arXiv subjects

Feng Wu

Publications and source records attributed to Feng Wu.

At least 181 records · Page 10Linked to original sources

arXiv2022

Automatic Reward Design via Learning Motivation-Consistent Intrinsic Rewards

Reward design is a critical part of the application of reinforcement learning, the performance of which strongly depends on how well the reward signal frames the goal of the designer and how well the signal assesses progress in reaching that goal. In many cases, the extrinsic rewards provided by the environment (e.g., win or loss of a game) are very sparse and make it difficult to train agents directly. Researchers usually assist the learning of agents by adding some auxiliary rewards in practice. However, designing auxiliary rewards is often turned to a trial-and-error search for reward settings that produces acceptable results. In this paper, we propose to automatically generate goal-consistent intrinsic rewards for the agent to learn, by maximizing which the expected accumulative extrinsic rewards can be maximized. To this end, we introduce the concept of motivation which captures the underlying goal of maximizing certain rewards and propose the motivation based reward design method. The basic idea is to shape the intrinsic rewards by minimizing the distance between the intrinsic and extrinsic motivations. We conduct extensive experiments and show that our method performs better than the state-of-the-art methods in handling problems of delayed reward, exploration, and credit assignment.

cs.LG↗

arXiv2022

Meta Reinforcement Learning with Successor Feature Based Context

Most reinforcement learning (RL) methods only focus on learning a single task from scratch and are not able to use prior knowledge to learn other tasks more effectively. Context-based meta RL techniques are recently proposed as a possible solution to tackle this. However, they are usually less efficient than conventional RL and may require many trial-and-errors during training. To address this, we propose a novel meta-RL approach that achieves competitive performance comparing to existing meta-RL algorithms, while requires significantly fewer environmental interactions. By combining context variables with the idea of decomposing reward in successor feature framework, our method does not only learn high-quality policies for multiple tasks simultaneously but also can quickly adapt to new tasks with a small amount of training. Compared with state-of-the-art meta-RL baselines, we empirically show the effectiveness and data efficiency of our method on several continuous control tasks.

cs.LG↗

arXiv2022

Towards Hybrid-Optimization Video Coding

Video coding is a mathematical optimization problem of rate and distortion essentially. To solve this complex optimization problem, two popular video coding frameworks have been developed: block-based hybrid video coding and end-to-end learned video coding. If we rethink video coding from the perspective of optimization, we find that the existing two frameworks represent two directions of optimization solutions. Block-based hybrid coding represents the discrete optimization solution because those irrelevant coding modes are discrete in mathematics. It searches for the best one among multiple starting points (i.e. modes). However, the search is not efficient enough. On the other hand, end-to-end learned coding represents the continuous optimization solution because the gradient descent is based on a continuous function. It optimizes a group of model parameters efficiently by the numerical algorithm. However, limited by only one starting point, it is easy to fall into the local optimum. To better solve the optimization problem, we propose to regard video coding as a hybrid of the discrete and continuous optimization problem, and use both search and numerical algorithm to solve it. Our idea is to provide multiple discrete starting points in the global space and optimize the local optimum around each point by numerical algorithm efficiently. Finally, we search for the global optimum among those local optimums. Guided by the hybrid optimization idea, we design a hybrid optimization video coding framework, which is built on continuous deep networks entirely and also contains some discrete modes. We conduct a comprehensive set of experiments. Compared to the continuous optimization framework, our method outperforms pure learned video coding methods. Meanwhile, compared to the discrete optimization framework, our method achieves comparable performance to HEVC reference software HM16.10 in PSNR.

eess.IV↗

arXiv2022

Learning Task-relevant Representations for Generalization via Characteristic Functions of Reward Sequence Distributions

