29,797 Hits in 6.0 sec

Learning for a Robot: Deep Reinforcement Learning, Imitation Learning, Transfer Learning

Jiang Hua, Liangcai Zeng, Gongfa Li, Zhaojie Ju
2021 Sensors  
The survey reveals that the latest research in deep learning and reinforcement learning has paved the way for highly complex tasks to be performed by robots.  ...  In view of the existing problem, this paper presents a state-of-the-art survey on an intelligent robot with the capability of autonomous deciding and learning.  ...  The policy-based algorithm could solve the problem of high cost in a real scenario and generate guided training samples by optimizing the trajectory distribution [62] .  ... 
doi:10.3390/s21041278 pmid:33670109 pmcid:PMC7916895 fatcat:ehzsevmddfg5zlyc2wms6yuhui

From Distributed Machine Learning to Federated Learning: A Survey [article]

Ji Liu, Jizhou Huang, Yang Zhou, Xuhong Li, Shilei Ji, Haoyi Xiong, Dejing Dou
2022 arXiv   pre-print
In this paper, we provide a comprehensive survey of existing works for federated learning. We propose a functional architecture of federated learning systems and a taxonomy of related techniques.  ...  Federated learning emerges as an efficient approach to exploit distributed data and computing resources, so as to collaboratively train machine learning models, while obeying the laws and regulations and  ...  Federated Learning (FL) emerges as an efficient approach to exploit the distributed resources to collaboratively train a machine learning model.  ... 
arXiv:2104.14362v3 fatcat:sgip6r7vy5djpfapk74jygldni

Transfer Learning in Deep Reinforcement Learning: A Survey [article]

Zhuangdi Zhu, Kaixiang Lin, Anil K. Jain, Jiayu Zhou
2022 arXiv   pre-print
Reinforcement learning is a learning paradigm for solving sequential decision-making problems.  ...  Specifically, we provide a framework for categorizing the state-of-the-art transfer learning approaches, under which we analyze their goals, methodologies, compatible reinforcement learning backbones,  ...  Although GAIL is more related to imitation learning than LfD, its philosophy of using expert demonstrations for distribution matching has inspired other LfD algorithms.  ... 
arXiv:2009.07888v5 fatcat:2rfeugb27ffv7jxh7siqn56s6e

Using Machine Learning on Sensor Data

Alexandra Moraru, Marko Pesko, Maria Porcius, Carolina Fortuna, Dunja Mladenic
2010 Journal of Computing and Information Technology  
We describe a vertical system integration of a sensor node and a toolkit of machine learning algorithms for predicting the number of persons located in a closed space.  ...  The dataset used as input for the learning algorithms is composed of automatically collected sensor data and additional manually introduced data.  ...  She started her collaboration with J. Stefan Institute in 2008, with a 2 months internship program, and since 2009 she is a student there.  ... 
doi:10.2498/cit.1001913 fatcat:wzgnylsuj5b3xdizwc6sflgiau

Augmented learning roads for Internet routing

John McCaffery, Alan Miller, Iain Oliver, Colin Allison
2014 2014 IEEE Frontiers in Education Conference (FIE) Proceedings  
Routing protocols can be included in the Computer Science curriculum in distributed systems, computer networking, algorithms, data structures, and graph theory.  ...  to augment conventional practice.  ...  Many smaller islands contain either single user sandbox areas, canned simulations for demonstration purposes, and shared simulation areas where multiple learners can collaboratively build a network.  ... 
doi:10.1109/fie.2014.7044337 dblp:conf/fie/McCafferyMOA14 fatcat:gxpp6d6y3ffbfj43r77wn76wcm

Semi-Supervised Imitation Learning of Team Policies from Suboptimal Demonstrations [article]

Sangwon Seo, Vaibhav V. Unhelkar
2022 arXiv   pre-print
Further, to allow for sample- and label-efficient policy learning from small datasets, BTIL employs a Bayesian perspective and is capable of learning from semi-supervised demonstrations.  ...  We present Bayesian Team Imitation Learner (BTIL), an imitation learning algorithm to model the behavior of teams performing sequential tasks in Markovian domains.  ...  Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.  ... 
arXiv:2205.02959v5 fatcat:ep5asleudbfe3psiyjbh77a4ji

Assessment of Learning in Digital Interactive Social Networks: A Learning Analytics Approach

Mark Wilson, Kathleen Scalise, Perman Gochyyev
2016 Online Learning  
This is followed by a description of the development of a "learning progression" for this project, as well as the logic behind the instrument construction and data analytics, along with examples of each  ...  The paper concludes with a discussion of the next steps for this effort.  ...  The tour through the site for the ATC21S demonstration scenario is conceived as a "collaboration contest," or virtual treasure hunt (see Figure 4 for a sample screen).  ... 
doi:10.24059/olj.v20i2.799 fatcat:m6o3iusvz5fmjotv6cuxgrnqxa

Learning Policy for Robot Anomaly Recovery Based on Robot Introspection [chapter]

Xuefeng Zhou, Hongmin Wu, Juan Rojas, Zhihao Xu, Shuai Li
2020 Nonparametric Bayesian Learning for Collaborative Robot Multimodal Introspection  
Then, we heuristically generate a set of synthetic demonstrations for augmenting the learning by appending a multivariate Gaussian noise distribution with mean equal to zeros and covariance equal to ones  ...  Zhou et al., Nonparametric Bayesian Learning for Collaborative Robot Multimodal Introspection, https://doi.  ...  • Learning from Demonstration Learning from demonstration (LfD) has been extensively used to program the robots, which aiming to provide a natural way to transfer human skills to robot.  ... 
doi:10.1007/978-981-15-6263-1_6 fatcat:v6cjejtmebdprayfwbjmgeoeku

