arXiv subjects

Donglin Wang

Publications and source records attributed to Donglin Wang.

At least 145 records · Page 8Linked to original sources

arXiv2023

Evaluating the Impact of Numerology and Retransmission on 5G NR V2X Communication at A System-Level Simulation

In recent years, Vehicle-to-Everything (V2X) communication opens an ample amount of opportunities to increase the safety of drivers and passengers and improve traffic efficiency which enables direct communication between vehicles. The Third Generation Partnership Project (3GPP) has specified a 5G New Radio (NR) Sidelink (SL) PC5 interface for supporting Cellular V2X (C-V2X) communication in Release 16 in 2017. 5G NR V2X communication is expected to provide high reliability, extra-low latency, and a high data rate for vehicular networks. In this paper, the newly introduced features of 5G NR standards like flexible numerology, variable Subcarrier Spacing (SCS), and allocated Physical Resource Blocks (PRBs) have been inspected in 5G NR V2X communications. Moreover, the 5G NR V2X data packet will be distributed to all nearby User Equipment (UE) by the Transmitter (Tx). However, there may be instances where certain UEs fail to receive the data packets in a single transmission. Unfortunately, the SL Tx lacks a feedback channel to verify if the Receivers (Rxs) have received the information. To meet the stringent reliability and latency requirements of C-V2X communication, we suggest and assess a retransmission scheme along with a scheme that incorporates varying resource allocations for retransmission in NR V2X communication. The effect of retransmission schemes on NR V2X communication systems has been detected.

eess.SY↗

arXiv2023

STRAPPER: Preference-based Reinforcement Learning via Self-training Augmentation and Peer Regularization

Preference-based reinforcement learning (PbRL) promises to learn a complex reward function with binary human preference. However, such human-in-the-loop formulation requires considerable human effort to assign preference labels to segment pairs, hindering its large-scale applications. Recent approache has tried to reuse unlabeled segments, which implicitly elucidates the distribution of segments and thereby alleviates the human effort. And consistency regularization is further considered to improve the performance of semi-supervised learning. However, we notice that, unlike general classification tasks, in PbRL there exits a unique phenomenon that we defined as similarity trap in this paper. Intuitively, human can have diametrically opposite preferredness for similar segment pairs, but such similarity may trap consistency regularization fail in PbRL. Due to the existence of similarity trap, such consistency regularization improperly enhances the consistency possiblity of the model's predictions between segment pairs, and thus reduces the confidence in reward learning, since the augmented distribution does not match with the original one in PbRL. To overcome such issue, we present a self-training method along with our proposed peer regularization, which penalizes the reward model memorizing uninformative labels and acquires confident predictions. Empirically, we demonstrate that our approach is capable of learning well a variety of locomotion and robotic manipulation behaviors using different semi-supervised alternatives and peer regularization.

cs.LG↗

arXiv2023

Beyond OOD State Actions: Supported Cross-Domain Offline Reinforcement Learning

Offline reinforcement learning (RL) aims to learn a policy using only pre-collected and fixed data. Although avoiding the time-consuming online interactions in RL, it poses challenges for out-of-distribution (OOD) state actions and often suffers from data inefficiency for training. Despite many efforts being devoted to addressing OOD state actions, the latter (data inefficiency) receives little attention in offline RL. To address this, this paper proposes the cross-domain offline RL, which assumes offline data incorporate additional source-domain data from varying transition dynamics (environments), and expects it to contribute to the offline data efficiency. To do so, we identify a new challenge of OOD transition dynamics, beyond the common OOD state actions issue, when utilizing cross-domain offline data. Then, we propose our method BOSA, which employs two support-constrained objectives to address the above OOD issues. Through extensive experiments in the cross-domain offline RL setting, we demonstrate BOSA can greatly improve offline data efficiency: using only 10\% of the target data, BOSA could achieve {74.4\%} of the SOTA offline RL performance that uses 100\% of the target data. Additionally, we also show BOSA can be effortlessly plugged into model-based offline RL and noising data augmentation techniques (used for generating source-domain data), which naturally avoids the potential dynamics mismatch between target-domain data and newly generated source-domain data.

