Artificial Intelligence and Robotics Commons^™

Generative Flows With Invertible Attentions, Rhea Sanjay Sukthanker, Zhiwu Huang, Suryansh Kumar, Radu Timofte, Luc Van Gool

Research Collection School Of Computing and Information Systems

Flow-based generative models have shown an excellent ability to explicitly learn the probability density function of data via a sequence of invertible transformations. Yet, learning attentions in generative flows remains understudied, while it has made breakthroughs in other domains. To fill the gap, this paper introduces two types of invertible attention mechanisms, i.e., map-based and transformer-based attentions, for both unconditional and conditional generative flows. The key idea is to exploit a masked scheme of these two attentions to learn long-range data dependencies in the context of generative flows. The masked scheme allows for invertible attention modules with tractable Jacobian determinants, …

Officers: Operational Framework For Intelligent Crime-And-Emergency Response Scheduling, Jonathan David Chase, Siong Thye Goh, Tran Phong, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

In the quest to achieve better response times in dense urban environments, law enforcement agencies are seeking AI-driven planning systems to inform their patrol strategies. In this paper, we present a framework, OFFICERS, for deployment planning that learns from historical data to generate deployment schedules on a daily basis. We accurately predict incidents using ST-ResNet, a deep learning technique that captures wide-ranging spatio-temporal dependencies, and solve a large-scale optimization problem to schedule deployment, significantly improving its scalability through a simulated annealing solver. Methodologically, our approach outperforms our previous works where prediction was done using Generative Adversarial Networks, and optimization was …

Multimodal Zero-Shot Hateful Meme Detection, Jiawen Zhu, Roy Ka-Wei Lee, Wen Haw Chong

Research Collection School Of Computing and Information Systems

Facebook has recently launched the hateful meme detection challenge, which garnered much attention in academic and industry research communities. Researchers have proposed multimodal deep learning classification methods to perform hateful meme detection. While the proposed methods have yielded promising results, these classification methods are mostly supervised and heavily rely on labeled data that are not always available in the real-world setting. Therefore, this paper explores and aims to perform hateful meme detection in a zero-shot setting. Working towards this goal, we propose Target-Aware Multimodal Enhancement (TAME), which is a novel deep generative framework that can improve existing hateful meme classification …

High-Resolution Face Swapping Via Latent Semantics Disentanglement, Yangyang Xu, Bailin Deng, Junle Wang, Yanqing Jing, Jia Pan, Shengfeng He

Research Collection School Of Computing and Information Systems

We present a novel high-resolution face swapping method using the inherent prior knowledge of a pre-trained GAN model. Although previous research can leverage generative priors to produce high-resolution results, their quality can suffer from the entangled semantics of the latent space. We explicitly disentangle the latent semantics by utilizing the progressive nature of the generator, deriving structure at-tributes from the shallow layers and appearance attributes from the deeper ones. Identity and pose information within the structure attributes are further separated by introducing a landmark-driven structure transfer latent direction. The disentangled latent code produces rich generative features that incorporate feature blending …

Reinforcement Learning Approach To Solve Dynamic Bi-Objective Police Patrol Dispatching And Rescheduling Problem, Waldy Joe, Hoong Chuin Lau, Jonathan Pan

Research Collection School Of Computing and Information Systems

Police patrol aims to fulfill two main objectives namely to project presence and to respond to incidents in a timely manner. Incidents happen dynamically and can disrupt the initially-planned patrol schedules. The key decisions to be made will be which patrol agent to be dispatched to respond to an incident and subsequently how to adapt the patrol schedules in response to such dynamically-occurring incidents whilst still fulfilling both objectives; which sometimes can be conflicting. In this paper, we define this real-world problem as a Dynamic Bi-Objective Police Patrol Dispatching and Rescheduling Problem and propose a solution approach that combines Deep …

One-Stage Blind Source Separation Via A Sparse Autoencoder Framework, Jason Anthony Dabin

Dissertations

Blind source separation (BSS) is the process of recovering individual source transmissions from a received mixture of co-channel signals without a priori knowledge of the channel mixing matrix or transmitted source signals. The received co-channel composite signal is considered to be captured across an antenna array or sensor network and is assumed to contain sparse transmissions, as users are active and inactive aperiodically over time. An unsupervised machine learning approach using an artificial feedforward neural network sparse autoencoder with one hidden layer is formulated for blindly recovering the channel matrix and source activity of co-channel transmissions. The BSS sparse autoencoder …

