Filters








150 Hits in 4.6 sec

Lamarckian Evolution and the Baldwin Effect in Evolutionary Neural Networks [article]

P.A. Castillo, M.G. Arenas, J.G. Castellano, J.J. Merelo, A. Prieto, V. Rivas, G. Romero
2006 arXiv   pre-print
In the case of Darwinian evolution, the Baldwin effect, that is, the progressive incorporation of learned characteristics to the genotypes, can be observed and leveraged to improve the search.  ...  Two ways of combining learning and genetic search are explored: one exploits the Baldwin effect, while the other uses a Lamarckian strategy.  ...  Acknowledgements This work has been supported in part by CI-CYT TIC99-0550, INTAS 97-30950, HPRN-CT-2000-00068 and IST-1999-12679.  ... 
arXiv:cs/0603004v1 fatcat:4dzlfofpr5df5lmyzfwxfraszi

Meta-learning by the baldwin effect

Chrisantha Fernando, Jakub Sygnowski, Simon Osindero, Jane Wang, Tom Schaul, Denis Teplyashin, Pablo Sprechmann, Alexander Pritzel, Andrei Rusu
2018 Proceedings of the Genetic and Evolutionary Computation Conference Companion on - GECCO '18  
Whilst in simple cases MAML is more data efficient than the Baldwin effect, the Baldwin effect is more general in that it does not require gradients to be backpropagated to the reference parameters or  ...  hyperparameters, and permits effectively any number of gradient updates in the inner loop.  ...  The Baldwin effect in neural network learning has been investigated in several papers following the original work of Hinton and Nowlan [7] .  ... 
doi:10.1145/3205651.3205763 dblp:conf/gecco/FernandoSOWSTSP18 fatcat:gdtvorekirgsvhhs5acq3dufaq

Adding Learning to the Cellular Development of Neural Networks: Evolution and the Baldwin Effect

Frédéric Gruau, Darrell Whitley
1993 Evolutionary Computation  
Our results suggest that merely using learning to change the fimess landscape can be as effective as Lamarckian strategies at improving search. genetic algorithms, neural networks, learning developmental  ...  Three ways of adding learning to the development process are explored. Two of these exploit the Baldwin e -e c t by changing the fitness landscape without using Lamarckian evolution.  ...  Whitley acknowledges the contributions that Rik Belew and David Ackley have made to this work through their publications on learning and evolution as well as the ideas they have shared in informal conversations  ... 
doi:10.1162/evco.1993.1.3.213 fatcat:5y3kad2fybel7nh5s6v7ubw5we

Evolving Differentiable Gene Regulatory Networks [article]

Dennis G Wilson, Kyle Harrington, Sylvain Cussat-Blanc, Hervé Luga
2018 arXiv   pre-print
Using a standard machine learning dataset, we evaluate the ways in which evolution and SGD can be combined to further GRN optimization.  ...  We compare these approaches with neural network models trained by SGD and with support vector machines.  ...  The evolutionary advantage of plasticity was first proposed in [Baldwin, 1896] and is now referred to as the Baldwin Effect.  ... 
arXiv:1807.05948v1 fatcat:wudsfwd2sbahhf7asgw5xseizm

Evolution, Learning, and Instinct: 100 Years of the Baldwin Effect

Peter Turney, Darrell Whitley, Russell W. Anderson
1996 Evolutionary Computation  
Batali and Grundy systematically explore the evolution of six different neural network architectures in three different simulated worlds.  ...  This "recovery effect" is presented as a special case of the "relearning effect" described by neural network researchers (Hinton and Sejnowski, 1986; Harvey, 1996) .  ...  Darrell Whitley was supported by NSF grants IRI-9312748 and IRI-9503366.  ... 
doi:10.1162/evco.1996.4.3.iv fatcat:s2ecjwvbanbrnlxlg5qt4r6bfe

The Baldwin Effect Revisited: Three Steps Characterized by the Quantitative Evolution of Phenotypic Plasticity [chapter]

Reiji Suzuki, Takaya Arita
2003 Lecture Notes in Computer Science  
For this purpose, we investigate the evolution of connection weights in a neural network under the assumption of epistatic interactions.  ...  An interaction between evolution and learning called the Baldwin effect has been known for a century, but it is still poorly appreciated.  ...  Fig. 1 . 1 Two steps of the Baldwin effect Fig. 2 . 2 The topology of neural network in case of N = 5, M = 3 uals in a population and each individual has a feed-forward multi-layer neural network which  ... 
doi:10.1007/978-3-540-39432-7_42 fatcat:24n5hy6gbrd6zndng6nbfhj2uu

Language evolution and the Baldwin effect

Yusuke Watanabe, Reiji Suzuki, Takaya Arita
2008 Artificial Life and Robotics  
In order to examine whether and how the Baldwin effect might occur, we use a mechanism for the evolution of the plasticity (learnability) of each weight in the neural network as we 3 did in [6] . 7  ...  Key words: language evolution, Baldwin effect, genetic algorithm, recurrent neural network, artificial life.  ...  Here, we investigate the evolutionary dynamics which governs the Baldwin effect.  ... 
doi:10.1007/s10015-007-0443-y fatcat:njpc2wsopjdc3hrqptfxnujbta

