Searcharxiv⌕ Search

arXiv subjects

Wenxin Li

Publications and source records attributed to Wenxin Li.

54 records · Page 3Linked to original sources
arXiv2022

Evolutionary Game-Theoretical Analysis for General Multiplayer Asymmetric Games

Evolutionary game theory has been a successful tool to combine classical game theory with learning-dynamical descriptions in multiagent systems. Provided some symmetric structures of interacting players, many studies have been focused on using a simplified heuristic payoff table as input to analyse the dynamics of interactions. Nevertheless, even for the state-of-the-art method, there are two limits. First, there is inaccuracy when analysing the simplified payoff table. Second, no existing work is able to deal with 2-population multiplayer asymmetric games. In this paper, we fill the gap between heuristic payoff table and dynamic analysis without any inaccuracy. In addition, we propose a general framework for $m$ versus $n$ 2-population multiplayer asymmetric games. Then, we compare our method with the state-of-the-art in some classic games. Finally, to illustrate our method, we perform empirical game-theoretical analysis on Wolfpack as well as StarCraft II, both of which involve complex multiagent interactions.

cs.AI↗
arXiv2022

Performance Analysis of Modified SRPT in Multiple-Processor Multitask Scheduling

In this paper we study the multiple-processor multitask scheduling problem in both deterministic and stochastic models, where each job have several tasks and is complete only when all its tasks are finished. We consider and analyze Modified Shortest Remaining Processing Time (M-SRPT) scheduling algorithm, a simple modification of SRPT, which always schedules jobs according to SRPT whenever possible, while processes tasks in an arbitrary order. The M-SRPT algorithm is proved to achieve a competitive ratio of $Θ(\log α+β)$ for minimizing response time, where $α$ denotes the ratio between maximum job workload and minimum job workload, $β$ represents the ratio between maximum non-preemptive task workload and minimum job workload. In addition, the competitive ratio achieved is shown to be optimal (up to a constant factor), when there are constant number of machines. We further consider the problem under Poisson arrival and general workload distribution (\ie, M/GI/$N$ system), and show that M-SRPT achieves asymptotic optimal mean response time when the traffic intensity $ρ$ approaches $1$, if job size distribution has finite support. Beyond finite job workload, the asymptotic optimality of M-SRPT also holds for infinite job size distributions with certain probabilistic assumptions, for example, M/M/$N$ system with finite task workload. As a special case, we show that M-SRPT is asymptotic optimal in M/M/$1$ model, in which the task size distribution is allowed to have infinite support.

cs.PF↗
arXiv2022

Submodular Maximization in Clean Linear Time

In this paper, we provide the first deterministic algorithm that achieves the tight $1-1/e$ approximation guarantee for submodular maximization under a cardinality (size) constraint while making a number of queries that scales only linearly with the size of the ground set $n$. To complement our result, we also show strong information-theoretic lower bounds. More specifically, we show that when the maximum cardinality allowed for a solution is constant, no algorithm making a sub-linear number of function evaluations can guarantee any constant approximation ratio. Furthermore, when the constraint allows the selection of a constant fraction of the ground set, we show that any algorithm making fewer than $Ω(n/\log(n))$ function evaluations cannot perform better than an algorithm that simply outputs a uniformly random subset of the ground set of the right size. We then provide a variant of our deterministic algorithm for the more general knapsack constraint, which is the first linear-time algorithm that achieves $1/2$-approximation guarantee for this constraint. Finally, we extend our results to the general case of maximizing a monotone submodular function subject to the intersection of a $p$-set system and multiple knapsack constraints. We extensively evaluate the performance of our algorithms on multiple real-life machine learning applications, including movie recommendation, location summarization, twitter text summarization and video summarization.