Learning Algorithms for Active Learning [article]

Philip Bachman, Alessandro Sordoni, Adam Trischler
2017 arXiv   pre-print
We introduce a model that learns active learning algorithms via metalearning.  ...  For a distribution of related tasks, our model jointly learns: a data representation, an item selection heuristic, and a method for constructing prediction functions from labeled training sets.  ...  We demonstrate empirically that our proposed model learns effective active learning algorithms in an end-to-end fashion.  ... 
arXiv:1708.00088v1 fatcat:stki7ktg6bdrtfupfrkavqagdi

Distributed and Collaborative High Speed Inference Deep Learning for Mobile Edge with Topological Dependencies

Shagufta Henna, Alan Davy
2020 IEEE Transactions on Cloud Computing  
To bring more intelligence to the edge under topological dependencies, compared to optimization heuristics, this work proposes a novel collaborative distributed DL approach.  ...  By exploiting edge collaborative learning using stochastic gradient (SGD), the proposed approach called CGNN-edge ensures fast convergence and high accuracy.  ...  The system is based on an incentive-based mechanism for distributed resource allocation coupled with multi-stage stochastic programming. It also offers an auctionbased migration for load-balancing.  ... 
doi:10.1109/tcc.2020.2978846 fatcat:aw2lsygdpzg77fwdtivdhjwopq

Mutual Reinforcement Learning [article]

Sayanti Roy, Emily Kieson, Charles Abramson, Christopher Crick
2019 arXiv   pre-print
In this paper we demonstrate the application and effectiveness of a new approach called mutual reinforcement learning (MRL), where both humans and autonomous agents act as reinforcement learners in a skill  ...  While teaching skills in a physical (block-building) (n=34) or simulated (Tetris) environment (n=31), the expert tries to identify appropriate reward channels preferred by each individual and adapts itself  ...  ACKNOWLEDGMENTS This work was supported by NSF award #1527828 (NRI: Collaborative Goal and Policy Learning from Human Operators of Construction Co-Robots).  ... 
arXiv:1907.06725v3 fatcat:t6am5r6jzfc3pj5p664rrz7asy

Big Learning with Bayesian Methods [article]

Jun Zhu, Jianfei Chen, Wenbo Hu, Bo Zhang
2017 arXiv   pre-print
, regularized Bayesian inference for improving the flexibility via posterior regularization, and scalable algorithms and systems based on stochastic subsampling and distributed computing for dealing with  ...  This article provides a survey of the recent advances in Big learning with Bayesian methods, termed Big Bayesian Learning, including nonparametric Bayesian methods for adaptively inferring model complexity  ...  ACKNOWLEDGEMENTS The work is supported by National 973 Projects (2013CB329403), NSF of China Projects (61322308, 61332007), and Tsinghua Initiative Scientific Research Program (20121088071).  ... 
arXiv:1411.6370v2 fatcat:zmxse4kkqjgffkricevyumaoiu

Lessons Learned from Challenging Data Science Case Studies [chapter]

Kurt Stockinger, Martin Braschler, Thilo Stadelmann
2019 Applied Data Science  
by them.  ...  Secondly, the chapter serves as a digested, systematic summary of data science lessons that are relevant for data science practitioners.  ...  and the impact of data distribution on the runtime of SQL queries or machine learning algorithms.  ... 
doi:10.1007/978-3-030-11821-1_24 fatcat:6azhc4aon5eofi572joxaas5xq

Boosting Deep Transfer Learning for COVID-19 Classification [article]

Fouzia Altaf, Syed M.S. Islam, Naeem K. Janjua, Naveed Akhtar
2021 arXiv   pre-print
This paper provides an affirmative answer, devising a novel 'model' augmentation technique that allows a considerable performance boost to transfer learning for the task.  ...  Our method systematically reduces the distributional shift between the source and target domains and considers augmenting deep learning with complementary representation learning techniques.  ...  We further reduce the learning rate 10 times and fined-tuned these layers for 5 more epochs by augmenting the data with a random rotation in [-7,7] degrees, horizontal flip and cropping.  ... 
arXiv:2102.08085v1 fatcat:ceowqiriwnajrhwdmvenbnjykm

Learning multiple collaborative tasks with a mixture of Interaction Primitives

Marco Ewerton, Gerhard Neumann, Rudolf Lioutikov, Heni Ben Amor, Jan Peters, Guilherme Maeda
2015 2015 IEEE International Conference on Robotics and Automation (ICRA)  
Robots that interact with humans must learn to not only adapt to different human partners but also to new interactions. Such a form of learning can be achieved by demonstrations and imitation.  ...  A recently introduced method to learn interactions from demonstrations is the framework of Interaction Primitives.  ...  Based on this idea, Interaction Primitive (IP) is a framework that has been recently proposed to alleviate the problem of programming a robot for physical collaboration and assistive tasks [1] , [2]  ... 
doi:10.1109/icra.2015.7139393 dblp:conf/icra/EwertonNLA0M15 fatcat:3h4r57rjmrgtnjtcojcohq2mpu
« Previous Showing results 1 — 15 out of 29,797 results