Generalization across different environments with the same tasks is critical for successful applications of visual reinforcement learning (RL) in real scenarios. However, visual distractions -- which are common in real scenes -- from high-dimensional observations can be hurtful to the learned representations in visual RL, thus degrading the performance of generalization. To tackle this problem, we propose a novel approach, namely Characteristic Reward Sequence Prediction (CRESP), to extract the task-relevant information by learning reward sequence distributions (RSDs), as the reward signals are task-relevant in RL and invariant to visual distractions. Specifically, to effectively capture the task-relevant information via RSDs, CRESP introduces an auxiliary task -- that is, predicting the characteristic functions of RSDs -- to learn task-relevant representations, because we can well approximate the high-dimensional distributions by leveraging the corresponding characteristic functions. Experiments demonstrate that CRESP significantly improves the performance of generalization on unseen environments, outperforming several state-of-the-arts on DeepMind Control tasks with different visual distractions.

cs.LG↗

arXiv2022

Quantum Instruction Set Design for Performance

A quantum instruction set is where quantum hardware and software meet. We develop new characterization and compilation techniques for non-Clifford gates to accurately evaluate different quantum instruction set designs. We specifically apply them to our fluxonium processor that supports mainstream instruction $\mathrm{iSWAP}$ by calibrating and characterizing its square root $\mathrm{SQiSW}$. We measure a gate fidelity of up to $99.72\%$ with an average of $99.31\%$ and realize Haar random two-qubit gates using $\mathrm{SQiSW}$ with an average fidelity of $96.38\%$. This is an average error reduction of $41\%$ for the former and a $50\%$ reduction for the latter compared to using $\mathrm{iSWAP}$ on the same processor. This shows designing the quantum instruction set consisting of $\mathrm{SQiSW}$ and single-qubit gates on such platforms leads to a performance boost at almost no cost.

quant-ph↗

arXiv2022

A new stable and avoiding inversion iteration for computing matrix square root

The objective of this research was to compute the principal matrix square root with sparse approximation. A new stable iterative scheme avoiding fully matrix inversion (SIAI) is provided. The analysis on the sparsity and error of the matrices involved during the iterative process is given. Based on the bandwidth and error analysis, a more efficient algorithm combining the SIAI with the filtering technique is proposed. The high computational efficiency and accuracy of the proposed method are demonstrated by computing the principal square roots of different matrices to reveal its applicability over the existing methods.

math.NA↗

arXiv2022

Exploiting Global Semantic Similarities in Knowledge Graphs by Relational Prototype Entities

Knowledge graph (KG) embedding aims at learning the latent representations for entities and relations of a KG in continuous vector spaces. An empirical observation is that the head (tail) entities connected by the same relation often share similar semantic attributes -- specifically, they often belong to the same category -- no matter how far away they are from each other in the KG; that is, they share global semantic similarities. However, many existing methods derive KG embeddings based on the local information, which fail to effectively capture such global semantic similarities among entities. To address this challenge, we propose a novel approach, which introduces a set of virtual nodes called \textit{\textbf{relational prototype entities}} to represent the prototypes of the head and tail entities connected by the same relations. By enforcing the entities' embeddings close to their associated prototypes' embeddings, our approach can effectively encourage the global semantic similarities of entities -- that can be far away in the KG -- connected by the same relation. Experiments on the entity alignment and KG completion tasks demonstrate that our approach significantly outperforms recent state-of-the-arts.

cs.CL↗

arXiv2022

Self-Adaptive Label Augmentation for Semi-supervised Few-shot Classification

Few-shot classification aims to learn a model that can generalize well to new tasks when only a few labeled samples are available. To make use of unlabeled data that are more abundantly available in real applications, Ren et al. \shortcite{ren2018meta} propose a semi-supervised few-shot classification method that assigns an appropriate label to each unlabeled sample by a manually defined metric. However, the manually defined metric fails to capture the intrinsic property in data. In this paper, we propose a \textbf{S}elf-\textbf{A}daptive \textbf{L}abel \textbf{A}ugmentation approach, called \textbf{SALA}, for semi-supervised few-shot classification. A major novelty of SALA is the task-adaptive metric, which can learn the metric adaptively for different tasks in an end-to-end fashion. Another appealing feature of SALA is a progressive neighbor selection strategy, which selects unlabeled data with high confidence progressively through the training phase. Experiments demonstrate that SALA outperforms several state-of-the-art methods for semi-supervised few-shot classification on benchmark datasets.

cs.CV↗

arXiv2022

Feudal Multi-Agent Reinforcement Learning with Adaptive Network Partition for Traffic Signal Control

