Open Access. Powered by Scholars. Published by Universities.®

Artificial Intelligence and Robotics Commons™

Open Access. Powered by Scholars. Published by Universities.®

Discipline
Institution
Keyword
Publication Year
Publication
Publication Type
File Type

Articles 5191 - 5220 of 11187

Full-Text Articles in Artificial Intelligence and Robotics

Interdisciplinary Communication By Plausible Analogies: The Case Of Buddhism And Artificial Intelligence, Michael Cooper Jun 2022

Interdisciplinary Communication By Plausible Analogies: The Case Of Buddhism And Artificial Intelligence, Michael Cooper

USF Tampa Graduate Theses and Dissertations

Communicating interdisciplinary information is difficult, even when two fields are ostensibly discussing the same topic. In this work, I’ll discuss the capacity for analogical reasoning to provide a framework for developing novel judgments utilizing similarities in separate domains. I argue that analogies are best modeled after Paul Bartha’s By Parallel Reasoning, and that they can be used to create a Toulmin-style warrant that expresses a generalization. I argue that these comparisons provide insights into interdisciplinary research. In order to demonstrate this concept, I will demonstrate that fruitful comparisons can be made between Buddhism and Artificial Intelligence research.

Go to article

Explainable And Cooperative Autonomy Across Networks Of Distributed Systems, Peter Joseph Jorgensen Jun 2022

Explainable And Cooperative Autonomy Across Networks Of Distributed Systems, Peter Joseph Jorgensen

USF Tampa Graduate Theses and Dissertations

Large networks of complex systems-of-systems are commonplace and evermore present in both mundane and extraordinary facets of human existence. From the exponential growth of connectivity via the internet and other information networks, to the miniaturization of computers and sensors, to cross-domain sensor and communication networks, these networks of distributed systems-of-systems (NDSS) present incredible benefits and challenges. Autonomy is perhaps the most important and most difficult to achieve enabling technology for efficient performance of the NDSS. Giving each individual agent in a network the ability to manage its internal state in dynamic operating environments and in pursuit of multiple complex and …

Go to article

Edgenext: Efficiently Amalgamated Cnn-Transformer Architecture For Mobile Vision Applications, Muhammad Maaz, Abdelrahman Shaker, Hisham Cholakkal, Salman Khan, Syed Waqas Zamir, Rao Anwer, Fahad Shahbaz Khan Jun 2022

Edgenext: Efficiently Amalgamated Cnn-Transformer Architecture For Mobile Vision Applications, Muhammad Maaz, Abdelrahman Shaker, Hisham Cholakkal, Salman Khan, Syed Waqas Zamir, Rao Anwer, Fahad Shahbaz Khan

Computer Vision Faculty Publications

In the pursuit of achieving ever-increasing accuracy, large and complex neural networks are usually developed. Such models demand high computational resources and therefore cannot be deployed on edge devices. It is of great interest to build resource-efficient general purpose networks due to their usefulness in several application areas. In this work, we strive to effectively combine the strengths of both CNN and Transformer models and propose a new efficient hybrid architecture EdgeNeXt. Specifically in EdgeNeXt, we introduce split depth-wise transpose attention (SDTA) encoder that splits input tensors into multiple channel groups and utilizes depth-wise convolution along with self-attention across channel …

Go to article

Dynamic Prototype Convolution Network For Few-Shot Semantic Segmentation, Jie Liu, Yanqi Bao, Guo-Sen Xie, Huan Xiong, Jan-Jakob Sonke, Efstratios Gavves Jun 2022

Dynamic Prototype Convolution Network For Few-Shot Semantic Segmentation, Jie Liu, Yanqi Bao, Guo-Sen Xie, Huan Xiong, Jan-Jakob Sonke, Efstratios Gavves