cs.LG↗

arXiv2023

Beyond Reward: Offline Preference-guided Policy Optimization

This study focuses on the topic of offline preference-based reinforcement learning (PbRL), a variant of conventional reinforcement learning that dispenses with the need for online interaction or specification of reward functions. Instead, the agent is provided with fixed offline trajectories and human preferences between pairs of trajectories to extract the dynamics and task information, respectively. Since the dynamics and task information are orthogonal, a naive approach would involve using preference-based reward learning followed by an off-the-shelf offline RL algorithm. However, this requires the separate learning of a scalar reward function, which is assumed to be an information bottleneck of the learning process. To address this issue, we propose the offline preference-guided policy optimization (OPPO) paradigm, which models offline trajectories and preferences in a one-step process, eliminating the need for separately learning a reward function. OPPO achieves this by introducing an offline hindsight information matching objective for optimizing a contextual policy and a preference modeling objective for finding the optimal context. OPPO further integrates a well-performing decision policy by optimizing the two objectives iteratively. Our empirical results demonstrate that OPPO effectively models offline preferences and outperforms prior competing baselines, including offline RL algorithms performed over either true or pseudo reward function specifications. Our code is available on the project website: https://sites.google.com/view/oppo-icml-2023 .

cs.LG↗

arXiv2023

Reference-Limited Compositional Zero-Shot Learning

Compositional zero-shot learning (CZSL) refers to recognizing unseen compositions of known visual primitives, which is an essential ability for artificial intelligence systems to learn and understand the world. While considerable progress has been made on existing benchmarks, we suspect whether popular CZSL methods can address the challenges of few-shot and few referential compositions, which is common when learning in real-world unseen environments. To this end, we study the challenging reference-limited compositional zero-shot learning (RL-CZSL) problem in this paper, i.e., given limited seen compositions that contain only a few samples as reference, unseen compositions of observed primitives should be identified. We propose a novel Meta Compositional Graph Learner (MetaCGL) that can efficiently learn the compositionality from insufficient referential information and generalize to unseen compositions. Besides, we build a benchmark with two new large-scale datasets that consist of natural images with diverse compositional labels, providing more realistic environments for RL-CZSL. Extensive experiments in the benchmarks show that our method achieves state-of-the-art performance in recognizing unseen compositions when reference is limited for compositional learning.

cs.CV↗

arXiv2023

BVIP Guiding System with Adaptability to Individual Differences

Guiding robots can not only detect close-range obstacles like other guiding tools, but also extend its range to perceive the environment when making decisions. However, most existing works over-simplified the interaction between human agents and robots, ignoring the differences between individuals, resulting in poor experiences for different users. To solve the problem, we propose a data-driven guiding system to cope with the effect brighten by individual differences. In our guiding system, we design a Human Motion Predictor (HMP) and a Robot Dynamics Model (RDM) based on deep neural network, the time convolutional network (TCN) is verified to have the best performance, to predict differences in interaction between different human agents and robots. To train our models, we collected datasets that records the interactions from different human agents. Moreover, given the predictive information of the specific user, we propose a waypoints selector that allows the robot to naturally adapt to the user's state changes, which are mainly reflected in the walking speed. We compare the performance of our models with previous works and achieve significant performance improvements. On this basis, our guiding system demonstrated good adaptability to different human agents. Our guiding system is deployed on a real quadruped robot to verify the practicability.

cs.RO↗

arXiv2023

VoP: Text-Video Co-operative Prompt Tuning for Cross-Modal Retrieval

Many recent studies leverage the pre-trained CLIP for text-video cross-modal retrieval by tuning the backbone with additional heavy modules, which not only brings huge computational burdens with much more parameters, but also leads to the knowledge forgetting from upstream models. In this work, we propose the VoP: Text-Video Co-operative Prompt Tuning for efficient tuning on the text-video retrieval task. The proposed VoP is an end-to-end framework with both video & text prompts introducing, which can be regarded as a powerful baseline with only 0.1% trainable parameters. Further, based on the spatio-temporal characteristics of videos, we develop three novel video prompt mechanisms to improve the performance with different scales of trainable parameters. The basic idea of the VoP enhancement is to model the frame position, frame context, and layer function with specific trainable prompts, respectively. Extensive experiments show that compared to full fine-tuning, the enhanced VoP achieves a 1.4% average R@1 gain across five text-video retrieval benchmarks with 6x less parameter overhead. The code will be available at https://github.com/bighuang624/VoP.

