present a complete formation flight system that effectively integrates real-world constraints into aerial formation navigation. This paper proposes a differentiable graph-based metric to quantify the overall similarity error between formations. This metric is invariant to rotation, translation, and scaling, providing more freedom for formation coordination. We design a distributed trajectory optimization framework that considers formation similarity, obstacle avoidance, and dynamic feasibility. The optimization is decoupled to make large-scale formation flights computationally feasible. To improve the elasticity of formation navigation in highly constrained scenes, we present a swarm reorganization method which adaptively adjusts the formation parameters and task assignments by generating local navigation goals. A novel swarm agreement strategy called global-remap-local-replan and a formation-level path planner is proposed in this work to coordinate the swarm global planning and local trajectory optimizations efficiently. To validate the proposed method, we design comprehensive benchmarks and simulations with other cutting-edge works in terms of adaptability, predictability, elasticity, resilience, and efficiency. Finally, integrated with palm-sized swarm platforms with onboard computers and sensors, the proposed method demonstrates its efficiency and robustness by achieving the largest scale formation flight in dense outdoor environments.

* Submitted for IEEE Transactions on Robotics 
Viaarxiv icon

Meeting-Merging-Mission: A Multi-robot Coordinate Framework for Large-Scale Communication-Limited Exploration

Sep 16, 2021
Yuman Gao, Yingjian Wang, Xingguang Zhong, Tiankai Yang, Mingyang Wang, Zhixiong Xu, Yongchao Wang, Chao Xu, Fei Gao

Figure 1 for Meeting-Merging-Mission: A Multi-robot Coordinate Framework for Large-Scale Communication-Limited Exploration
Figure 2 for Meeting-Merging-Mission: A Multi-robot Coordinate Framework for Large-Scale Communication-Limited Exploration
Figure 3 for Meeting-Merging-Mission: A Multi-robot Coordinate Framework for Large-Scale Communication-Limited Exploration
Figure 4 for Meeting-Merging-Mission: A Multi-robot Coordinate Framework for Large-Scale Communication-Limited Exploration

This letter presents a complete framework Meeting-Merging-Mission for multi-robot exploration under communication restriction. Considering communication is limited in both bandwidth and range in the real world, we propose a lightweight environment presentation method and an efficient cooperative exploration strategy. For lower bandwidth, each robot utilizes specific polytopes to maintains free space and super frontier information (SFI) as the source for exploration decision-making. To reduce repeated exploration, we develop a mission-based protocol that drives robots to share collected information in stable rendezvous. We also design a complete path planning scheme for both centralized and decentralized cases. To validate that our framework is practical and generic, we present an extensive benchmark and deploy our system into multi-UGV and multi-UAV platforms.

Viaarxiv icon

Height estimation from single aerial images using a deep ordinal regression network

Jun 04, 2020
Xiang Li, Mingyang Wang, Yi Fang

Figure 1 for Height estimation from single aerial images using a deep ordinal regression network
Figure 2 for Height estimation from single aerial images using a deep ordinal regression network
Figure 3 for Height estimation from single aerial images using a deep ordinal regression network
Figure 4 for Height estimation from single aerial images using a deep ordinal regression network

Understanding the 3D geometric structure of the Earth's surface has been an active research topic in photogrammetry and remote sensing community for decades, serving as an essential building block for various applications such as 3D digital city modeling, change detection, and city management. Previous researches have extensively studied the problem of height estimation from aerial images based on stereo or multi-view image matching. These methods require two or more images from different perspectives to reconstruct 3D coordinates with camera information provided. In this paper, we deal with the ambiguous and unsolved problem of height estimation from a single aerial image. Driven by the great success of deep learning, especially deep convolution neural networks (CNNs), some researches have proposed to estimate height information from a single aerial image by training a deep CNN model with large-scale annotated datasets. These methods treat height estimation as a regression problem and directly use an encoder-decoder network to regress the height values. In this paper, we proposed to divide height values into spacing-increasing intervals and transform the regression problem into an ordinal regression problem, using an ordinal loss for network training. To enable multi-scale feature extraction, we further incorporate an Atrous Spatial Pyramid Pooling (ASPP) module to extract features from multiple dilated convolution layers. After that, a post-processing technique is designed to transform the predicted height map of each patch into a seamless height map. Finally, we conduct extensive experiments on ISPRS Vaihingen and Potsdam datasets. Experimental results demonstrate significantly better performance of our method compared to the state-of-the-art methods.

* 5 pages, 3 figures 
Viaarxiv icon

Density-Aware Convolutional Networks with Context Encoding for Airborne LiDAR Point Cloud Classification

Oct 14, 2019
Xiang Li, Mingyang Wang, Congcong Wen, Lingjing Wang, Nan Zhou, Yi Fang

Figure 1 for Density-Aware Convolutional Networks with Context Encoding for Airborne LiDAR Point Cloud Classification
Figure 2 for Density-Aware Convolutional Networks with Context Encoding for Airborne LiDAR Point Cloud Classification
Figure 3 for Density-Aware Convolutional Networks with Context Encoding for Airborne LiDAR Point Cloud Classification
Figure 4 for Density-Aware Convolutional Networks with Context Encoding for Airborne LiDAR Point Cloud Classification

To better address challenging issues of the irregularity and inhomogeneity inherently present in 3D point clouds, researchers have been shifting their focus from the design of hand-craft point feature towards the learning of 3D point signatures using deep neural networks for 3D point cloud classification. Recent proposed deep learning based point cloud classification methods either apply 2D CNN on projected feature images or apply 1D convolutional layers directly on raw point sets. These methods cannot adequately recognize fine-grained local structures caused by the uneven density distribution of the point cloud data. In this paper, to address this challenging issue, we introduced a density-aware convolution module which uses the point-wise density to re-weight the learnable weights of convolution kernels. The proposed convolution module is able to fully approximate the 3D continuous convolution on unevenly distributed 3D point sets. Based on this convolution module, we further developed a multi-scale fully convolutional neural network with downsampling and upsampling blocks to enable hierarchical point feature learning. In addition, to regularize the global semantic context, we implemented a context encoding module to predict a global context encoding and formulated a context encoding regularizer to enforce the predicted context encoding to be aligned with the ground truth one. The overall network can be trained in an end-to-end fashion with the raw 3D coordinates as well as the height above ground as inputs. Experiments on the International Society for Photogrammetry and Remote Sensing (ISPRS) 3D labeling benchmark demonstrated the superiority of the proposed method for point cloud classification. Our model achieved a new state-of-the-art performance with an average F1 score of 71.2% and improved the performance by a large margin on several categories.

Viaarxiv icon