Machine Learning Faculty Publications

The key challenge for few-shot semantic segmentation (FSS) is how to tailor a desirable interaction among sup-port and query features and/or their prototypes, under the episodic training scenario. Most existing FSS methods im-plement such support/query interactions by solely leveraging plain operations - e.g., cosine similarity and feature concatenation - for segmenting the query objects. How-ever, these interaction approaches usually cannot well capture the intrinsic object details in the query images that are widely encountered in FSS, e.g., if the query object to be segmented has holes and slots, inaccurate segmentation al-most always happens. To this end, we propose a dynamic …

Go to article

An Unsupervised Deep Neural Network For Image Fusion, Peipei Zhou, Xinglin Hou Jun 2022

An Unsupervised Deep Neural Network For Image Fusion, Peipei Zhou, Xinglin Hou

Journal of System Simulation

Abstract: Due to the low dynamic range of camera, can not be expressed in the different region of the high dynamic scene a single-exposure image. An unsupervised depth neural network is constructed to fuse the multi-exposure images into a high dynamic image. Based on the VGG-Net, encoding and decoding sub-networks are designed. Guided by the structural similarity of the images before and after fusion, a loss function suitable for image fusion is designed by introducing the weight factors based on the local image information, and the valid information of the different input images is given consideration. Compared with the …

Go to article

Research And Simulation Of Internet Of Vehicles Task Offloading Based On Mobile Edge Computing, Peng Cheng, Wenzhu Zhang, Shuhan Xie, Zixuan Yang Jun 2022

Research And Simulation Of Internet Of Vehicles Task Offloading Based On Mobile Edge Computing, Peng Cheng, Wenzhu Zhang, Shuhan Xie, Zixuan Yang

Journal of System Simulation

Abstract: In order to use the computing resources of edge devices to provide high-quality services, a joint resource allocation and task offloading mechanism is designed for the Internet of Vehicles architecture based on mobile edge computing. In the mechanism, the original problem is decomposed into two sub-problems of resource allocation and offloading decision. The original problem is simplified into the resource allocation of maximizing system capacity, and the initial offloading set is obtained through the proportional resource allocation algorithm; the above problem is solved by the joint offloading decision-making and resource allocation mechanism. The stable experimental results are obtained …

Go to article

Design And Realization Of 6-Dof Parachuting Simulation Training System, Xiaoguang Zhou, Peng Zhu, Yuanyuan Zhang, Huan Lu, Yuan Zhou Jun 2022

Design And Realization Of 6-Dof Parachuting Simulation Training System, Xiaoguang Zhou, Peng Zhu, Yuanyuan Zhang, Huan Lu, Yuan Zhou

Journal of System Simulation

Abstract: Aiming at restoring the 6-DOF motion process of each stage during parachuting, a 6-DOF parachuting simulation training system is designed and implemented. The architecture of the simulator is designed, and the realization of the sub-systems such as the motion calculation, 6-DOF motion platform, control loading system, virtual reality scene, somatosensory system and management console is explained. Compared with the same type of parachute simulator, this system has introduced a 6-DOF motion platform which can drive the trainees to simulate the various postures of parachuting. It can also help the trainees master the control methods of parachute and enhance the …

Go to article

A High Spectral-Efficiency Maritime Very-High-Frequency Communication Technology And Simulation, Xinyu Dou, Xiaohui Chen, Dequn Liang, Bin Lin Jun 2022

A High Spectral-Efficiency Maritime Very-High-Frequency Communication Technology And Simulation, Xinyu Dou, Xiaohui Chen, Dequn Liang, Bin Lin

Journal of System Simulation

Abstract: Marine communications cannot follow the development route on land of developing high frequency resources and high spectral-efficiency modulation technologies in 5G and 6G high-speed communication. The data rate and spectral-efficiency of maritime communications are low and are difficult to be improved. A high spectral-efficiency maritime very-high-frequency (VHF) communication technology based on multi-carrier time-delay overlapping modulation (MC-TDOP) is proposed. The core is delaying the subcarriers in turn and directly overlapping them in time domain. The orthogonality between the subcarriers can be neglected, thus the spectral-efficiency can be further enhanced from the fundamental information modulations. The results show that the proposed …