Multi-agent reinforcement learning (MARL) has been applied and shown great potential in multi-intersections traffic signal control, where multiple agents, one for each intersection, must cooperate together to optimize traffic flow. To encourage global cooperation, previous work partitions the traffic network into several regions and learns policies for agents in a feudal structure. However, static network partition fails to adapt to dynamic traffic flow, which will changes frequently over time. To address this, we propose a novel feudal MARL approach with adaptive network partition. Specifically, we first partition the network into several regions according to the traffic flow. To do this, we propose two approaches: one is directly to use graph neural network (GNN) to generate the network partition, and the other is to use Monte-Carlo tree search (MCTS) to find the best partition with criteria computed by GNN. Then, we design a variant of Qmix using GNN to handle various dimensions of input, given by the dynamic network partition. Finally, we use a feudal hierarchy to manage agents in each partition and promote global cooperation. By doing so, agents are able to adapt to the traffic flow as required in practice. We empirically evaluate our method both in a synthetic traffic grid and real-world traffic networks of three cities, widely used in the literature. Our experimental results confirm that our method can achieve better performance, in terms of average travel time and queue length, than several leading methods for traffic signal control.

cs.MA↗

arXiv2022

Compressing Deep Graph Neural Networks via Adversarial Knowledge Distillation

Deep graph neural networks (GNNs) have been shown to be expressive for modeling graph-structured data. Nevertheless, the over-stacked architecture of deep graph models makes it difficult to deploy and rapidly test on mobile or embedded systems. To compress over-stacked GNNs, knowledge distillation via a teacher-student architecture turns out to be an effective technique, where the key step is to measure the discrepancy between teacher and student networks with predefined distance functions. However, using the same distance for graphs of various structures may be unfit, and the optimal distance formulation is hard to determine. To tackle these problems, we propose a novel Adversarial Knowledge Distillation framework for graph models named GraphAKD, which adversarially trains a discriminator and a generator to adaptively detect and decrease the discrepancy. Specifically, noticing that the well-captured inter-node and inter-class correlations favor the success of deep GNNs, we propose to criticize the inherited knowledge from node-level and class-level views with a trainable discriminator. The discriminator distinguishes between teacher knowledge and what the student inherits, while the student GNN works as a generator and aims to fool the discriminator. To our best knowledge, GraphAKD is the first to introduce adversarial training to knowledge distillation in graph domains. Experiments on node-level and graph-level classification benchmarks demonstrate that GraphAKD improves the student performance by a large margin. The results imply that GraphAKD can precisely transfer knowledge from a complicated teacher GNN to a compact student GNN.

cs.LG↗

arXiv2022

Evaluating Pest Management Strategies: A Robust Method and its Application to Strawberry Disease Management

Farmers use pesticides to reduce yield losses. The efficacies of pesticide treatments are often evaluated by analyzing the average treatment effects and risks. The stochastic efficiency with respect to a function is often employed in such evaluations through ranking the certainty equivalents of each treatment. The main challenge of using this method is gathering an adequate number of observations to produce results with statistical power. However, in many cases, only a limited number of trials are replicated in field experiments, leaving an inadequate number of observations. In addition, this method focuses only on the farmer's profit without incorporating the impact of disease pressure on yield and profit. The objective of our study is to propose a methodology to address the issue of an insufficient number of observations using simulations and take into account the effect of disease pressure on yield through a quantile regression model. We apply this method to the case of strawberry disease management in Florida.

