arXiv subjects

Chi Zhang

Publications and source records attributed to Chi Zhang.

At least 649 records · Page 36Linked to original sources

arXiv2022

Learning Practical Communication Strategies in Cooperative Multi-Agent Reinforcement Learning

In Multi-Agent Reinforcement Learning, communication is critical to encourage cooperation among agents. Communication in realistic wireless networks can be highly unreliable due to network conditions varying with agents' mobility, and stochasticity in the transmission process. We propose a framework to learn practical communication strategies by addressing three fundamental questions: (1) When: Agents learn the timing of communication based on not only message importance but also wireless channel conditions. (2) What: Agents augment message contents with wireless network measurements to better select the game and communication actions. (3) How: Agents use a novel neural message encoder to preserve all information from received messages, regardless of the number and order of messages. Simulating standard benchmarks under realistic wireless network settings, we show significant improvements in game performance, convergence speed and communication efficiency compared with state-of-the-art.

cs.AI↗

arXiv2022

CRCNet: Few-shot Segmentation with Cross-Reference and Region-Global Conditional Networks

Few-shot segmentation aims to learn a segmentation model that can be generalized to novel classes with only a few training images. In this paper, we propose a Cross-Reference and Local-Global Conditional Networks (CRCNet) for few-shot segmentation. Unlike previous works that only predict the query image's mask, our proposed model concurrently makes predictions for both the support image and the query image. Our network can better find the co-occurrent objects in the two images with a cross-reference mechanism, thus helping the few-shot segmentation task. To further improve feature comparison, we develop a local-global conditional module to capture both global and local relations. We also develop a mask refinement module to refine the prediction of the foreground regions recurrently. Experiments on the PASCAL VOC 2012, MS COCO, and FSS-1000 datasets show that our network achieves new state-of-the-art performance.

cs.CV↗

arXiv2022

Multi-agent Databases via Independent Learning

Machine learning is rapidly being used in database research to improve the effectiveness of numerous tasks included but not limited to query optimization, workload scheduling, physical design, etc. Currently, the research focus has been on replacing a single database component responsible for one task by its learning-based counterpart. However, query performance is not simply determined by the performance of a single component, but by the cooperation of multiple ones. As such, learning based database components need to collaborate during both training and execution in order to develop policies that meet end performance goals. Thus, the paper attempts to address the question "Is it possible to design a database consisting of various learned components that cooperatively work to improve end-to-end query latency?". To answer this question, we introduce MADB (Multi-Agent DB), a proof-of-concept system that incorporates a learned query scheduler and a learned query optimizer. MADB leverages a cooperative multi-agent reinforcement learning approach that allows the two components to exchange the context of their decisions with each other and collaboratively work towards reducing the query latency. Preliminary results demonstrate that MADB can outperform the non-cooperative integration of learned components.

cs.DB↗

arXiv2022

Buffer Pool Aware Query Scheduling via Deep Reinforcement Learning

In this extended abstract, we propose a new technique for query scheduling with the explicit goal of reducing disk reads and thus implicitly increasing query performance. We introduce SmartQueue, a learned scheduler that leverages overlapping data reads among incoming queries and learns a scheduling strategy that improves cache hits. SmartQueue relies on deep reinforcement learning to produce workload-specific scheduling strategies that focus on long-term performance benefits while being adaptive to previously-unseen data access patterns. We present results from a proof-of-concept prototype, demonstrating that learned schedulers can offer significant performance improvements over hand-crafted scheduling heuristics. Ultimately, we make the case that this is a promising research direction at the intersection of machine learning and databases.

cs.DB↗

arXiv2022

Sobolev Training for Implicit Neural Representations with Approximated Image Derivatives

Recently, Implicit Neural Representations (INRs) parameterized by neural networks have emerged as a powerful and promising tool to represent different kinds of signals due to its continuous, differentiable properties, showing superiorities to classical discretized representations. However, the training of neural networks for INRs only utilizes input-output pairs, and the derivatives of the target output with respect to the input, which can be accessed in some cases, are usually ignored. In this paper, we propose a training paradigm for INRs whose target output is image pixels, to encode image derivatives in addition to image values in the neural network. Specifically, we use finite differences to approximate image derivatives. We show how the training paradigm can be leveraged to solve typical INRs problems, i.e., image regression and inverse rendering, and demonstrate this training paradigm can improve the data-efficiency and generalization capabilities of INRs. The code of our method is available at \url{https://github.com/megvii-research/Sobolev_INRs}.

cs.CV↗

arXiv2022

Learning Algebraic Representation for Systematic Generalization in Abstract Reasoning