Go to article

Simulation Method Of Virtual Human Pose Optimization Based On Vr Peripherals, Muqing Wang, Lei Zhang, Xiumin Fan, Xiaomeng Luo, Wenmin Zhu Jun 2022

Simulation Method Of Virtual Human Pose Optimization Based On Vr Peripherals, Muqing Wang, Lei Zhang, Xiumin Fan, Xiaomeng Luo, Wenmin Zhu

Journal of System Simulation

Abstract: The commonly used VR peripheral is the position tracker worn on the operator body and is not very convenient. Focus on the problem, a visual sensor combination-based scheme is proposed to realize the simple and real-time unmarked virtual human driven simulation. On the basis of the parameterized human model SMPL, by minimizing the objective function of multiple error terms, the driven accuracy of the virtual human is improved, and the fitted human body model conforming to the operator form and posture is calculated and the ergonomic evaluation is realized. The module is verified by a case of evaluation in …

Go to article

Multi-Person Interactive Globe System Based On Ar Technology, Yiling Sun, Yi Chen, Guihua Shan, Xiaoxing Li Jun 2022

Multi-Person Interactive Globe System Based On Ar Technology, Yiling Sun, Yi Chen, Guihua Shan, Xiaoxing Li

Journal of System Simulation

Abstract: The characteristic of multi-source, high-dimensional, time-varying and massive of the Earth big data is difficult to be understood and analyzed. Aiming at this problem and for the science popularization needs, an AR-based multi-person interactive globe system is proposed and implemented. A system architecture integrating AR technology is proposed to realize the seamless overlay combination effect of the virtual information and the physical globe. A data visualization display scheme is designed to realize the visualization of the Earth big data in three-dimensional space. A lightweight multi-person multi-terminal collaboration mechanism is proposed to improve the practicality and interestingness of the system. …

Go to article

Multi-Agent Simulation For Online Fresh Food Autonomous Delivery, Miaojia Lu, Chengyuan Huang, Jing Teng Jun 2022

Multi-Agent Simulation For Online Fresh Food Autonomous Delivery, Miaojia Lu, Chengyuan Huang, Jing Teng

Journal of System Simulation

Abstract: Autonomous delivery can solve the last-mile delivery problems of low efficiency, high manual cost, and potential safety hazard. The autonomous delivery of the online fresh food in urban communities is discussed and a data-driven agent-based platform with the actual spatial-temporal demand is built. Three kinds of agents including the autonomous vehicles, customers, and distribution center and the simulation environment based on the actual road network are construct. To achieve the objectives of the minimum total operating costs and maximum customer satisfaction, the different static and dynamic order dispatch strategies and the route planning strategies with the principle of …

Go to article

Denoising Algorithm Based On Multi-Feature Non-Local Mean Filtering For Monte Carlo Rendered Images, Kai Yang, Chunyi Chen, Xiaojuan Hu, Haiyang Yu Jun 2022

Denoising Algorithm Based On Multi-Feature Non-Local Mean Filtering For Monte Carlo Rendered Images, Kai Yang, Chunyi Chen, Xiaojuan Hu, Haiyang Yu

Journal of System Simulation

Abstract: Aiming at the rendering noise in Monte Carlo synthesized images induced by the low light-path sampling rate, a denoising algorithm based on the multi-feature non-local-mean filtering is proposed. The gradient image of the scene's albedo information is calculatedwith the canny operator, and a guided filter together with the said gradient image is employed to prefilter the normal vector image. The structural similarity of the sub-blocks in the prefiltered normal vector image is calculated and the improved weights of the non-local mean filter are computed according to the logarithmic value of the reciprocal of the structural similarity. The improved …

Go to article

Multi-Uavs 3d Path Planning Method Based On Random Strategy Search, Sen Zhang, Mengyan Zhang, Jingping Shao, Jiexin Pu Jun 2022

