Multiagent learning in adaptive dynamic systems

Burkov, Andriy; Chaib-draa, Brahim

doi:10.1145/1329125.1329174

Cited by 6 publications

(3 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Learning a model of state dynamics can result in a pre-trained hidden layer structure that reduces the training time in reinforce learning problems ( Anderson et al., 2015 ), and learning the deep Q networks from human demonstrators also helps to give a relatively good initial model and predict the dynamics ( Gabriel et al., 2019 ). There are many other applications of smart initialization on policy gradient methods ( Yun et al., 2017 ) and Q-learning methods ( Burkov and Chaib-Draa, 2007 ; Song et al., 2012 ), which speed up the learning and level up the performance ( Finn et al., 2016 ).…”

Section: Methods To Integrate Human Knowledgementioning

confidence: 99%

Integrating Machine Learning with Human Knowledge

et al. 2020

View full text Add to dashboard Cite

Summary Machine learning has been heavily researched and widely used in many disciplines. However, achieving high accuracy requires a large amount of data that is sometimes difficult, expensive, or impractical to obtain. Integrating human knowledge into machine learning can significantly reduce data requirement, increase reliability and robustness of machine learning, and build explainable machine learning systems. This allows leveraging the vast amount of human knowledge and capability of machine learning to achieve functions and performance not available before and will facilitate the interaction between human beings and machine learning systems, making machine learning decisions understandable to humans. This paper gives an overview of the knowledge and its representations that can be integrated into machine learning and the methodology. We cover the fundamentals, current status, and recent progress of the methods, with a focus on popular and new topics. The perspectives on future directions are also discussed.

show abstract

Section: Methods To Integrate Human Knowledgementioning

confidence: 99%

Integrating Machine Learning with Human Knowledge

et al. 2020

View full text Add to dashboard Cite

show abstract

“…The ADL algorithm fits multiple base classifiers to the training data during training. Each training iteration involves creating a new instance of the base classifier, fitting it to the training data, and storing the trained model [26].…”

Section: Mathematical Theory Of Adaptive Decision Learner (Adl)mentioning

confidence: 99%

From Deep Learning Maze to Neural Network Waltz: Unveiling Peak Performance in Stellar Classification (Using SDSS DR17)

Chatterjee,

Ghosh

2024

Preprint

View full text Add to dashboard Cite

Stellar classification based on spectral characteristics plays a pivotal role in astronomy, facilitating the study of celestial bodies’ composition and evolution. In this research, we assess the performance of ten distinct machine learning algorithms in classifying stellar objects using data from the Sloan Digital Sky Survey (SDSS). Leveraging features such as ’u’, ’g’, ’r’, ’i’, ’z’, and ’redshift’, with ’class’ as the target variable, we evaluate the accuracy of algorithms including XGBoost, CNN, RNN, AdaBoost, Adaptive Decision Learner, LSTM Networks, GRU, Random Forest Classifier, SVM, and Logistic Regression. Our findings reveal that the Random Forest Classifier outperforms other algorithms with an accuracy of 97.805%, showcasing its efficacy in capturing the complex spectral patterns of stellar objects. Moreover, other algorithms such as XGBoost, RNN, Adaptive Decision Learner, and GRU demonstrate notable accuracies ranging from 96.609% to 97.395%. This study underscores the utility of machine learning in stellar classification, offering valuable insights for astronomical research and enhancing our comprehension of the cosmos.

show abstract

“…to get desirable results [1,13,22]. Burkov and Chaib-draa [5] recently reported that mutual cooperation in PD games was realized just by using past action sequences as states of Q-learning.…”

Section: Related Workmentioning

confidence: 99%

Utility based Q-learning to facilitate cooperation in Prisoner's Dilemma games

Moriyama

2009

Web Intelligence and Agent Systems: An International Journal

View full text Add to dashboard Cite

This work deals with Q-learning in a multiagent environment. There are many multiagent Q-learning methods, and most of them aim to converge to a Nash equilibrium, which is not desirable in games like the Prisoner's Dilemma (PD). However, normal Q-learning agents that use a stochastic method in choosing actions to avoid local optima may yield mutual cooperation in a PD game. Although such mutual cooperation usually occurs singly, it can be facilitated if the Q-function of cooperation becomes larger than that of defection after the cooperation. This work derives a theorem on how many consecutive repetitions of mutual cooperation are needed to make the Q-function of cooperation larger than that of defection. In addition, from the perspective of the author's previous works that discriminate utilities from rewards and use utilities for learning in PD games, this work also derives a corollary on how much utility is necessary to make the Q-function larger by one-shot mutual cooperation.

show abstract

Multiagent learning in adaptive dynamic systems

Cited by 6 publications

References 4 publications

Integrating Machine Learning with Human Knowledge

Integrating Machine Learning with Human Knowledge

From Deep Learning Maze to Neural Network Waltz: Unveiling Peak Performance in Stellar Classification (Using SDSS DR17)

Utility based Q-learning to facilitate cooperation in Prisoner's Dilemma games

Contact Info

Product

Resources

About