Is intelligence realized by connectionist or classicist? While connectionist approaches have achieved superhuman performance, there has been growing evidence that such task-specific superiority is particularly fragile in systematic generalization. This observation lies in the central debate between connectionist and classicist, wherein the latter continually advocates an algebraic treatment in cognitive architectures. In this work, we follow the classicist's call and propose a hybrid approach to improve systematic generalization in reasoning. Specifically, we showcase a prototype with algebraic representation for the abstract spatial-temporal reasoning task of Raven's Progressive Matrices (RPM) and present the ALgebra-Aware Neuro-Semi-Symbolic (ALANS) learner. The ALANS learner is motivated by abstract algebra and the representation theory. It consists of a neural visual perception frontend and an algebraic abstract reasoning backend: the frontend summarizes the visual information from object-based representation, while the backend transforms it into an algebraic structure and induces the hidden operator on the fly. The induced operator is later executed to predict the answer's representation, and the choice most similar to the prediction is selected as the solution. Extensive experiments show that by incorporating an algebraic treatment, the ALANS learner outperforms various pure connectionist models in domains requiring systematic generalization. We further show the generative nature of the learned algebraic representation; it can be decoded by isomorphism to generate an answer.

cs.AI↗

arXiv2022

A Structured Method for Compilation of QAOA Circuits in Quantum Computing

Quantum Approximation Optimization Algorithm (QAOA) is a highly advocated variational algorithm for solving the combinatorial optimization problem. One critical feature in the quantum circuit of QAOA algorithm is that it consists of two-qubit operators that commute. The flexibility in reordering the two-qubit gates allows compiler optimizations to generate circuits with better depths, gate count, and fidelity. However, it also imposes significant challenges due to additional freedom exposed in the compilation. Prior studies lack the following: (1) Performance guarantee, (2) Scalability, and (3) Awareness of regularity in scalable hardware. We propose a structured method that ensures linear depth for any compiled QAOA circuit on multi-dimensional quantum architectures. We also demonstrate how our method runs on Google Sycamore and IBM Non-linear architectures in a scalable manner and in linear time. Overall, we can compile a circuit with up to 1024 qubits in 10 seconds with a 3.8X speedup in depth, 17% reduction in gate count, and 18X improvement for circuit ESP.

quant-ph↗

arXiv2022

Few-shot Open-set Recognition Using Background as Unknowns

Few-shot open-set recognition aims to classify both seen and novel images given only limited training data of seen classes. The challenge of this task is that the model is required not only to learn a discriminative classifier to classify the pre-defined classes with few training data but also to reject inputs from unseen classes that never appear at training time. In this paper, we propose to solve the problem from two novel aspects. First, instead of learning the decision boundaries between seen classes, as is done in standard close-set classification, we reserve space for unseen classes, such that images located in these areas are recognized as the unseen classes. Second, to effectively learn such decision boundaries, we propose to utilize the background features from seen classes. As these background regions do not significantly contribute to the decision of close-set classification, it is natural to use them as the pseudo unseen classes for classifier learning. Our extensive experiments show that our proposed method not only outperforms multiple baselines but also sets new state-of-the-art results on three popular benchmarks, namely tieredImageNet, miniImageNet, and Caltech-USCD Birds-200-2011 (CUB).

cs.CV↗

arXiv2022

Level-set based pre-processing algorithm for particle-based methods

Obtaining high quality particle distribution representing clean geometry in pre-processing is essential for the simulation accuracy of the particle-based methods. In this paper, several level-set based techniques for cleaning up `dirty' geometries automatically and generating homogeneous particle distributions are presented. First, a non-resolved structure identifying method based on level-set field is employed to detect the tiny fragments which make the geometry `dirty' under a given resolutions. Second, a re-distance algorithm is proposed to remove the tiny fragments and reconstruct clean and smooth geometries. Third, a `static confinement' boundary condition is developed in the particle relaxation process. By complementing the kernel support for the particles near the geometric surface, the boundary condition achieves better body-fitted particle distribution on the narrow region with high curvature. Several numerical examples include a 2D airfoil 30P30N, 3D SPHinXsys symbol, a skyscraper with a flagpole and an inferior vena cava demonstrate that the present method not only cleans up the `dirty' geometries efficiently, but also provides better body-fitted homogeneous particle distribution for complex geometry.

cs.CE↗

arXiv2022

TENET: Transformer Encoding Network for Effective Temporal Flow on Motion Prediction