Multi-Uavs 3d Path Planning Method Based On Random Strategy Search, Sen Zhang, Mengyan Zhang, Jingping Shao, Jiexin Pu

Journal of System Simulation

Abstract: In view of the difficulty of the traditional path planning method without energy consumption constraints to meet the emergency rescue requirements in the complex mountain operation environment, a three-dimensional path planning algorithm for multi-UAVs is proposed based on LSTM-DPPO(long short-term memory-distributed proximal policy optimization) framework. The LSTM long and short-term memory neural network is used to extract the important characteristic state information sequence of the multiple unmanned aerial vehicles in their respective flight process. After repeated iteration and updating, an optimal network parameter model is obtained. Combined with the energy consumption, the optimal 3D detection path is generated. …

Go to article

Simulation Of Multi-Layer Ship Evacuation System Based On Improved A* Algorithm, Dun Meng, Zhuo Hu, Huajun Zhang Jun 2022

Simulation Of Multi-Layer Ship Evacuation System Based On Improved A* Algorithm, Dun Meng, Zhuo Hu, Huajun Zhang

Journal of System Simulation

Abstract: Aiming at the low efficiency of emergency evacuation at sea, an emergency evacuation system based on improved A* algorithm is proposed. Based on the network flow model, the traversal mode of the adjacency node is used to complete the path search, and the influence of the path personnel density and path obstacles is added to the calculation of the cost, which makes the algorithm more practical. In order to improve the efficiency of the algorithm, the node optimization of the network is carried out, and a multi-path optimal scheme is proposed in the case of single layer with multiple …

Go to article

Image Center Layout Optimization Method Based On Improved Genetic Algorithm, Zhijie Li, Haoqi Shi, Changhua Li, Jie Zhang Jun 2022

Image Center Layout Optimization Method Based On Improved Genetic Algorithm, Zhijie Li, Haoqi Shi, Changhua Li, Jie Zhang

Journal of System Simulation

Abstract: Aiming at the layout optimization methods of image center being influenced by the subjective factors and low level of automation, a method of combining systematic layout planning(SLP) with the improved genetic algorithm is proposed. The layout scheme generated by SLP improves the initial population of the genetic algorithm and increases the diversity of the initial population. In order to improve the efficiency of optimization, the improved algorithm updates the crossover probability and mutation probability adaptively according to the evolution stages and the fitness value of the individuals. On the basis of the layout area model and multi-objective optimization mathematical …

Go to article

Airport Flight Transit Support Time Collaborative Planning Modeling Based On Stn, Bin Chen, Yue Liu, Yalei Yang Jun 2022

Airport Flight Transit Support Time Collaborative Planning Modeling Based On Stn, Bin Chen, Yue Liu, Yalei Yang

Journal of System Simulation

Abstract: Under the constraint of resources, the collaborative planning of airport flight transit support time is one of the effective methods to improve airport operation efficiency. Based on Simple Temporal Network (STN), a planning model of flight transit support time is established. Based on the temporal decoupling, the shortest path matrix simplification, and the distance graph solving of STN task model considering resources, the method of collaborative planning of flight transit support time for airport considering resources is obtained. The comparison results of the simulation and the actual data show that STN task model considering resources can optimize the airport …

Go to article

Vehicle Detection Method Based On Multi Scale Feature Fusion, Yin Wang, Feixiang Wang, Qianlai Sun Jun 2022

Vehicle Detection Method Based On Multi Scale Feature Fusion, Yin Wang, Feixiang Wang, Qianlai Sun

Journal of System Simulation

Abstract: Vehicle detection is the important research content and hotspot in the intelligent transportation. Aiming at the low detection accuracy and poor small-scale recognition effect of the traditional vehicle detection algorithm, an improved detection method based on YOLOv4(you only look once v4) is proposed to improve the detection performance of small target vehicles in traffic scenes. By redesigning the YOLOv4 network, the MobileNetv2 deep separable convolution module is used to replace the traditional convolution, and the convolutional block attention module (CBAM) attention module is integrated into the feature extraction network to ensure the detection accuracy of the model and reduce …