cs.CV↗

arXiv2023

Effect of Variable Physical Numerologies on Link-Level Performance of 5G NR V2X

With technology and societal development, the 5th generation wireless communication (5G) contributes significantly to different societies like industries or academies. Vehicle-to-Everything (V2X) communication technology has been one of the leading services for 5G which has been applied in vehicles. It is used to exchange their status information with other traffic and traffic participants to increase traffic safety and efficiency. Cellular-V2X (C-V2X) is one of the emerging technologies to enable V2X communications. The first Long-Term Evolution (LTE) based C-V2X was released on the 3rd Generation Partnership Project (3GPP) standard. 3GPP is working towards the development of New Radio (NR) systems that it is called 5G NR V2X. One single numerology in LTE cannot satisfy most performance requirements because of the variety of deployment options and scenarios. For this reason, in order to meet the diverse requirements, the 5G NR Physical Layer (PHY) is designed to provide a highly flexible framework. Scalable Orthogonal Frequency-Division Multiplexing (OFDM) numerologies make flexibility possible. The term numerology refers to the PHY waveform parametrization and allows different Subcarrier Spacings (SCSs), symbols, and slot duration. This paper implements the Link-Level (LL) simulations of LTE C-V2X communication and 5G NR V2X communication where simulation results are used to compare similarities and differences between LTE and 5G NR. We detect the effect of variable PHY Numerologies of 5G NR on the LL performance of V2X. The simulation results show that the performance of 5G NR improved by using variable numerologies.

cs.NI↗

arXiv2023

RotoGBML: Towards Out-of-Distribution Generalization for Gradient-Based Meta-Learning

Gradient-based meta-learning (GBML) algorithms are able to fast adapt to new tasks by transferring the learned meta-knowledge, while assuming that all tasks come from the same distribution (in-distribution, ID). However, in the real world, they often suffer from an out-of-distribution (OOD) generalization problem, where tasks come from different distributions. OOD exacerbates inconsistencies in magnitudes and directions of task gradients, which brings challenges for GBML to optimize the meta-knowledge by minimizing the sum of task gradients in each minibatch. To address this problem, we propose RotoGBML, a novel approach to homogenize OOD task gradients. RotoGBML uses reweighted vectors to dynamically balance diverse magnitudes to a common scale and uses rotation matrixes to rotate conflicting directions close to each other. To reduce overhead, we homogenize gradients with the features rather than the network parameters. On this basis, to avoid the intervention of non-causal features (e.g., backgrounds), we also propose an invariant self-information (ISI) module to extract invariant causal features (e.g., the outlines of objects). Finally, task gradients are homogenized based on these invariant causal features. Experiments show that RotoGBML outperforms other state-of-the-art methods on various few-shot image classification benchmarks.

cs.CV↗

arXiv2023

Mitigating Unnecessary Handovers in Ultra-Dense Networks through Machine Learning-based Mobility Prediction

In 5G wireless communication, Intelligent Transportation Systems (ITS) and automobile applications, such as autonomous driving, are widely examined. These applications have strict requirements and often require high Quality of Service (QoS). In an urban setting, Ultra-Dense Networks (UDNs) have the potential to not only provide optimal QoS but also increase system capacity and frequency reuse. However, the current architecture of 5G UDN of dense Small Cell Nodes (SCNs) deployment prompts increased delay, handover times, and handover failures. In this paper, we propose a Machine Learning (ML) supported Mobility Prediction (MP) strategy to predict future Vehicle User Equipment (VUE) mobility and handover locations. The primary aim of the proposed methodology is to minimize Unnecessary Handover (UHO) while ensuring VUEs take full advantage of the deployed UDN. We evaluate and validate our approach on a downlink system-level simulator. We predict mobility using Support Vector Machine (SVM), Decision Tree Classifier (DTC), and Random Forest Classifier (RFC). The simulation results show an average reduction of 30% in handover times by utilizing ML-based MP, with RFC showing the most reduction up to 70% in some cases.