A Self-Learning Intersection Control System For Connected And Automated Vehicles, Ardeshir Mirbakhsh

Dissertations

This study proposes a Decentralized Sparse Coordination Learning System (DSCLS) based on Deep Reinforcement Learning (DRL) to control intersections under the Connected and Automated Vehicles (CAVs) environment. In this approach, roadway sections are divided into small areas; vehicles try to reserve their desired area ahead of time, based on having a common desired area with other CAVs; the vehicles would be in an independent or coordinated state. Individual CAVs are set accountable for decision-making at each step in both coordinated and independent states. In the training process, CAVs learn to minimize the overall delay at the intersection. Due to the …

Local Learning Algorithms For Stochastic Spiking Neural Networks, Bleema Rosenfeld

Dissertations

This dissertation focuses on the development of machine learning algorithms for spiking neural networks, with an emphasis on local three-factor learning rules that are in keeping with the constraints imposed by current neuromorphic hardware. Spiking neural networks (SNNs) are an alternative to artificial neural networks (ANNs) that follow a similar graphical structure but use a processing paradigm more closely modeled after the biological brain in an effort to harness its low power processing capability. SNNs use an event based processing scheme which leads to significant power savings when implemented in dedicated neuromorphic hardware such as Intel’s Loihi chip.

This work …

Optimization Opportunities In Human In The Loop Computational Paradigm, Dong Wei

Dissertations

An emerging trend is to leverage human capabilities in the computational loop at different capacities, ranging from tapping knowledge from a richly heterogeneous pool of knowledge resident in the general population to soliciting expert opinions. These practices are, in general, termed human-in-the-loop (HITL) computations.

A HITL process requires holistic treatment and optimization from multiple standpoints considering all stakeholders: a. applications, b. platforms, c. humans. In application-centric optimization, the factors of interest usually are latency (how long it takes for a set of tasks to finish), cost (the monetary or computational expenses incurred in the process), and quality of the completed …

Representation Learning In Finance, Ajim Uddin

Dissertations

Finance studies often employ heterogeneous datasets from different sources with different structures and frequencies. Some data are noisy, sparse, and unbalanced with missing values; some are unstructured, containing text or networks. Traditional techniques often struggle to combine and effectively extract information from these datasets. This work explores representation learning as a proven machine learning technique in learning informative embedding from complex, noisy, and dynamic financial data. This dissertation proposes novel factorization algorithms and network modeling techniques to learn the local and global representation of data in two specific financial applications: analysts’ earnings forecasts and asset pricing.

Financial analysts’ earnings forecast …

Nusax: Multilingual Parallel Sentiment Dataset For 10 Indonesian Local Languages, Genta Indra Winata, Alham Fikri Aji, Samuel Cahyawijaya, Rahmad Mahendra, Fajri Koto, Ade Romadhony, Kemal Kurniawan, David Moeljadi, Radityo Eko Prasojo, Pascale Fung, Timothy Baldwin, Jey Han Lau

Natural Language Processing Faculty Publications

Natural language processing (NLP) has a significant impact on society via technologies such as machine translation and search engines. Despite its success, NLP technology is only widely available for high-resource languages such as English and Chinese, while it remains inaccessible to many languages due to the unavailability of data resources and benchmarks. In this work, we focus on developing resources for languages in Indonesia. Despite being the second most linguistically diverse country, most languages in Indonesia are categorized as endangered and some are even extinct. We develop the first-ever parallel resource for 10 low-resource languages in Indonesia. Our resource includes …

Un-Fair Trojan: Targeted Backdoor Attacks Against Model Fairness, Nicholas Furth

Theses

Machine learning models have been shown to be vulnerable against various backdoor and data poisoning attacks that adversely affect model behavior. Additionally, these attacks have been shown to make unfair predictions with respect to certain protected features. In federated learning, multiple local models contribute to a single global model communicating only using local gradients, the issue of attacks become more prevalent and complex. Previously published works revolve around solving these issues both individually and jointly. However, there has been little study on the effects of attacks against model fairness. Demonstrated in this work, a flexible attack, which we call Un-Fair …

