Filters








58,402 Hits in 2.7 sec

Incremental Multi-Step Q-Learning [chapter]

Jing Peng, Ronald J. Williams
1994 Machine Learning Proceedings 1994  
This paper presents a novel incremental algorithm that combines Q-learning, a well-known dynamicprogramming based reinforcement learning method, with the TD(A) return estimation process, which is typically  ...  The resulting algorithm, Q(A)-learning, thus combines some of the best features of the Q-learning and actor-critic learning paradigms.  ...  Introduction The incremental multi-step Q-learning (Q(A)-learning) method is a new direct (or modelfree) algorithm that extends the one-step Q-learning algorithm (Watkins, 1989 ) by combining it with  ... 
doi:10.1016/b978-1-55860-335-6.50035-0 dblp:conf/icml/PengW94 fatcat:vgztqp6jxbd3habgyvwaku56uu

Incremental multi-step Q-learning

Jing Peng, Ronald J. Williams
1996 Machine Learning  
This paper presents a novel incremental algorithm that combines Q-learning, a well-known dynamicprogramming based reinforcement learning method, with the TD(A) return estimation process, which is typically  ...  The resulting algorithm, Q(A)-learning, thus combines some of the best features of the Q-learning and actor-critic learning paradigms.  ...  Introduction The incremental multi-step Q-learning (Q(A)-learning) method is a new direct (or modelfree) algorithm that extends the one-step Q-learning algorithm (Watkins, 1989 ) by combining it with  ... 
doi:10.1007/bf00114731 fatcat:uy3yg4t3cbeirjjvyqx35g5yiy

Using multi-agent systems for learning optimal policies for complex problems

Andreas Lommatzsch, Sahin Albayrak
2007 Proceedings of the 45th annual southeast regional conference on - ACM-SE 45  
The basic idea is to apply reinforcement learning and to incrementally acquire knowledge about the implicit parameters dependencies. Based on the obtained data an optimal strategy is learned.  ...  For speeding up the learning process a multi-agent architecture is applied, that supports the simultaneous analysis of alternative strategies.  ...  Moreover Q-Learning works incrementally and improves the discovered solutions over time.  ... 
doi:10.1145/1233341.1233385 dblp:conf/ACMse/LommatzschA07 fatcat:ql4wbfmn6zamhlg2ii35gk7k74

Incremental Multiple Classifier Active Learning for Concept Indexing in Images and Videos [chapter]

Bahjat Safadi, Yubing Tong, Georges Quénot
2011 Lecture Notes in Computer Science  
In this paper, we propose a new incremental active learning algorithm based on multiple SVM for image and video annotation.  ...  Active learning with multiple classifiers has shown good performance for concept indexing in images or video shots in the case of highly imbalanced data.  ...  Fig. 1 . 1 The framework of the proposed incremental method with 40 steps.  ... 
doi:10.1007/978-3-642-17832-0_23 fatcat:ioap2si34zchbm5eoilt226xki

Incremental reinforcement learning for designing multi-agent systems

Olivier Buffet, Alain Dutech, François Charpillet
2001 Proceedings of the fifth international conference on Autonomous agents - AGENTS '01  
One solution to automatically build such large Multi-Agent Systems is to use decentralized learning : each agent learns by itself its own behavior.  ...  Our results show that incremental learning leads to better performances than more classical techniques. We then discuss several improvements which could lead to even better performances.  ...  Another aspect of using Q-Learning in a multi-agent system like we did still puzzles us.  ... 
doi:10.1145/375735.375826 dblp:conf/agents/BuffetDC01 fatcat:qsczgp7lejhmfitkmpelopgz7q

Fast Transformer Decoding: One Write-Head is All You Need [article]

Noam Shazeer
2019 arXiv   pre-print
requirements of incremental decoding.  ...  Multi-head attention layers, as used in the Transformer neural sequence model, are a powerful alternative to RNNs for moving information across and between sequences.  ...  part of the model took 222ms and each incremental step of the decoder took 47ms.  ... 
arXiv:1911.02150v1 fatcat:qwgqn3mcvnfv3b4kwemuoh4v4a

Knowledge transfer in multi-agent reinforcement learning with incremental number of agents

Wenzhang Liu, Lu Dong, Jian Liu, Changyin Sun
2022 Journal of Systems Engineering and Electronics  
In this paper, the reinforcement learning method for cooperative multi-agent systems (MAS) with incremental number of agents is studied.  ...  The existing multi-agent reinforcement learning approaches deal with the MAS with a specific number of agents, and can learn well-performed policies.  ...  reinforcement learning with incremental number of agents  ... 
doi:10.23919/jsee.2022.000045 fatcat:tw3nziv7ezdhpaveqyrih57oke

Developing Communication Strategy for Multi-Agent Systems with Incremental Fuzzy Model

Sam Hamzeloo, Mansoor Zolghadri
2018 International Journal of Advanced Computer Science and Applications  
An incremental method is also presented to create and tune our fuzzy model that reduces the high computational complexity of the multi-agent systems.  ...  its sequence of action-observation from previous communication, up to the current time step.  ...  Incremental Learning An incremental learning is a method that creates a model by recursively extracting required information from sequence of incoming data.  ... 
doi:10.14569/ijacsa.2018.090822 fatcat:rwld75dp7jhd5bnv5nlj7tdrae