cs.DS↗
arXiv2021

FAPR: Fast and Accurate Program Repair for Introductory Programming Courses

In introductory programming courses, it is challenging for instructors to provide debugging feedback on students' incorrect programs. Some recent tools automatically offer program repair feedback by identifying any differences between incorrect and correct programs, but suffer from issues related to scalability, accuracy, and cross-language portability. This paper presents FAPR -- our novel approach that suggests repairs based on program differences in a fast and accurate manner. FAPR is different from current tools in three aspects. First, it encodes syntactic information into token sequences to enable high-speed comparison between incorrect and correct programs. Second, to accurately extract program differences, FAPR adopts a novel matching algorithm that maximizes token-level matches and minimizes statement-level differences. Third, FAPR relies on testing instead of static/dynamic analysis to validate and refine candidate repairs, so it eliminates the language dependency or high runtime overhead incurred by complex program analysis. We implemented FAPR to suggest repairs for both C and C++ programs; our experience shows the great cross-language portability of FAPR. More importantly, we empirically compared FAPR with a state-of-the-art tool Clara. FAPR suggested repairs for over 95.5% of incorrect solutions. We sampled 250 repairs among FAPR's suggestions, and found 89.6% of the samples to be minimal and correct. FAPR outperformed Clara by suggesting repairs for more cases, creating smaller repairs, producing higher-quality fixes, and causing lower runtime overheads. Our results imply that FAPR can potentially help instructors or TAs to effectively locate bugs in incorrect code, and to provide debugging hints/guidelines based on those generated repairs.

cs.SE↗
arXiv2020

Nearly Linear Time Deterministic Algorithms for Submodular Maximization Under Knapsack Constraint and Beyond

In this work, we study the classic submodular maximization problem under knapsack constraints and beyond. We first present an $(7/16-\varepsilon)$-approximate algorithm for single knapsack constraint, which requires $O(n\cdot\max\{\varepsilon^{-1},\log\log n\})$ queries, and two passes in the streaming setting. This provides improvements in approximation ratio, query complexity and number of passes on the stream. We next show that there exists an $(1/2-\varepsilon)$-approximate deterministic algorithm for constant number of binary packing constraints, which achieves a query complexity of $O_{\varepsilon}(n\cdot\log \log n)$. One salient feature of our deterministic algorithm is, both its approximation ratio and time complexity are independent of the number of constraints. Lastly we present nearly linear time algorithms for the intersection of $p$-system and $d$ knapsack constraint, we achieve approximation ratio of $(1/(p+\frac{7}{4}d+1)-\varepsilon)$ for monotone objective and $(p/(p+1)(2p+\frac{7}{4}d+1)-\varepsilon)$ for non-monotone objective.

cs.DS↗
arXiv2020

A Note on Optimizing the Ratio of Monotone Supermodular Functions

We show that for the problem of minimizing (or maximizing) the ratio of two supermodular functions, no bounded approximation ratio can be achieved via polynomial number of queries, if the two supermodular functions are both monotone non-decreasing or non-increasing.

cs.LG↗
arXiv2020

A Faster FPTAS for Knapsack Problem With Cardinality Constraint

We study the $K$-item knapsack problem (i.e., $1.5$-dimensional KP), which is a generalization of the famous 0-1 knapsack problem (i.e., $1$-dimensional KP) in which an upper bound $K$ is imposed on the number of items selected. This problem is of fundamental importance and is known to have a broad range of applications in various fields. It is well known that, there is no FPTAS for the $d$-dimensional knapsack problem when $d\geq 2$, unless P $=$ NP. While the $K$-item knapsack problem is known to admit an FPTAS, the complexity of all existing FPTASs have a high dependency on the cardinality bound $K$ and approximation error $\varepsilon$, which could result in inefficiencies especially when $K$ and $\varepsilon^{-1}$ increase. The current best results are due to Mastrolilli and Hutter (2006), in which two schemes are presented exhibiting a space-time tradeoff--one scheme with time complexity $O(n+Kz^{2}/\varepsilon^{2})$ and space complexity $O(n+z^{3}/\varepsilon)$, while another scheme requires $O(n+(Kz^{2}+z^{4})/\varepsilon^{2})$ run-time but only needs $O(n+z^{2}/\varepsilon)$ space, where $z=\min\{K,1/\varepsilon\}$. In this paper we close the space-time tradeoff exhibited in the state-of-the-art by designing a new FPTAS with a run-time of $\widetilde{O}(n+z^{2}/\varepsilon^{2})$, while simultaneously reaching the $O(n+z^{2}/\varepsilon)$ space bound. Our scheme provides $\widetilde{O}(K)$ and $O(z)$ improvements on the state-of-the-art algorithms in time and space complexity respectively, and is the first scheme that achieves a run-time that is independent of cardinality bound $K$ (up to logarithmic factors) under fixed $\varepsilon$. Another salient feature of our scheme is that it is the first FPTAS that achieves better time and space complexity bounds than the very first standard FPTAS over all parameter regimes.