This technical report presents an effective method for motion prediction in autonomous driving. We develop a Transformer-based method for input encoding and trajectory prediction. Besides, we propose the Temporal Flow Header to enhance the trajectory encoding. In the end, an efficient K-means ensemble method is used. Using our Transformer network and ensemble method, we win the first place of Argoverse 2 Motion Forecasting Challenge with the state-of-the-art brier-minFDE score of 1.90.

cs.CV↗

arXiv2022

Few-shot Segmentation with Optimal Transport Matching and Message Flow

We tackle the challenging task of few-shot segmentation in this work. It is essential for few-shot semantic segmentation to fully utilize the support information. Previous methods typically adopt masked average pooling over the support feature to extract the support clues as a global vector, usually dominated by the salient part and lost certain essential clues. In this work, we argue that every support pixel's information is desired to be transferred to all query pixels and propose a Correspondence Matching Network (CMNet) with an Optimal Transport Matching module to mine out the correspondence between the query and support images. Besides, it is critical to fully utilize both local and global information from the annotated support images. To this end, we propose a Message Flow module to propagate the message along the inner-flow inside the same image and cross-flow between support and query images, which greatly helps enhance the local feature representations. Experiments on PASCAL VOC 2012, MS COCO, and FSS-1000 datasets show that our network achieves new state-of-the-art few-shot segmentation performance.

cs.CV↗

arXiv2022

Origin of Nonlinear Damping due to Mode Coupling in Auto-Oscillatory Modes Strongly Driven by Spin-Orbit Torque

We investigate the physical origin of nonlinear damping due to mode coupling between several auto-oscillatory modes driven by spin-orbit torque in constricted Py/Pt heterostructures by examining the dependence of auto-oscillation on temperature and applied field orientation. We observe a transition in the nonlinear damping of the auto-oscillation modes extracted from the total oscillation power as a function of drive current, which coincides with the onset of power redistribution amongst several modes and the crossover from linewidth narrowing to linewidth broadening in all individual modes. This indicates the activation of another relaxation process by nonlinear magnon-magnon scattering within the modes. We also find that both nonlinear damping and threshold current in the mode-interaction damping regime at high drive current after transition are temperature independent, suggesting that the mode coupling occurs dominantly through a non-thermal magnon scattering process via a dipole or exchange interaction rather than thermally excited magnon-mediated scattering. This finding presents a promising pathway to overcome the current limitations of efficiently controlling the interaction between two highly nonlinear magnetic oscillators to prevent mode crosstalk or inter-mode energy transfer and deepens understanding of complex nonlinear spin dynamics in multimode spin wave systems.

cond-mat.mes-hall↗

arXiv2022

Automatic Generation of Product-Image Sequence in E-commerce

Product images are essential for providing desirable user experience in an e-commerce platform. For a platform with billions of products, it is extremely time-costly and labor-expensive to manually pick and organize qualified images. Furthermore, there are the numerous and complicated image rules that a product image needs to comply in order to be generated/selected. To address these challenges, in this paper, we present a new learning framework in order to achieve Automatic Generation of Product-Image Sequence (AGPIS) in e-commerce. To this end, we propose a Multi-modality Unified Image-sequence Classifier (MUIsC), which is able to simultaneously detect all categories of rule violations through learning. MUIsC leverages textual review feedback as the additional training target and utilizes product textual description to provide extra semantic information. Based on offline evaluations, we show that the proposed MUIsC significantly outperforms various baselines. Besides MUIsC, we also integrate some other important modules in the proposed framework, such as primary image selection, noncompliant content detection, and image deduplication. With all these modules, our framework works effectively and efficiently in JD.com recommendation platform. By Dec 2021, our AGPIS framework has generated high-standard images for about 1.5 million products and achieves 13.6% in reject rate.

cs.CV↗

arXiv2022

A Lagrangian free-stream boundary condition for weakly compressible smoothed particle hydrodynamics

This paper proposes the first free-stream boundary condition in a purely Lagrangian framework for weakly-compressible smoothed particle hydrodynamics (WCSPH). The boundary condition is implemented based on several numerical techniques, including a spatio-temporal method and an ensuring technique for surface particle identification, far-field corrections of density and velocity near the free-stream boundary, and improved in- and outlet conditions which guarantee smooth inflow profile and free outflow particles without the issue of the truncated kernel. The accuracy and effectiveness of the method have been validated by the reference solutions of flow past a flat plate and circular cylinder, respectively. The method has also been tested with more complex flows to demonstrate its applicability, such as fluid-structure interaction (FSI) and 3D flow past a sphere.