Mnemonics Training: Multi-Class Incremental Learning without Forgetting [article]

Yaoyao Liu, Yuting Su, An-An Liu, Bernt Schiele, Qianru Sun
2020 arXiv   pre-print
Multi-Class Incremental Learning (MCIL) aims to learn new concepts by incrementally updating a model trained on previous concepts.  ...  However, there is an inherent trade-off to effectively learning new concepts without catastrophic forgetting of previous ones.  ...  Our work is conducted on the benchmarks of the latter one called multi-class incremental learning (MCIL).  ... 
arXiv:2002.10211v4 fatcat:zz4vaugqxreo5hl7phs46qhu5i

Incremental Sequence Learning [article]

Edwin D. de Jong
2016 arXiv   pre-print
Incremental Sequence Learning starts out by using only the first few steps of each sequence as training data.  ...  We study incremental learning in the context of sequence learning, using generative RNNs in the form of multi-layer recurrent Mixture Density Networks.  ...  Sequence Learning We study incremental learning in the context of sequence learning. The aim in sequence learning is to predict, given a step of the sequence, what the next step will be.  ... 
arXiv:1611.03068v2 fatcat:bai3txvuavdsllqwdi2rsz6mlu

Stochastic Gradient Trees [article]

Henry Gouk, Bernhard Pfahringer, Eibe Frank
2019 arXiv   pre-print
with batch multi-instance learning methods.  ...  In contrast to previous approaches to gradient-based tree learning, our method operates in the incremental learning setting rather than the batch learning setting, and does not make use of soft splits  ...  Multi-Instance Learning The evaluation metric used for multi-instance learning is the 10-fold cross-validation accuracy.  ... 
arXiv:1901.07777v3 fatcat:ecs4omwtkbg3lni7umr7tc56nu

A Clinical Decision Support Framework for Incremental Polyps Classification in Virtual Colonoscopy

Mariette Awad, Yuichi Motai, Janne Näppi, Hiroyuki Yoshida
2010 Algorithms  
The experimental results obtained from iteratively applying WP-SVM to improve detection sensitivity demonstrate its viability for incremental learning, thereby motivating further follow on research to  ...  Incremental data are incorporated in the WP-SVM as a weighted vector space, and the only storage requirements are the hyperplane parameters.  ...  To the best of our knowledge, no prior work has addressed incremental multi-classification of polyps using a support vector machine (SVM) approach within the framework of dynamic learning.  ... 
doi:10.3390/a3010001 fatcat:a76fwwu6nnajxcjmms2gzosuk4

Incremental multi-classifier learning algorithm on grid'5000 for large scale image annotation

Yubing Tong, Bahjat Safadi, Georges Quénot
2010 Proceedings of the international workshop on Very-large-scale multimedia corpus, mining and retrieval - VLS-MCMR '10  
In this paper, we proposed an incremental multi-classifier (SVM classifiers were used) learning algorithm for large scale imbalanced image annotation.  ...  With our previous research, active learning with multi-classifier showed considering performance in large scale data but much calculation was involved.  ...  PARALLEL INCREMENTAL MULTI-CLASSIFIER LEARNING ALGORITHM Incremental Multi-classifier Learning Algorithm For solving the large scale imbalanced data, active learning with multi-classifier was adopted  ... 
doi:10.1145/1878137.1878139 fatcat:xugz6e7lavasdjdcpjy37sfhbi

Flexible Workshop Scheduling Optimization Based On Multi-agent Technology

Jiang Xuesong, Tao Sun, Tao Qiaoyun, Jian Wang
2016 International Journal of Hybrid Information Technology  
This paper proposed an algorithm which was a combination of the ant colony algorithm and Q-learning algorithm.  ...  Considering the complexity of flexible workshop scheduling, combined with plant production process characteristics and constraints, we constructed a multi-agent system model to solve multi-objective flexible  ...  Q-Learning Algorithm Q-learning is one of the main algorithms, which was firstly proposed by Watkins in his doctoral dissertation in 1989 [10] .  ... 
doi:10.14257/ijhit.2016.9.5.25 fatcat:onbehq7mpbdxxejcxdoe345fhi

Learning without Memorizing [article]

Prithviraj Dhar, Rajat Vikram Singh, Kuan-Chuan Peng, Ziyan Wu, Rama Chellappa
2019 arXiv   pre-print
However, this is impractical as it increases the memory requirement at every incremental step, which makes it impossible to implement IL algorithms on edge devices with limited memory.  ...  Incremental learning (IL) is an important task aimed at increasing the capability of a trained model, in terms of the number of classes recognizable by the model.  ...  Here k is the number of incremental steps.  ... 
arXiv:1811.08051v2 fatcat:w4rqi5g4sjhhvimar6uuzkilyi
« Previous Showing results 1 — 15 out of 58,402 results