Go to article

A Hybrid Algorithm Based On Seeker Optimization Algorithm And Salp Swarm Algorithm For Pid Parameters Optimization, Shaomi Duan, Huilong Luo, Haipeng Liu Jun 2022

A Hybrid Algorithm Based On Seeker Optimization Algorithm And Salp Swarm Algorithm For Pid Parameters Optimization, Shaomi Duan, Huilong Luo, Haipeng Liu

Journal of System Simulation

Abstract: Aiming at the premature convergence of seeker optimization algorithm(SOA) during optimizing the global problems, a new SOA-SSA hybrid algorithm based on seeker optimization algorithm and salp swarm algorithm (SSA) is proposed.The SOA-SSA algorithm is based on a double population evolution strategy, in which some individuals of the population are evolved by seeker optimization algorithm and the rest are evolved from salp swarm algorithm. The individuals in SOA and SSA both employ an information sharing mechanism to realize the coevolution. These strategies increase the diversity of the population and avoid the premature convergence. The experimental results show that …

Go to article

Application Of Improved Q Learning Algorithm In Job Shop Scheduling Problem, Yejian Zhao, Yanhong Wang, Jun Zhang, Hongxia Yu, Zhongda Tian Jun 2022

Application Of Improved Q Learning Algorithm In Job Shop Scheduling Problem, Yejian Zhao, Yanhong Wang, Jun Zhang, Hongxia Yu, Zhongda Tian

Journal of System Simulation

Abstract: Aiming at the job shop scheduling in a dynamic environment, a dynamic scheduling algorithm based on an improved Q learning algorithm and dispatching rules is proposed. The state space of the dynamic scheduling algorithm is described with the concept of "the urgency of remaining tasks" and a reward function with the purpose of "the higher the slack, the higher the penalty" is disigned. In view of the problem that the greedy strategy will select the sub-optimal actions in the later stage of learning, the traditional Q learning algorithm is improved by introducing an action selection strategy based on the …

Go to article

Visual Inspection Model Of Uav Cluster Based On Improved Pigeon Flock Hierarchy, Qi Chen, Haoyang Cui Jun 2022

Visual Inspection Model Of Uav Cluster Based On Improved Pigeon Flock Hierarchy, Qi Chen, Haoyang Cui

Journal of System Simulation

Abstract: Aim at UAV being vulnerable to the environmental interference and the low efficiency of the traditional single-person-UAV model in the transmission line inspection, a visual inspection model for the power line by UAV is proposed based on the improved pigeon flock hierarchy. The initial landmark point of the UAV is generated based on GPS coordinates of the aircraft-carrying vehicle and the tower to be inspected, and the movement trajectory is planned. The return point of the UAV is used to update the initial landmark of onward UAV, which realizes the dynamic handover between the work-exchanging UAV, and the landmark …

Go to article

Design Of Interactive Simulated Water Gun Fire Fighting Training System Based On Steam Vr, Cheng Lu, Xuesheng Jin Jun 2022

Design Of Interactive Simulated Water Gun Fire Fighting Training System Based On Steam Vr, Cheng Lu, Xuesheng Jin

Journal of System Simulation

Abstract: In order to save the fire fighting training resources and increase the immersion and experience of VR training, an interactive simulated water gun fire fighting training system based on Steam VR is designed. By using the Hall sensors and signal conversion circuit boards to collect and transmit the signal of the simulated water gun, and by using the Unity3D engine combined with the VIVE head-mounted display to build and present VR fire scene. The gun is controlled through C# programming to complete the interaction with the virtual fire scene. The system is evaluated by a post-questionnaire survey …

Go to article

Application Of Observability In Performance Evaluation Of Photosynthesis Model, Hongnai Gao, Lijiang Fu, Qian Xia, Ya Guo Jun 2022