Evolutionary Computation: An Overview

Melanie Mitchell, Charles E. Taylor
1999 Annual Review of Ecology and Systematics  
Yet these simple rules are thought to be responsible for the extraordinary variety and complexity we see in the biosphere.  ...  Other common forms of evolutionary algorithms will be described in the third section.  ...  The Baldwin effect has been observed in evolutionary computation studies (see, e.g., 1, 30) .  ... 
doi:10.1146/annurev.ecolsys.30.1.593 fatcat:k4tagcnyzvcjtbbaetp56ar7di

Knowledge Incorporation in Evolutionary Computation [Book Review]

Qingfu Zhang
2006 IEEE Computational Intelligence Magazine  
The first chapter in this part, Chapter 17, studies the Baldwin effect and Lamarckian evolution in a cellular genetic algorithm for optimization of recurrent neural networks.  ...  It is concluded that the Lamarckian evolution is more efficient in the optimization of recurrent neural networks.  ... 
doi:10.1109/mci.2006.329699 fatcat:op7hlvz2mzgozmdrhwmxbuzrj4

Evolutionary Speech Recognition [chapter]

Anne Spalanzani
2007 Robust Speech Recognition and Understanding  
Hybridization Three major approaches combining neural networks and evolutionary algorithms are presented in the literature: finding the optimal weights of the neural network (using evolution only or combined  ...  In gray, recognition rate of the "neural network population", in black, the recognition rate of the "evolutionary neural network population" One can notice the regularity of the performances of the evoluted  ...  How to reference In order to correctly reference this scholarly work, feel free to copy and paste the following: Anne Spalanzani (2007) .  ... 
doi:10.5772/4744 fatcat:nv4nfhbit5derdfhvv6m3zdkzm

How to Shift Bias: Lessons from the Baldwin Effect

Peter Turney
1996 Evolutionary Computation  
The Baldwin effect was proposed in 1896, to explain how phenomena that might appear to require Lamarckian evolution (inheritance of acquired characteristics) can arise from purely Darwinian evolution.  ...  Hinton and Nowlan presented a computational model of the Baldwin effect in 1987.  ...  In a typical hybrid of evolutionary computation and neural networks, evolutionary search is responsible for determining the architecture of the network (the connections and the size of the hidden layer  ... 
doi:10.1162/evco.1996.4.3.271 fatcat:h6lhg3nbcnbzxmclildohcd5fe

Learning and Evolution of Autonomous Adaptive Agents [chapter]

Vladimir G. Red'ko, Danil V. Prokhorov
2010 Studies in Computational Intelligence  
We show that the Baldwin effect can be observed in our model, viz., originally acquired adaptive policy of best agent-brokers becomes inherited over the course of the evolution.  ...  Each agent is equipped with a neural network adaptive critic design for behavioral adaptation. We discuss three cases in which either learning, or evolution, or both, are active in our model.  ...  This work is partially supported by the Russian Foundation for Basic Research, Grant No 07-01-00180 and the Program of the Presidium of the Russian Academy of Science "Intelligent informational technologies  ... 
doi:10.1007/978-3-642-05177-7_25 fatcat:mrarapojmbbsnlhzlxxpynjj2y

Investigations into Lamarckism, Baldwinism and Local Search in Grammatical Evolution Guided by Reinforcement

Jack Mario Mingo, Ricardo Aler, Darío Maravall, Javier de Lope Asiaín
2013 Computing and informatics  
In this paper the role of the Lamarck Hypothesis is reviewed and a solution inspired only in the Baldwin effect is included as well.  ...  Grammatical Evolution Guided by Reinforcement is an extension of Grammatical Evolution that tries to improve the evolutionary process adding a learning process for all the individuals in the population  ...  Ackley and Littman proposed an adaptation strategy known as Evolutionary Reinforcement Learning (ERL) which combined evolution with neural network learning and they studied how several populations behaved  ... 
dblp:journals/cai/MingoAMA13 fatcat:dugwc2lsurgr3g7fzq2fw2aldu

Evolving Learnable Neural Networks Under Changing Environments with Various Rates of Inheritance of Acquired Characters: Comparison of Darwinian and Lamarckian Evolution

Takahiro Sasaki, Mario Tokoro
1999 Artificial Life  
In this article, we study the relationship between learning and evolution in a simple abstract model, where neural networks capable of learning are evolved using genetic algorithms (GAs).  ...  The processes of adaptation in natural organisms consist of two complementary phases: learning, occurring within each individual's lifetime, and evolution, occurring over successive generations of the  ...  By contrast, in deliberately constructed artificial evolutionary systems, Lamarckian evolution can be both easy to implement and potentially more effective" (p. 3).  ... 
doi:10.1162/106454699568746 pmid:10648951 fatcat:sm73lxypzrctxiqnk7ckmnwpay

Genetics-Based Machine Learning [chapter]

Tim Kovacs
2012 Handbook of Natural Computing  
from weak to strong bias See [119] for a Lamarckian LCS Baldwin effect: smoothing Baldwin effect I: smoothing fitness landscape Phenotypic Plasticity: the ability to adapt (e.g. learn) during lifetime  ...  behaviour becomes instinctive Allows learned behaviours to become inherited without Lamarckian inheritance Baldwin effect: bias Baldwin effect and bias [293] Weak bias = learning Strong bias  ... 
doi:10.1007/978-3-540-92910-9_30 fatcat:rm5bx5lwdvfalolrky6lpyt67a
« Previous Showing results 1 — 15 out of 150 results