Searcharxiv⌕ Search

arXiv subjects

Xiao Ma

Publications and source records attributed to Xiao Ma.

At least 127 records · Page 7Linked to original sources
arXiv2022

RPM: Generalizable Behaviors for Multi-Agent Reinforcement Learning

Despite the recent advancement in multi-agent reinforcement learning (MARL), the MARL agents easily overfit the training environment and perform poorly in the evaluation scenarios where other agents behave differently. Obtaining generalizable policies for MARL agents is thus necessary but challenging mainly due to complex multi-agent interactions. In this work, we model the problem with Markov Games and propose a simple yet effective method, ranked policy memory (RPM), to collect diverse multi-agent trajectories for training MARL policies with good generalizability. The main idea of RPM is to maintain a look-up memory of policies. In particular, we try to acquire various levels of behaviors by saving policies via ranking the training episode return, i.e., the episode return of agents in the training environment; when an episode starts, the learning agent can then choose a policy from the RPM as the behavior policy. This innovative self-play training framework leverages agents' past policies and guarantees the diversity of multi-agent interaction in the training data. We implement RPM on top of MARL algorithms and conduct extensive experiments on Melting Pot. It has been demonstrated that RPM enables MARL agents to interact with unseen agents in multi-agent generalization evaluation scenarios and complete given tasks, and it significantly boosts the performance up to 402% on average.

cs.MA↗
arXiv2022

Boosting Offline Reinforcement Learning via Data Rebalancing

Offline reinforcement learning (RL) is challenged by the distributional shift between learning policies and datasets. To address this problem, existing works mainly focus on designing sophisticated algorithms to explicitly or implicitly constrain the learned policy to be close to the behavior policy. The constraint applies not only to well-performing actions but also to inferior ones, which limits the performance upper bound of the learned policy. Instead of aligning the densities of two distributions, aligning the supports gives a relaxed constraint while still being able to avoid out-of-distribution actions. Therefore, we propose a simple yet effective method to boost offline RL algorithms based on the observation that resampling a dataset keeps the distribution support unchanged. More specifically, we construct a better behavior policy by resampling each transition in an old dataset according to its episodic return. We dub our method ReD (Return-based Data Rebalance), which can be implemented with less than 10 lines of code change and adds negligible running time. Extensive experiments demonstrate that ReD is effective at boosting offline RL performance and orthogonal to decoupling strategies in long-tailed classification. New state-of-the-arts are achieved on the D4RL benchmark.

cs.LG↗
arXiv2022

From Earth to Space: A First Deployment of 5G Core Network on Satellite

Recent developments in the aerospace industry have led to a dramatic reduction in the manufacturing and launch costs of low Earth orbit satellites. The new trend enables the paradigm shift of satellite-terrestrial integrated networks with global coverage. In particular, the integration of 5G communication systems and satellites has the potential to restructure next-generation mobile networks. By leveraging the network function virtualization and network slicing, the orbital 5G core networks will facilitate the coordination and management of network functions in satellite-terrestrial integrated networks. We are the first to deploy a lightweight 5G core network on a real-world satellite to investigate its feasibility. We conducted experiments to validate the onboard 5G core network functions. The validated procedures include registration and session setup procedures. The results show that the 5G core network can function normally and generate correct signaling.

cs.NI↗
arXiv2022

Benchmarking of DL Libraries and Models on Mobile Devices

Deploying deep learning (DL) on mobile devices has been a notable trend in recent years. To support fast inference of on-device DL, DL libraries play a critical role as algorithms and hardware do. Unfortunately, no prior work ever dives deep into the ecosystem of modern DL libs and provides quantitative results on their performance. In this paper, we first build a comprehensive benchmark that includes 6 representative DL libs and 15 diversified DL models. We then perform extensive experiments on 10 mobile devices, which help reveal a complete landscape of the current mobile DL libs ecosystem. For example, we find that the best-performing DL lib is severely fragmented across different models and hardware, and the gap between those DL libs can be rather huge. In fact, the impacts of DL libs can overwhelm the optimizations from algorithms or hardware, e.g., model quantization and GPU/DSP-based heterogeneous computing. Finally, atop the observations, we summarize practical implications to different roles in the DL lib ecosystem.

cs.LG↗
arXiv2022

Almost sharp wave kinetic theory of multidimensional KdV type equations with $d\ge 3$