Application Of Observability In Performance Evaluation Of Photosynthesis Model, Hongnai Gao, Lijiang Fu, Qian Xia, Ya Guo

Journal of System Simulation

Abstract: The photochemical reaction of photosynthesis involves a variety of physiological substances that cannot be directly measured. By modeling the control system, the state of these physiological substances can be es-timated based on the chlorophyll fluorescence, but the reliability of the state estimation is not given in all the reference documents. In response to this problem, based on the photochemical reaction kinetic model, the observability of the nonlinear system is introduced to evaluate the reliability of the state estimation. Aiming at the existing observability methods lacking the direct comparability due to the different dimensions of the components of different states, …

Go to article

Fuzzy Super-Twisting Second Order Sliding Mode Trajectory Tracking Control For Robotic Manipulator, Baolin Du, Dachang Zhu, Yihua Pan Jun 2022

Fuzzy Super-Twisting Second Order Sliding Mode Trajectory Tracking Control For Robotic Manipulator, Baolin Du, Dachang Zhu, Yihua Pan

Journal of System Simulation

Abstract: A fuzzy super-twisting second order sliding mode control method is proposed for the uncertainties of the model error and external disturbance on the trajectory tracking accuracy of robotic manipulator. Based on the dynamic model of the robotic, a new non-singular terminal sliding mode manifold is designed, and an improved super-twisting algorithm is used to design the second order sliding mode controller. In order to solve the problem that the matching disturbance can only be compensated under the condition of the known disturbance boundary in the sliding mode control, the fuzzy logic algorithm is used to carryout the online compensation …

Go to article

Research On Fire Emergency Evacuation Simulation Based On Cooperative Deep Reinforcement Learning, Lingjia Ni, Xiaoxia Huang, Hongga Li, Zibo Zhang Jun 2022

Research On Fire Emergency Evacuation Simulation Based On Cooperative Deep Reinforcement Learning, Lingjia Ni, Xiaoxia Huang, Hongga Li, Zibo Zhang

Journal of System Simulation

Abstract: The fire accident is a major threat to the public safety, in which the high temperature, toxic and harmful gases seriously interfer the selection of the evacuation routes. Deep reinforcement learning is introduced into the research of emergency evacuation simulation, and a cooperative double deep Q network algorithm is proposed for the multi-agent environment. A fire scene model that changes dynamically over time is established to provide the real-time information on the distribution of the dangerous areas for the evacuation. The independent agent neural networks are integrated and the multi-agent unified deep neural network is established to realize the …

Go to article

Machine Learning With Kay, Lasith Niroshan, James Carswell Jun 2022

Machine Learning With Kay, Lasith Niroshan, James Carswell

Conference Papers

Computational power is very important when training Deep Learning (DL) models with large amounts of data (Wooldridge, 2021). Hence, High-Performance Computing (HPC) can be leveraged to reduce computational cost, and the Irish Centre for High-End Computing (ICHEC) provides significant infrastructure and services for research and development to both academia and industry. A portion of ICHEC's HPC system has been allocated for institutional access, and this paper presents a case study of how to use Kay (Ireland's national supercomputer) in the remote sensing domain. Specifically, this study uses clusters of Kay Graphics Processing Units (GPUs) for training DL models to extract …

Go to article

Learning To Generalize Dispatching Rules On The Job Shop Scheduling, Zangir Iklassov, Dmitrii Medvedev, Ruben Solozabal, Martin Takac Jun 2022

Learning To Generalize Dispatching Rules On The Job Shop Scheduling, Zangir Iklassov, Dmitrii Medvedev, Ruben Solozabal, Martin Takac