Collaborative Design And Simulation Integrated Method Of Civil Aircraft Take-Off Scenarios Based On X Language, Pengfei Gu, Lin Zhang, Zhen Chen, Junjie Ye

Journal of System Simulation

Abstract: For the large and complex products, the current traditional model-based systems engineering (MBSE) method of the integrated implementation of multiple modeling and simulation languages and platforms for system design and simulation verification can not ensure the efficient and accurate feedback of system design to realize the quick design optimization. X language, a new generation of integrated modeling and simulation language based on complex systems and supporting MBSE, is used to realize the integrated modeling and simulation on cross-domain subsystems of civil aircraft for the take-off scenarios. From the demand analysis of the take-off process of civil aircraft, the system-level …

Cross Level Switching Technology For Multi-Resolution Model Of Complex Products, Wei Li, Wenjia Zhang, Heming Zhang

Journal of System Simulation

Abstract: In the R&D of complex products, different design stages have different design goals and different simulation tasks, which need different resolution complex products simulation models. The models and interfaces for the multi-resolution characteristics are defined and thus the resolution control mechanism is studied. The description mechanisms for the system structure status and model resolution state are established separately, the control mechanism for the system resolution is proposed, and a cross level switching technology is sorted out. The experimental results show that the method can effectively solve the problem of model resolution switching, which ensures the simulation accuracy, improves …

A Cyber-Physical Integrated Modeling Method Oriented For Motion Simulation Of Complex Systems, Wenzheng Liu, Heming Zhang

Journal of System Simulation

Abstract: The traditional virtual modeling of motion simulation lacks the dynamic modeling of cyber subsystems and physical subsystems in complex systems. The advantages of the traditional kinematic virtual modeling and cyber calculation are combined, and aiming at the problem that the accuracy and real-time property of motion simulation cannot meet the actual industrial manufacturing requirements, a cyber physical integrated modeling method for the motion simulation of complex systems is proposed. The inconsistency between real robotic driving and virtual robot motion is solved, which is verified by a case study of mechanical arm motion control. A virtuality-reality mapping platform for complex …

Uav Formation Recovery And Consistency Simulation Based On Improved Potential Field, Ning Wang, Jiyang Dai, Jin Ying

Journal of System Simulation

Abstract: Aiming at the problems of multi-UAV formation collision avoidance, formation recovery and the consistency of position and velocity convergence, a distributed cooperative formation control algorithm based on the improved potential field principle and consistency theory is proposed. The static obstacle model, UAV particle model and the second-order system dynamic model are established; the coordination potential field function with coordination factors and communication weights is defined, which can achieve the control objectives of collision avoidance and formation recovery; on the basic consistency protocol, the formation center reference vector, expected speed and speed stabilization items are introduced to achieve the convergence …

Multi-Objective Optimization Configuration Of Agv System Based On Response Surface And Nsga-Ii, Jianlin Fu, Guofu Ding, Jian Zhang, Haifan Jiang, Peipei Guo

Journal of System Simulation

Abstract: Automated guided vehicle(AGV) system plays an important role in the production flexibility and efficiency in manufacturing systems. Due to the dynamic and stochastic characteristics of AGV system with many variables, its optimal configuration is relatively complex. A method combining system simulation, mathematical analysis and multi-objective optimization is proposed to optimize the configuration of AGV system. The discrete event simulation is used to simulate the operation of AGV system, the sensitivity analysis is used to separate design variables, the factorial experiments and response surface methods are used to build the fitting multi-objective optimization mathematical model, and the non-dominated sorting genetic …

Study On The Scale Characteristics Of Permeability Of Tpms Porous Materials, Tong Wu, Qinghui Wang, Zhijia Xu

Journal of System Simulation

Abstract: Triply Periodic Minimal Surface (TPMS) has been widely used in the design of porous materials, however, there is insufficient research on the scale characteristics of permeability. Four commonly used TPMS units are chosen as the research object, based on the introduction of their mathematical models and porosity control methods, a numerical simulation model based on CFD (computational fluid dynamics) is established; TPMS units and cubic porous structures with different porosities are analyzed, the quantitative correlation between their scales and permeability is clarified, i.e., within the selected scale range, the permeability of various TPMS units and cubic porous structures is …