econ.GN↗

arXiv2022

MFGNet: Dynamic Modality-Aware Filter Generation for RGB-T Tracking

Many RGB-T trackers attempt to attain robust feature representation by utilizing an adaptive weighting scheme (or attention mechanism). Different from these works, we propose a new dynamic modality-aware filter generation module (named MFGNet) to boost the message communication between visible and thermal data by adaptively adjusting the convolutional kernels for various input images in practical tracking. Given the image pairs as input, we first encode their features with the backbone network. Then, we concatenate these feature maps and generate dynamic modality-aware filters with two independent networks. The visible and thermal filters will be used to conduct a dynamic convolutional operation on their corresponding input feature maps respectively. Inspired by residual connection, both the generated visible and thermal feature maps will be summarized with input feature maps. The augmented feature maps will be fed into the RoI align module to generate instance-level features for subsequent classification. To address issues caused by heavy occlusion, fast motion and out-of-view, we propose to conduct a joint local and global search by exploiting a new direction-aware target driven attention mechanism. The spatial and temporal recurrent neural network is used to capture the direction-aware context for accurate global attention prediction. Extensive experiments on three large-scale RGB-T tracking benchmark datasets validated the effectiveness of our proposed algorithm. The source code of this paper is available at \textcolor{magenta}{\url{https://github.com/wangxiao5791509/MFG_RGBT_Tracking_PyTorch}}.

cs.CV↗

arXiv2022

Titanium Nitride Film on Sapphire Substrate with Low Dielectric Loss for Superconducting Qubits

Dielectric loss is one of the major decoherence sources of superconducting qubits. Contemporary high-coherence superconducting qubits are formed by material systems mostly consisting of superconducting films on substrate with low dielectric loss, where the loss mainly originates from the surfaces and interfaces. Among the multiple candidates for material systems, a combination of titanium nitride (TiN) film and sapphire substrate has good potential because of its chemical stability against oxidization, and high quality at interfaces. In this work, we report a TiN film deposited onto sapphire substrate achieving low dielectric loss at the material interface. Through the systematic characterizations of a series of transmon qubits fabricated with identical batches of TiN base layers, but different geometries of qubit shunting capacitors with various participation ratios of the material interface, we quantitatively extract the loss tangent value at the substrate-metal interface smaller than $8.9 \times 10^{-4}$ in 1-nm disordered layer. By optimizing the interface participation ratio of the transmon qubit, we reproducibly achieve qubit lifetimes of up to 300 $μ$s and quality factors approaching 8 million. We demonstrate that TiN film on sapphire substrate is an ideal material system for high-coherence superconducting qubits. Our analyses further suggest that the interface dielectric loss around the Josephson junction part of the circuit could be the dominant limitation of lifetimes for state-of-the-art transmon qubits.

quant-ph↗

arXiv2022

MNL-Bandits under Inventory and Limited Switches Constraints

Optimizing the assortment of products to display to customers is a key to increasing revenue for both offline and online retailers. To trade-off between exploring customers' preference and exploiting customers' choices learned from data, in this paper, by adopting the Multi-Nomial Logit (MNL) choice model to capture customers' choices over products, we study the problem of optimizing assortments over a planning horizon $T$ for maximizing the profit of the retailer. To make the problem setting more practical, we consider both the inventory constraint and the limited switches constraint, where the retailer cannot use up the resource inventory before time $T$ and is forbidden to switch the assortment shown to customers too many times. Such a setting suits the case when an online retailer wants to dynamically optimize the assortment selection for a population of customers. We develop an efficient UCB-like algorithm to optimize the assortments while learning customers' choices from data. We prove that our algorithm can achieve a sub-linear regret bound $\tilde{O}\left(T^{1-α/2}\right)$ if $O(T^α)$ switches are allowed. %, and our regret bound is optimal with respect to $T$. Extensive numerical experiments show that our algorithm outperforms baselines and the gap between our algorithm's performance and the theoretical upper bound is small.

cs.LG↗

arXiv2022

Duality-Induced Regularizer for Semantic Matching Knowledge Graph Embeddings

Semantic matching models -- which assume that entities with similar semantics have similar embeddings -- have shown great power in knowledge graph embeddings (KGE). Many existing semantic matching models use inner products in embedding spaces to measure the plausibility of triples and quadruples in static and temporal knowledge graphs. However, vectors that have the same inner products with another vector can still be orthogonal to each other, which implies that entities with similar semantics may have dissimilar embeddings. This property of inner products significantly limits the performance of semantic matching models. To address this challenge, we propose a novel regularizer -- namely, DUality-induced RegulArizer (DURA) -- which effectively encourages the entities with similar semantics to have similar embeddings. The major novelty of DURA is based on the observation that, for an existing semantic matching KGE model (primal), there is often another distance based KGE model (dual) closely associated with it, which can be used as effective constraints for entity embeddings. Experiments demonstrate that DURA consistently and significantly improves the performance of state-of-the-art semantic matching models on both static and temporal knowledge graph benchmarks.