In this work, we study the random series expansion of a multidimensional KdV type equation with a diffusion term, the so-called Zakharov-Kuznetsov (ZK) equation. We impose random initial data and periodic boundary condition with period $L$ on this equation. Using the random series expansion, we derive the $3$-wave kinetic equation on the inertial range for $t\lesssim L^{-\varepsilon}T_{\text{kin}}$. Our result reaches kinetic time scale up to $\varepsilon$ loss.

math.AP↗
arXiv2022

Imitation Learning via Differentiable Physics

Existing imitation learning (IL) methods such as inverse reinforcement learning (IRL) usually have a double-loop training process, alternating between learning a reward function and a policy and tend to suffer long training time and high variance. In this work, we identify the benefits of differentiable physics simulators and propose a new IL method, i.e., Imitation Learning via Differentiable Physics (ILD), which gets rid of the double-loop design and achieves significant improvements in final performance, convergence speed, and stability. The proposed ILD incorporates the differentiable physics simulator as a physics prior into its computational graph for policy learning. It unrolls the dynamics by sampling actions from a parameterized policy, simply minimizing the distance between the expert trajectory and the agent trajectory, and back-propagating the gradient into the policy via temporal physics operators. With the physics prior, ILD policies can not only be transferable to unseen environment specifications but also yield higher final performance on a variety of tasks. In addition, ILD naturally forms a single-loop structure, which significantly improves the stability and training speed. To simplify the complex optimization landscape induced by temporal physics operations, ILD dynamically selects the learning objectives for each state during optimization. In our experiments, we show that ILD outperforms state-of-the-art methods in a variety of continuous control tasks with Brax, requiring only one expert demonstration. In addition, ILD can be applied to challenging deformable object manipulation tasks and can be generalized to unseen configurations.

cs.LG↗
arXiv2022

Transmission of Bernoulli Sources Using Convolutional LDGM Codes

We propose in this paper to exploit convolutional low density generator matrix (LDGM) codes for transmission of Bernoulli sources over binary-input output-symmetric (BIOS) channels. To this end, we present a new framework to prove the coding theorems for linear codes, which unifies the channel coding theorem, the source coding theorem and the joint source-channel coding (JSCC) theorem. In the presented framework, the systematic bits and the corresponding parity-check bits play different roles. Precisely, the noisy systematic bits are used to limit the list size of typical codewords, while the noisy parity-check bits are used to select from the list the maximum likelihood codeword. This new framework for linear codes allows that the systematic bits and the parity-check bits are transmitted in different ways and over different channels. With this framework, we prove that the Bernoulli generator matrix codes (BGMCs) are capacity-achieving over BIOS channels, entropy-achieving for Bernoulli sources, and also system-capacity-achieving for JSCC applications. A lower bound on the bit-error rate (BER) is derived for linear codes, which can be used to predict the error floors and hence serves as a simple tool to design the JSCC system. Numerical results show that the convolutional LDGM codes perform well in the waterfall region and match well with the derived error floors, which can be lowered down if required by simply increasing the encoding memory.

cs.IT↗
arXiv2022

Hierarchical Reinforcement Learning under Mixed Observability

The framework of mixed observable Markov decision processes (MOMDP) models many robotic domains in which some state variables are fully observable while others are not. In this work, we identify a significant subclass of MOMDPs defined by how actions influence the fully observable components of the state and how those, in turn, influence the partially observable components and the rewards. This unique property allows for a two-level hierarchical approach we call HIerarchical Reinforcement Learning under Mixed Observability (HILMO), which restricts partial observability to the top level while the bottom level remains fully observable, enabling higher learning efficiency. The top level produces desired goals to be reached by the bottom level until the task is solved. We further develop theoretical guarantees to show that our approach can achieve optimal and quasi-optimal behavior under mild assumptions. Empirical results on long-horizon continuous control tasks demonstrate the efficacy and efficiency of our approach in terms of improved success rate, sample efficiency, and wall-clock training time. We also deploy policies learned in simulation on a real robot.