Triangular Mesh Boolean Operation Method For Finite Element Analysis, Yufei Guo, Kang Zhao, Yongqing Hai

Journal of System Simulation

Abstract: To shorten the cycle of finite element analysis (FEA), an adaptive triangular mesh Boolean operation method for finite element analysis is proposed. The ADT (alternating digital tree) data structure is applied to the intersection calculation of triangular meshes, which improves the efficiency of the intersection calculation of Boolean operations. A sphere packing algorithm and a node addition/deletion algorithm are used to remesh some remeshing regions, which ensures the efficiency of the method and the high-quality of remeshed meshes. An improved octree background grid is used to record and smooth the size field, which can generate size-adaptive meshes. The size …

Cellular Automata Model Of Mixed Traffic Flow Composed Of Intelligent Connected Vehicles’ Platoon, Yangsheng Jiang, Sichen Wang, Kuan Gao, Meng Liu, Zhihong Yao

Journal of System Simulation

Abstract: To solve the existing cellular automata model of automatic-manual driving that does not consider the behavior of vehicle platoon, a cellular automata model of mixed traffic flow with the intelligent connected vehicles platoon is proposed, and the characteristics of mixed traffic flow are analyzed. The existing car-following behaviors in mixed traffic flow are analyzed. Based on the characteristics of the car-following behaviors, the cellular automata rules of human-driven vehicles (HDV), adaptive cruise control (ACC), and cooperative adaptive cruise control (CACC) are developed, respectively. Based on the numerical simulation experiments, the mixed traffic flow characteristics and congestion conditions are analyzed …

Water Body Extraction From High Resolution Remote Sensing Images Based On Fused Visual Word Bags, Xin Wang, Mingjun Xu, Jian Xiao, Lizhong Xu

Journal of System Simulation

Abstract: Aiming at the problem that water body extraction is easily influenced by shadow or light in high resolution remote sensing images, an improved algorithm based on fusion of visual word bags is proposed. Based on the deep analysis of the characteristics of remote sensing water body targets, a spectral feature extraction approach is designed. To enhance the description ability of water body targets, a novel visual word bag fusion model based on local binary pattern and spectral feature is constructed. Based on the proposed visual word bag fusion model, a water body target classifier is presented. …

Study On Building Fire Evacuation Path Planning Based On Improved Ant Colony Algorithm, Jiangtao Liang, Huiqin Wang

Journal of System Simulation

Abstract: Aiming at the problem of dynamic planning of evacuation paths in comprehensive building fires, with the shortest escape time required for evacuees as the goal, considering the impact of fire products and crowd density on the evacuation speed of personnel, an evacuation path planning model based on improved ant colony algorithm is constructed. A evacuation network data model composed of an obstacle vertex grid is established, the inspiration function of the ant colony algorithm and the deadlock processing strategy are improved, the explosion operator in the fireworks algorithm is introduced to optimize the ant path, and a comprehensive building …

Teaching-Learning-Based Optimization Algorithm For Permutation Flowshop Scheduling, Qiwen Zhang, Bin Zhang

Journal of System Simulation

Abstract: A multi-classes teaching-learning-based optimization (MCTLBO) algorithm is proposed for the permutation flowshop scheduling problem (PFSP) by combining continuous algorithm with discrete strategy. An improved nawaz enscore ham (NEH) population initialization method based on permutation mutation is adopted, which takes into account the quality and diversity of initial solutions. In the teaching stage, discrete adaptive teaching with duplicate removal is introduced to avoid meaningless teaching processes. A new self-learning strategy based on Levy flight is added, and the self-learning in discrete stage is simulated by variable neighborhood search. Learner phase and class communication are combined to improve the efficiency of …

Supply Chain Delivery Model And Simulation Based On Product Experience, Genshang Xing, Fang Lu, Shushan Li, Dingti Luo

Journal of System Simulation

Abstract: With the higher demand of consumers for delivery time of products purchased online, it is becoming more important for retailers to determine the delivery time of products based on product experience and market position. By constructing a supply chain delivery model under the three channel rights structure, the impact of product experience on the optimal delivery time, the retailer’s decision mode selection considering the delivery time and the supply chain profit after delivery are analyzed, and the correctness and reliability of the model are verified by numerical simulation. Research shows that the optimal product delivery time of retailers …