cs.CL↗

arXiv2022

ProgressiveMotionSeg: Mutually Reinforced Framework for Event-Based Motion Segmentation

Dynamic Vision Sensor (DVS) can asynchronously output the events reflecting apparent motion of objects with microsecond resolution, and shows great application potential in monitoring and other fields. However, the output event stream of existing DVS inevitably contains background activity noise (BA noise) due to dark current and junction leakage current, which will affect the temporal correlation of objects, resulting in deteriorated motion estimation performance. Particularly, the existing filter-based denoising methods cannot be directly applied to suppress the noise in event stream, since there is no spatial correlation. To address this issue, this paper presents a novel progressive framework, in which a Motion Estimation (ME) module and an Event Denoising (ED) module are jointly optimized in a mutually reinforced manner. Specifically, based on the maximum sharpness criterion, ME module divides the input event into several segments by adaptive clustering in a motion compensating warp field, and captures the temporal correlation of event stream according to the clustered motion parameters. Taking temporal correlation as guidance, ED module calculates the confidence that each event belongs to real activity events, and transmits it to ME module to update energy function of motion segmentation for noise suppression. The two steps are iteratively updated until stable motion segmentation results are obtained. Extensive experimental results on both synthetic and real datasets demonstrate the superiority of our proposed approaches against the State-Of-The-Art (SOTA) methods.

cs.CV↗

arXiv2022

Development and Commissioning of a Compact Cosmic Ray Muon Imaging Prototype

Due to the muon tomography's capability of imaging high Z materials, some potential applications have been reported on inspecting smuggled nuclear materials in customs. A compact Cosmic Ray Muons (CRM) imaging prototype, Lanzhou University Muon Imaging System (LUMIS), is comprehensively introduced in this paper including the structure design, assembly, data acquisition and analysis, detector performance test, and material imaging commissioning etc. Casted triangular prism plastic scintillators (PS) were coupled with Si-PMs for sensitive detector components in system. LUMIS's experimental results show that the detection efficiency of an individual detector layer is about 98%, the position resolution for vertical incident muons is 2.5 mm and the angle resolution is 8.73 mrad given a separation distance of 40.5 cm. Moreover, the image reconstruction software was developed based on the Point of Closest Approach (PoCA) to detect lead bricks as our target. The reconstructed images indicate that the profile of the lead bricks in the image is highly consistent with the target. Subsequently, the capability of LUMIS to distinguish different materials, such as Pb, Cu, Fe, and Al, was investigated as well. The lower limit of response time for rapidly alarming high-Z materials is also given and discussed. The successful development and commissioning of the LUMIS prototype have provided a new solution option in technology and craftsmanship for developing compact CRM imaging systems that can be used in many applications.

physics.ins-det↗

arXiv2022

Polarization-controlled dynamically switchable high-harmonic generation from all-dielectric metasurfaces governed by dual bound states in the continuum

Tailoring optical nonlinear effects (e.g. harmonic generation, sum-frequency mixing, etc.) in the recently emerging all-dielectric platform is important for both the fundamental science and industrial development of high-efficiency, ultrafast, and miniaturized photonic devices. In this work, we propose a novel paradigm for dynamically switchable high-harmonic generation in Silicon nanodimer metasurfaces by exploiting polarization-controlled dual bound states in the continuum (BIC). Owing to the high-quality factor of BIC resonances, efficient harmonic signals including the third-harmonic generation and fifth-harmonic generation from a direct process as well as a cascaded process by degenerate four-wave mixing are obtained. Moreover, the BICs and their resonantly enhanced harmonics can be switched on or off with high selectivity respect to the fundamental pump polarization. Compared with previous reports, our work provide a simple but effective tuning strategy by fully exploring the structural symmetry and polarization degree of freedom rather than resorting to additional external stimuli, which would have great advantages in smart designing tunable and switchable nonlinear light source for chip-scale applications.

physics.optics↗