cs.DS↗
arXiv2020

Viral Pneumonia Screening on Chest X-ray Images Using Confidence-Aware Anomaly Detection

Cluster of viral pneumonia occurrences during a short period of time may be a harbinger of an outbreak or pandemic, like SARS, MERS, and recent COVID-19. Rapid and accurate detection of viral pneumonia using chest X-ray can be significantly useful in large-scale screening and epidemic prevention, particularly when other chest imaging modalities are less available. Viral pneumonia often have diverse causes and exhibit notably different visual appearances on X-ray images. The evolution of viruses and the emergence of novel mutated viruses further result in substantial dataset shift, which greatly limits the performance of classification approaches. In this paper, we formulate the task of differentiating viral pneumonia from non-viral pneumonia and healthy controls into an one-class classification-based anomaly detection problem, and thus propose the confidence-aware anomaly detection (CAAD) model, which consists of a shared feature extractor, an anomaly detection module, and a confidence prediction module. If the anomaly score produced by the anomaly detection module is large enough or the confidence score estimated by the confidence prediction module is small enough, we accept the input as an anomaly case (i.e., viral pneumonia). The major advantage of our approach over binary classification is that we avoid modeling individual viral pneumonia classes explicitly and treat all known viral pneumonia cases as anomalies to reinforce the one-class model. The proposed model outperforms binary classification models on the clinical X-VIRAL dataset that contains 5,977 viral pneumonia (no COVID-19) cases, 18,619 non-viral pneumonia cases, and 18,774 healthy controls.

eess.IV↗
arXiv2019

Learning to Design Games: Strategic Environments in Reinforcement Learning

In typical reinforcement learning (RL), the environment is assumed given and the goal of the learning is to identify an optimal policy for the agent taking actions through its interactions with the environment. In this paper, we extend this setting by considering the environment is not given, but controllable and learnable through its interaction with the agent at the same time. This extension is motivated by environment design scenarios in the real-world, including game design, shopping space design and traffic signal design. Theoretically, we find a dual Markov decision process (MDP) w.r.t. the environment to that w.r.t. the agent, and derive a policy gradient solution to optimizing the parametrized environment. Furthermore, discontinuous environments are addressed by a proposed general generative framework. Our experiments on a Maze game design task show the effectiveness of the proposed algorithms in generating diverse and challenging Mazes against various agent settings.

cs.AI↗
arXiv2019

Artificial Intelligence for Prosthetics - challenge solutions

In the NeurIPS 2018 Artificial Intelligence for Prosthetics challenge, participants were tasked with building a controller for a musculoskeletal model with a goal of matching a given time-varying velocity vector. Top participants were invited to describe their algorithms. In this work, we describe the challenge and present thirteen solutions that used deep reinforcement learning approaches. Many solutions use similar relaxations and heuristics, such as reward shaping, frame skipping, discretization of the action space, symmetry, and policy blending. However, each team implemented different modifications of the known algorithms by, for example, dividing the task into subtasks, learning low-level control, or by incorporating expert knowledge and using imitation learning.

cs.LG↗
arXiv2018

Layout Design for Intelligent Warehouse by Evolution with Fitness Approximation

With the rapid growth of the express industry, intelligent warehouses that employ autonomous robots for carrying parcels have been widely used to handle the vast express volume. For such warehouses, the warehouse layout design plays a key role in improving the transportation efficiency. However, this work is still done by human experts, which is expensive and leads to suboptimal results. In this paper, we aim to automate the warehouse layout designing process. We propose a two-layer evolutionary algorithm to efficiently explore the warehouse layout space, where an auxiliary objective fitness approximation model is introduced to predict the outcome of the designed warehouse layout and a two-layer population structure is proposed to incorporate the approximation model into the ordinary evolution framework. Empirical experiments show that our method can efficiently design effective warehouse layouts that outperform both heuristic-designed and vanilla evolution-designed warehouse layouts.