Murals Super-Resolution Reconstruction With The Stable Enhanced Generative Adversarial Network, Jianfang Cao, Yiming Jia, Minmin Yan, Xiaodong Tian

Journal of System Simulation

Abstract: Aiming at the problems of low resolution and unclear texture details of ancient murals, which led to insufficient viewing of murals and low research value, a stable enhanced super-resolution generative adversarial networks (SESRGAN) reconstruction algorithm is proposed. Based on the generative adversarial network, the generative network uses dense residual blocks to extract mural features, and uses the visual geometry group (VGG) network as the basic framework of the discriminating network to determine the authenticity of the input mural, and introduces perception loss, content loss and penalty loss to jointly optimize the model. Experimental results show that, compared with other …

Design Of Variable Stiffness Energy Storage Walking Assist Hip Exoskeleton And Simulation Of Assistance Effect, Bingshan Hu, Ke Cheng, Sheng Lu, Hongliu Yu

Journal of System Simulation

Abstract: Passive energy storage walking assist exoskeleton makes full use of the human’s own energy, reducing energy consumption when walking. Aiming at the present passive energy storage walking assist exoskeleton adopts fixed stiffness joint, a passive variable stiffness energy storage walking assist hip exoskeleton is designed, on the base of joint energy flow characteristics in the process of people walking and the change of stiffness characteristics. The human-exoskeletons coupling model is established, and the optimal stiffness that minimizes the power consumption of the human body walking on a flat surface, as well as the total metabolism and the main thigh …

Robot Path Planning Based On Bidirectional Aggregation Ant Colony Optimization, Xiangyang Deng, Limin Zhang, Wei Fang, Miao Tang

Journal of System Simulation

Abstract: Path planning is a key theoretical issue of the autonomous mobile robot technology. This paper utilizes an improved grid method to establish environment model, which involves a new priori advantage azimuth structure that includes two parts of the primary dominant grid cell and the subprime grid cell. It improves the pheromone mark ant colony optimization algorithm by putting forward a novel pheromone update strategy based on secondary path cognitive method, which is called bidirectional guidance strategies. It repeats an alternation of the starting point and the target point in each new round of iteration. The experimental results show that …

Research On Parametric Modeling And Deformation Method Of Human Muscle, Na Ni, Kunjin He, Xincheng Zhu, Zhengming Chen

Journal of System Simulation

Abstract: Due to the volume preserving constraint of human muscle in the deformation process, and the lack of three-dimensional multi-angle representation of muscle motion, a parametric modeling and deformation method of human muscle is proposed. Based on MRI (magnetic resonance imaging) data, the external contour line is extracted from the slice image generated by MRI data to construct a three-dimensional muscle model; muscle features are defined hierarchically, and the mapping relationship between semantic parameters is established to realize the volume preserving deformation of muscle; by establishing a vector-valued dynamic fourth-order differential equation, the feature curve dynamically simulates the process …

State Prediction Of Poverty Alleviation Objects Based On Hmm And Multidimensional Data, Jun He, Sunyan Hong, Yifang Zhou, Shikai Shen, Muquan Zou

Journal of System Simulation

Abstract: In order to solve the problems of inaccurate prediction of poverty, poverty reduction and poverty returen, and the difficulty in identifying the key factors affecting the state transition, 8 key features and 22 observed states are extracted from the poverty reduction basic data and multi-industry data. The relationship between observed state and implied state is constructed, and the hidden markov model (HMM) of poverty alleviation is established. Data of a deep poverty county for three consecutive years are used as samples for parameter training, test experiment and result verification. The results show that the method has a strong …

Simulation Of Navigation Process Based On Nonlinear Observer, Zhiwei Wang, Jizong Hu, Fengjie Wang, Jie Huang

Journal of System Simulation

Abstract: In order to solve the problem that the scope of using thetraditional observers is limited by assumptions, in the process of establishing the observer, a parameter projection relationship is designed and added to the observer to keep it under the condition of not being constrained by the assumptions. It is semi-globally stable, and the estimation process is more direct, which makes the parameter estimation process under nonlinear conditions converge faster. The simulation results show that the computational complexity of the nonlinear observer is reduced by nearly 80% compared with the multiplicative extended Kalman filter. The experimental results show that …