cs.RO↗
arXiv2022

Spectral radius and rainbow matchings of graphs

Let $n,m$ be integers such that $1\leq m\leq (n-2)/2$ and let $[n]=\{1,\ldots,n\}$. Let $\mathcal{G}=\{G_1,\ldots,G_{m+1}\}$ be a family of graphs on the same vertex set $[n]$. In this paper, we prove that if for any $i\in [m+1]$, the spectral radius of $G_i$ is not less than $\max\{2m,\frac{1}{2}(m-1+\sqrt{(m-1)^2+4m(n-m)})\}$, then $\mathcal{G}$ admits a rainbow matching, i.e. a choice of disjoint edges $e_i\in G_i$, unless $G_1=G_2=\ldots=G_{m+1}$ and $G_1\in \{K_{2m+1}\cup (n-2m-1)K_1, K_m\vee (n-m)K_1\}$.

math.CO↗
arXiv2022

Label Adversarial Learning for Skeleton-level to Pixel-level Adjustable Vessel Segmentation

You can have your cake and eat it too. Microvessel segmentation in optical coherence tomography angiography (OCTA) images remains challenging. Skeleton-level segmentation shows clear topology but without diameter information, while pixel-level segmentation shows a clear caliber but low topology. To close this gap, we propose a novel label adversarial learning (LAL) for skeleton-level to pixel-level adjustable vessel segmentation. LAL mainly consists of two designs: a label adversarial loss and an embeddable adjustment layer. The label adversarial loss establishes an adversarial relationship between the two label supervisions, while the adjustment layer adjusts the network parameters to match the different adversarial weights. Such a design can efficiently capture the variation between the two supervisions, making the segmentation continuous and tunable. This continuous process allows us to recommend high-quality vessel segmentation with clear caliber and topology. Experimental results show that our results outperform manual annotations of current public datasets and conventional filtering effects. Furthermore, such a continuous process can also be used to generate an uncertainty map representing weak vessel boundaries and noise.

eess.IV↗
arXiv2022

Towards Sustainable Satellite Edge Computing

Recently, Low Earth Orbit (LEO) satellites experience rapid development and satellite edge computing emerges to address the limitation of bent-pipe architecture in existing satellite systems. Introducing energy-consuming computing components in satellite edge computing increases the depth of battery discharge. This will shorten batteries' life and influences the satellites' operation in orbit. In this paper, we aim to extend batteries' life by minimizing the depth of discharge for Earth observation missions. Facing the challenges of wireless uncertainty and energy harvesting dynamics, our work develops an online energy scheduling algorithm within an online convex optimization framework. Our algorithm achieves sub-linear regret and the constraint violation asymptotically approaches zero. Simulation results show that our algorithm can reduce the depth of discharge significantly.

eess.SY↗
arXiv2022

Learning Latent Graph Dynamics for Visual Manipulation of Deformable Objects

Manipulating deformable objects, such as ropes and clothing, is a long-standing challenge in robotics, because of their large degrees of freedom, complex non-linear dynamics, and self-occlusion in visual perception. The key difficulty is a suitable representation, rich enough to capture the object shape, dynamics for manipulation and yet simple enough to be estimated reliably from visual observations. This work aims to learn latent Graph dynamics for DefOrmable Object Manipulation (G-DOOM). G-DOOM approximates a deformable object as a sparse set of interacting keypoints, which are extracted automatically from images via unsupervised learning. It learns a graph neural network that captures abstractly the geometry and the interaction dynamics of the keypoints. To handle object self-occlusion, G-DOOM uses a recurrent neural network to track the keypoints over time and condition their interactions on the history. We then train the resulting recurrent graph dynamics model through contrastive learning in a high-fidelity simulator. For manipulation planning, G-DOOM reasons explicitly about the learned dynamics model through model-predictive control applied at each keypoint. Preliminary experiments of G-DOOM on a set of challenging rope and cloth manipulation tasks indicate strong performance, compared with state-of-the-art methods. Although trained in a simulator, G-DOOM transfers directly to a real robot for both rope and cloth manipulation.

cs.RO↗
arXiv2022

On Exploring Pose Estimation as an Auxiliary Learning Task for Visible-Infrared Person Re-identification