cs.NI↗

arXiv2023

Behavior Proximal Policy Optimization

Offline reinforcement learning (RL) is a challenging setting where existing off-policy actor-critic methods perform poorly due to the overestimation of out-of-distribution state-action pairs. Thus, various additional augmentations are proposed to keep the learned policy close to the offline dataset (or the behavior policy). In this work, starting from the analysis of offline monotonic policy improvement, we get a surprising finding that some online on-policy algorithms are naturally able to solve offline RL. Specifically, the inherent conservatism of these on-policy algorithms is exactly what the offline RL method needs to overcome the overestimation. Based on this, we propose Behavior Proximal Policy Optimization (BPPO), which solves offline RL without any extra constraint or regularization introduced compared to PPO. Extensive experiments on the D4RL benchmark indicate this extremely succinct method outperforms state-of-the-art offline RL algorithms. Our implementation is available at https://github.com/Dragon-Zhuang/BPPO.

cs.LG↗

arXiv2023

The Effect of Variable Factors on the Handover Performance for Ultra Dense Network

With wireless communication technology development, the 5G New Radio (NR) has been proposed and developed for a decade. This advanced mobile communication technology has more advancements, such as higher system capacity, higher spectrum efficiency, higher data rates, and so on. In 5G, Ultra- Dense Network (UDN) is deployed for increasing the system capacity and frequency reuse to meet high application requirements. The architecture of 5G UDN is to realize the dense and flexible deployment of smaller general Node B (gNB). However, the increased capacity of applying UDN in 5G is anticipated at the cost of increased signal interference, increased handover times, and increased handover failures. The Time to Trigger (TTT) is one of the most important factors in handover frequency which is deserved to be detected. Moreover, the density of the 5G gNBs influences the handover times and performance as well. In this work, we provide a compendium of 5G handover management. A downlink system-level simulator for 5G handover is built and utilized to evaluate the effect of different TTT values and densities of gNBs on the 5G handover. In addition, different velocities of Traffic Users (TUs) have been applied to the simulation system. From the simulation results, the handover performance has been analyzed and optimized by applying adjustable TTT under different densities of gNBs which will help people have a better understanding of the selection and effect of proper TTT, UDN, and different velocities of TUs on 5G handover performance.

cs.NI↗

arXiv2023

Graph based Environment Representation for Vision-and-Language Navigation in Continuous Environments

Vision-and-Language Navigation in Continuous Environments (VLN-CE) is a navigation task that requires an agent to follow a language instruction in a realistic environment. The understanding of environments is a crucial part of the VLN-CE task, but existing methods are relatively simple and direct in understanding the environment, without delving into the relationship between language instructions and visual environments. Therefore, we propose a new environment representation in order to solve the above problems. First, we propose an Environment Representation Graph (ERG) through object detection to express the environment in semantic level. This operation enhances the relationship between language and environment. Then, the relational representations of object-object, object-agent in ERG are learned through GCN, so as to obtain a continuous expression about ERG. Sequentially, we combine the ERG expression with object label embeddings to obtain the environment representation. Finally, a new cross-modal attention navigation framework is proposed, incorporating our environment representation and a special loss function dedicated to training ERG. Experimental result shows that our method achieves satisfactory performance in terms of success rate on VLN-CE tasks. Further analysis explains that our method attains better cross-modal matching and strong generalization ability.

cs.CV↗

arXiv2022

Few-shot Domain Adaptation for IMU Denoising

Different application scenarios will cause IMU to exhibit different error characteristics which will cause trouble to robot application. However, most data processing methods need to be designed for specific scenario. To solve this problem, we propose a few-shot domain adaptation method. In this work, a domain adaptation framework is considered for denoising the IMU, a reconstitution loss is designed to improve domain adaptability. In addition, in order to further improve the adaptability in the case of limited data, a few-shot training strategy is adopted. In the experiment, we quantify our method on two datasets (EuRoC and TUM-VI) and two real robots (car and quadruped robot) with three different precision IMUs. According to the experimental results, the adaptability of our framework is verified by t-SNE. In orientation results, our proposed method shows the great denoising performance.

cs.RO↗

arXiv2022

GDA-HIN: A Generalized Domain Adaptive Model across Heterogeneous Information Networks