Machine Learning Faculty Publications

This paper introduces a Reinforcement Learning approach to better generalize heuristic dispatching rules on the Job-shop Scheduling Problem (JSP). Current models on the JSP do not focus on generalization, although, as we show in this work, this is key to learning better heuristics on the problem. A well-known technique to improve generalization is to learn on increasingly complex instances using Curriculum Learning (CL). However, as many works in the literature indicate, this technique might suffer from catastrophic forgetting when transferring the learned skills between different problem sizes. To address this issue, we introduce a novel Adversarial Curriculum Learning (ACL) strategy, …

Go to article

An Empirical Study On Sampling Approaches For 3d Image Classification Using Deep Learning, Nicholas Michelette Jun 2022

An Empirical Study On Sampling Approaches For 3d Image Classification Using Deep Learning, Nicholas Michelette

Theses and Dissertations

A 3D classification method requires more training data than a 2D image classification method to achieve good performance. These training data usually come in the form of multiple 2D images (e.g., slices in a CT scan) or point clouds (e.g., 3D CAD modeling) for volumetric object representation. The amount of data required to complete this higher dimension problem comes with the cost of requiring more processing time and space. This problem can be mitigated with data size reduction (i.e., sampling). In this thesis, we empirically study and compare the classification performance and deep learning training time of PointNet utilizing uniform …

Go to article

What-Is And How-To For Fairness In Machine Learning: A Survey, Reflection, And Perspective, Zeyu Tang, Jiji Zhang, Kun Zhang Jun 2022

What-Is And How-To For Fairness In Machine Learning: A Survey, Reflection, And Perspective, Zeyu Tang, Jiji Zhang, Kun Zhang

Machine Learning Faculty Publications

Algorithmic fairness has attracted increasing attention in the machine learning community. Various definitions are proposed in the literature, but the differences and connections among them are not clearly addressed. In this paper, we review and reflect on various fairness notions previously proposed in machine learning literature, and make an attempt to draw connections to arguments in moral and political philosophy, especially theories of justice. We also consider fairness inquiries from a dynamic perspective, and further consider the long-term impact that is induced by current prediction and decision. In light of the differences in the characterized fairness, we present a flowchart …

Go to article

Learning To Control Under Time-Varying Environment, Yuzhen Han, Ruben Solozabal, Jing Dong, Xingyu Zhou, Martin Takac, Bin Gu Jun 2022

Learning To Control Under Time-Varying Environment, Yuzhen Han, Ruben Solozabal, Jing Dong, Xingyu Zhou, Martin Takac, Bin Gu

Machine Learning Faculty Publications

This paper investigates the problem of regret minimization in linear time-varying (LTV) dynamical systems. Due to the simultaneous presence of uncertainty and non-stationarity, designing online control algorithms for unknown LTV systems remains a challenging task. At a cost of NP-hard offline planning, prior works have introduced online convex optimization algorithms, although they suffer from nonparametric rate of regret. In this paper, we propose the first computationally tractable online algorithm with regret guarantees that avoids offline planning over the state linear feedback policies. Our algorithm is based on the optimism in the face of uncertainty (OFU) principle in which we optimistically …

Go to article

Flecs: A Federated Learning Second-Order Framework Via Compression And Sketching, Artem Agafonov, Dmitry Kamzolov, Rachael Tappenden, Alexander Gasnikov, Martin Takac Jun 2022

Flecs: A Federated Learning Second-Order Framework Via Compression And Sketching, Artem Agafonov, Dmitry Kamzolov, Rachael Tappenden, Alexander Gasnikov, Martin Takac

Machine Learning Faculty Publications

Inspired by the recent work FedNL (Safaryan et al, FedNL: Making Newton-Type Methods Applicable to Federated Learning), we propose a new communication efficient second-order framework for Federated learning, namely FLECS. The proposed method reduces the high-memory requirements of FedNL by the usage of an L-SR1 type update for the Hessian approximation which is stored on the central server. A low dimensional 'sketch' of the Hessian is all that is needed by each device to generate an update, so that memory costs as well as number of Hessian-vector products for the agent are low. Biased and unbiased compressions are utilized to …

Go to article

First Prev Next Last