Visible-infrared person re-identification (VI-ReID) has been challenging due to the existence of large discrepancies between visible and infrared modalities. Most pioneering approaches reduce intra-class variations and inter-modality discrepancies by learning modality-shared and ID-related features. However, an explicit modality-shared cue, i.e., body keypoints, has not been fully exploited in VI-ReID. Additionally, existing feature learning paradigms imposed constraints on either global features or partitioned feature stripes, which neglect the prediction consistency of global and part features. To address the above problems, we exploit Pose Estimation as an auxiliary learning task to assist the VI-ReID task in an end-to-end framework. By jointly training these two tasks in a mutually beneficial manner, our model learns higher quality modality-shared and ID-related features. On top of it, the learnings of global features and local features are seamlessly synchronized by Hierarchical Feature Constraint (HFC), where the former supervises the latter using the knowledge distillation strategy. Experimental results on two benchmark VI-ReID datasets show that the proposed method consistently improves state-of-the-art methods by significant margins. Specifically, our method achieves nearly 20$\%$ mAP improvements against the state-of-the-art method on the RegDB dataset. Our intriguing findings highlight the usage of auxiliary task learning in VI-ReID.

cs.CV↗
arXiv2021

Tiansuan Constellation: An Open Research Platform

Satellite network is the first step of interstellar voyages. It can provide global Internet connectivity everywhere on earth, where most areas cannot access the Internet by the terrestrial infrastructure due to the geographic accessibility and high cost. The space industry experiences a rise in large low-earth-orbit satellite constellations to achieve universal connectivity. The research community is also urgent to do some leading research to bridge the connectivity divide. Researchers now conduct their work by simulation, which is far from enough. However, experiments on real satellites are blocked by the high threshold of space technology, such as deployment cost and unknown risks. To solve the above dilemma, we are eager to contribute to the universal connectivity and build an open research platform, Tiansuan constellation to support experiments on real satellite networks. We discuss the potential research topics that would benefit from Tiansuan constellation. We provide two case studies that have already deployed in two experimental satellites of Tiansuan constellation.

cs.DC↗
arXiv2021

From Cloud to Edge: A First Look at Public Edge Platforms

Public edge platforms have drawn increasing attention from both academia and industry. In this study, we perform a first-of-its-kind measurement study on a leading public edge platform that has been densely deployed in China. Based on this measurement, we quantitatively answer two critical yet unexplored questions. First, from end users' perspective, what is the performance of commodity edge platforms compared to cloud, in terms of the end-to-end network delay, throughput, and the application QoE. Second, from the edge service provider's perspective, how are the edge workloads different from cloud, in terms of their VM subscription, monetary cost, and resource usage. Our study quantitatively reveals the status quo of today's public edge platforms, and provides crucial insights towards developing and operating future edge services.

cs.NI↗
arXiv2021

State-based Episodic Memory for Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning (MARL) algorithms have made promising progress in recent years by leveraging the centralized training and decentralized execution (CTDE) paradigm. However, existing MARL algorithms still suffer from the sample inefficiency problem. In this paper, we propose a simple yet effective approach, called state-based episodic memory (SEM), to improve sample efficiency in MARL. SEM adopts episodic memory (EM) to supervise the centralized training procedure of CTDE in MARL. To the best of our knowledge, SEM is the first work to introduce EM into MARL. We can theoretically prove that, when using for MARL, SEM has lower space complexity and time complexity than state and action based EM (SAEM), which is originally proposed for single-agent reinforcement learning. Experimental results on StarCraft multi-agent challenge (SMAC) show that introducing episodic memory into MARL can improve sample efficiency and SEM can reduce storage cost and time cost compared with SAEM.

cs.LG↗
arXiv2021

Sleep, Sense or Transmit: Energy-Age Tradeoff for Status Update with Two-Thresholds Optimal Policy

Age-of-Information (AoI), or simply age, which measures the data freshness, is essential for real-time Internet-of-Things (IoT) applications. On the other hand, energy saving is urgently required by many energy-constrained IoT devices. This paper studies the energy-age tradeoff for status update from a sensor to a monitor over an error-prone channel. The sensor can sleep, sense and transmit a new update, or retransmit by considering both sensing energy and transmit energy. An infinite-horizon average cost problem is formulated as a Markov decision process (MDP) with the objective of minimizing the weighted sum of average AoI and average energy consumption. By solving the associated discounted cost problem and analyzing the Markov chain under the optimal policy, we prove that there exists a threshold optimal stationary policy with only two thresholds, i.e., one threshold on the AoI at the transmitter (AoIT) and the other on the AoI at the receiver (AoIR). Moreover, the two thresholds can be efficiently found by a line search. Numerical results show the performance of the optimal policies and the tradeoff curves with different parameters. Comparisons with the conventional policies show that considering sensing energy is of significant impact on the policy design, and introducing sleep mode greatly expands the tradeoff range.

cs.IT↗