physics.flu-dyn↗

arXiv2022

DETR++: Taming Your Multi-Scale Detection Transformer

Convolutional Neural Networks (CNN) have dominated the field of detection ever since the success of AlexNet in ImageNet classification [12]. With the sweeping reform of Transformers [27] in natural language processing, Carion et al. [2] introduce the Transformer-based detection method, i.e., DETR. However, due to the quadratic complexity in the self-attention mechanism in the Transformer, DETR is never able to incorporate multi-scale features as performed in existing CNN-based detectors, leading to inferior results in small object detection. To mitigate this issue and further improve performance of DETR, in this work, we investigate different methods to incorporate multi-scale features and find that a Bi-directional Feature Pyramid (BiFPN) works best with DETR in further raising the detection precision. With this discovery, we propose DETR++, a new architecture that improves detection results by 1.9% AP on MS COCO 2017, 11.5% AP on RICO icon detection, and 9.1% AP on RICO layout extraction over existing baselines.

cs.CV↗

arXiv2022

Quantum computation in a hybrid array of molecules and Rydberg atoms

We show that an array of polar molecules interacting with Rydberg atoms is a promising hybrid system for scalable quantum computation. Quantum information is stored in long-lived hyperfine or rotational states of molecules which interact indirectly through resonant dipole-dipole interactions with Rydberg atoms. A two-qubit gate based on this interaction has a duration of 1 $μ$s and an achievable fidelity of 99.9%. The gate has little sensitivity to the motional states of the particles -- the molecules can be in thermal states, the atoms do not need to be trapped during Rydberg excitation, the gate does not heat the molecules, and heating of the atoms has a negligible effect. Within a large, static array, the gate can be applied to arbitrary pairs of molecules separated by tens of micrometres, making the scheme highly scalable. The molecule-atom interaction can also be used for rapid qubit initialization and efficient, non-destructive qubit readout, without driving any molecular transitions. Single qubit gates are driven using microwave pulses alone, exploiting the strong electric dipole transitions between rotational states. Thus, all operations required for large scale quantum computation can be done without moving the molecules or exciting them out of their ground electronic states.

quant-ph↗

arXiv2022

Constrained Reinforcement Learning for Short Video Recommendation

The wide popularity of short videos on social media poses new opportunities and challenges to optimize recommender systems on the video-sharing platforms. Users provide complex and multi-faceted responses towards recommendations, including watch time and various types of interactions with videos. As a result, established recommendation algorithms that concern a single objective are not adequate to meet this new demand of optimizing comprehensive user experiences. In this paper, we formulate the problem of short video recommendation as a constrained Markov Decision Process (MDP), where platforms want to optimize the main goal of user watch time in long term, with the constraint of accommodating the auxiliary responses of user interactions such as sharing/downloading videos. To solve the constrained MDP, we propose a two-stage reinforcement learning approach based on actor-critic framework. At stage one, we learn individual policies to optimize each auxiliary response. At stage two, we learn a policy to (i) optimize the main response and (ii) stay close to policies learned at the first stage, which effectively guarantees the performance of this main policy on the auxiliaries. Through extensive simulations, we demonstrate effectiveness of our approach over alternatives in both optimizing the main goal as well as balancing the others. We further show the advantage of our approach in live experiments of short video recommendations, where it significantly outperforms other baselines in terms of watch time and interactions from video views. Our approach has been fully launched in the production system to optimize user experiences on the platform.

cs.LG↗

arXiv2022

Energy Efficient Federated Learning over Heterogeneous Mobile Devices via Joint Design of Weight Quantization and Wireless Transmission

Federated learning (FL) is a popular collaborative distributed machine learning paradigm across mobile devices. However, practical FL over resource constrained mobile devices confronts multiple challenges, e.g., the local on-device training and model updates in FL are power hungry and radio resource intensive for mobile devices. To address these challenges, in this paper, we attempt to take FL into the design of future wireless networks and develop a novel joint design of wireless transmission and weight quantization for energy efficient FL over mobile devices. Specifically, we develop flexible weight quantization schemes to facilitate on-device local training over heterogeneous mobile devices. Based on the observation that the energy consumption of local computing is comparable to that of model updates, we formulate the energy efficient FL problem into a mixed-integer programming problem where the quantization and spectrum resource allocation strategies are jointly determined for heterogeneous mobile devices to minimize the overall FL energy consumption (computation + transmissions) while guaranteeing model performance and training latency. Since the optimization variables of the problem are strongly coupled, an efficient iterative algorithm is proposed, where the bandwidth allocation and weight quantization levels are derived. Extensive simulations are conducted to verify the effectiveness of the proposed scheme.

cs.NI↗