cs.AI↗
arXiv2018

N2RPP: An Adversarial Network to Rebuild Plantar Pressure for ACLD Patients

Foot is a vital part of human, and lots of valuable information is embedded. Plantar pressure is one of which contains this information and it describes human walking features. It is proved that once one has trouble with lower limb, the distribution of plantar pressure will change to some degree. Plantar pressure can be converted into images according to some simple standards. In this paper, we take full advantage of these plantar pressure images for medical usage. We present N2RPP, a generative adversarial network (GAN) based method to rebuild plantar pressure images of anterior cruciate ligament deficiency (ACLD) patients from low dimension features, which are extracted from an autoencoder. Through the result of experiments, the extracted features are a useful representation to describe and rebuild plantar pressure images. According to N2RPP's results, we find out that there are several noteworthy differences between normal people and patients. This can provide doctors a rough direction of adjusting plantar pressure to a better distribution to reduce patients' sore and pain during the rehabilitation treatment for ACLD.

cs.CV↗
arXiv2018

A Model for Medical Diagnosis Based on Plantar Pressure

The process of determining which disease or condition explains a person's symptoms and signs can be very complicated and may be inaccurate in some cases. The general belief is that diagnosing diseases relies on doctors' keen intuition, rich experience and professional equipment. In this work, we employ ideas from recent advances in plantar pressure research and from the powerful capacity of the convolutional neural network for learning representations. Here, we propose a model using convolutional neural network based on plantar pressure for medical diagnosis. Our model learns a network that maps plantar pressure data to its corresponding medical diagnostic label. We then apply our model to make the medical diagnosis on datasets we collected from cooperative hospital and achieve an accuracy of 98.36%. We demonstrate that the model base on the convolutional neural network is competitive in medical diagnosis.

cs.CV↗
arXiv2018

ICFVR 2017: 3rd International Competition on Finger Vein Recognition

In recent years, finger vein recognition has become an important sub-field in biometrics and been applied to real-world applications. The development of finger vein recognition algorithms heavily depends on large-scale real-world data sets. In order to motivate research on finger vein recognition, we released the largest finger vein data set up to now and hold finger vein recognition competitions based on our data set every year. In 2017, International Competition on Finger Vein Recognition(ICFVR) is held jointly with IJCB 2017. 11 teams registered and 10 of them joined the final evaluation. The winner of this year dramatically improved the EER from 2.64% to 0.483% compared to the winner of last year. In this paper, we introduce the process and results of ICFVR 2017 and give insights on development of state-of-art finger vein recognition algorithms.

cs.CV↗
arXiv2017

Managing Risk of Bidding in Display Advertising

In this paper, we deal with the uncertainty of bidding for display advertising. Similar to the financial market trading, real-time bidding (RTB) based display advertising employs an auction mechanism to automate the impression level media buying; and running a campaign is no different than an investment of acquiring new customers in return for obtaining additional converted sales. Thus, how to optimally bid on an ad impression to drive the profit and return-on-investment becomes essential. However, the large randomness of the user behaviors and the cost uncertainty caused by the auction competition may result in a significant risk from the campaign performance estimation. In this paper, we explicitly model the uncertainty of user click-through rate estimation and auction competition to capture the risk. We borrow an idea from finance and derive the value at risk for each ad display opportunity. Our formulation results in two risk-aware bidding strategies that penalize risky ad impressions and focus more on the ones with higher expected return and lower risk. The empirical study on real-world data demonstrates the effectiveness of our proposed risk-aware bidding strategies: yielding profit gains of 15.4% in offline experiments and up to 17.5% in an online A/B test on a commercial RTB platform over the widely applied bidding strategies.

cs.GT↗
arXiv2016

A Fusion Method Based on Decision Reliability Ratio for Finger Vein Verification

Finger vein verification has developed a lot since its first proposal, but there is still not a perfect algorithm. It is proved that algorithms with the same overall accuracy may have different misclassified patterns. We could make use of this complementation to fuse individual algorithms together for more precise result. According to our observation, algorithm has different confidence on its decisions but it is seldom considered in fusion methods. Our work is first to define decision reliability ratio to quantify this confidence, and then propose the Maximum Decision Reliability Ratio (MDRR) fusion method incorporating Weighted Voting. Experiment conducted on a data set of 1000 fingers and 5 images per finger proves the effectiveness of the method. The classifier obtained by MDRR method gets an accuracy of 99.42% while the maximum accuracy of the original individual classifiers is 97.77%. The experiment results also show the MDRR outperforms the traditional fusion methods as Voting, Weighted Voting, Sum and Weighted Sum.

cs.CV↗