Domain adaptation using graph-structured networks learns label-discriminative and network-invariant node embeddings by sharing graph parameters. Most existing works focus on domain adaptation of homogeneous networks. The few works that study heterogeneous cases only consider shared node types but ignore private node types in individual networks. However, for given source and target heterogeneous networks, they generally contain shared and private node types, where private types bring an extra challenge for graph domain adaptation. In this paper, we investigate Heterogeneous Information Networks (HINs) with both shared and private node types and propose a Generalized Domain Adaptive model across HINs (GDA-HIN) to handle the domain shift between them. GDA-HIN can not only align the distribution of identical-type nodes and edges in two HINs but also make full use of different-type nodes and edges to improve the performance of knowledge transfer. Extensive experiments on several datasets demonstrate that GDA-HIN can outperform state-of-the-art methods in various domain adaptation tasks across heterogeneous networks.

cs.LG↗

arXiv2022

Can Offline Reinforcement Learning Help Natural Language Understanding?

Pre-training has been a useful method for learning implicit transferable knowledge and it shows the benefit of offering complementary features across different modalities. Recent work mainly focuses on the modalities such as image and text, for example, studies show that visual features learned from images can help visual-grounded language understanding. In this paper, we consider investigating the potential connection between offline reinforcement learning (RL) and language modeling (LM). Intuitively, RL and LM are similar in predicting the next states based on the current and previous states, which rely on both local and long-range dependency across states. To validate such an assumption, we pre-trained different offline RL tasks using Transformer and then evaluate these models on various language-related tasks. Experimental results show that our RL pre-trained models can give close performance compared with the models using the LM training objective, showing that there exist common useful features across these two modalities. To further explore the potential relationship, we investigate some factors such as Markov property and the sequential nature of RL trajectory.

cs.CL↗

arXiv2022

KSG: Knowledge and Skill Graph

The knowledge graph (KG) is an essential form of knowledge representation that has grown in prominence in recent years. Because it concentrates on nominal entities and their relationships, traditional knowledge graphs are static and encyclopedic in nature. On this basis, event knowledge graph (Event KG) models the temporal and spatial dynamics by text processing to facilitate downstream applications, such as question-answering, recommendation and intelligent search. Existing KG research, on the other hand, mostly focuses on text processing and static facts, ignoring the vast quantity of dynamic behavioral information included in photos, movies, and pre-trained neural networks. In addition, no effort has been done to include behavioral intelligence information into the knowledge graph for deep reinforcement learning (DRL) and robot learning. In this paper, we propose a novel dynamic knowledge and skill graph (KSG), and then we develop a basic and specific KSG based on CN-DBpedia. The nodes are divided into entity and attribute nodes, with entity nodes containing the agent, environment, and skill (DRL policy or policy representation), and attribute nodes containing the entity description, pre-train network, and offline dataset. KSG can search for different agents' skills in various environments and provide transferable information for acquiring new skills. This is the first study that we are aware of that looks into dynamic KSG for skill retrieval and learning. Extensive experimental results on new skill learning show that KSG boosts new skill learning efficiency.

cs.AI↗

arXiv2022

Tree Structure-Aware Few-Shot Image Classification via Hierarchical Aggregation

In this paper, we mainly focus on the problem of how to learn additional feature representations for few-shot image classification through pretext tasks (e.g., rotation or color permutation and so on). This additional knowledge generated by pretext tasks can further improve the performance of few-shot learning (FSL) as it differs from human-annotated supervision (i.e., class labels of FSL tasks). To solve this problem, we present a plug-in Hierarchical Tree Structure-aware (HTS) method, which not only learns the relationship of FSL and pretext tasks, but more importantly, can adaptively select and aggregate feature representations generated by pretext tasks to maximize the performance of FSL tasks. A hierarchical tree constructing component and a gated selection aggregating component is introduced to construct the tree structure and find richer transferable knowledge that can rapidly adapt to novel classes with a few labeled images. Extensive experiments show that our HTS can significantly enhance multiple few-shot methods to achieve new state-of-the-art performance on four benchmark datasets. The code is available at: https://github.com/remiMZ/HTS-ECCV22.

cs.CV↗