Filters








2,905 Hits in 1.8 sec

Temporal binding as an inducer for connectionist recruitment learning over delayed lines

Cengiz Günay, Anthony S. Maida
2003 Neural Networks  
Second, we explore the practicality of temporal binding to drive a process of long-term memory formation based on a recruitment learning method (Feldman, 1982) .  ...  Complementing similar approaches, we implement a continuous-time learning procedure allowing computation with spiking neurons.  ...  Overview of Recruitment Learning Briefly, recruitment learning is a concept production scheme which allocates units from a pool of free units that are embedded in a random graph (Feldman, 1982) .  ... 
doi:10.1016/s0893-6080(03)00117-5 pmid:12850012 fatcat:eqltqfld5rfl7pfwjsaczt37am

Using temporal binding for connectionist recruitment learning over delayed lines

C. Gunay, A.S. Maida
Proceedings of the International Joint Conference on Neural Networks, 2003.  
Second, we explore the practicality of temporal binding to drive a process of long-term memory formation based on a recruitment learning method [2].  ...  Complementing similar approaches, we implement a continuous-time learning procedure allowing computation with spiking neurons.  ...  Why Recruitment Learning?  ... 
doi:10.1109/ijcnn.2003.1223348 fatcat:oirjwtkcpzbyxhhwdulbko2l3y

Simulating Stages of Human Cognitive Development with Connectionist Models [chapter]

Thomas R. Shultz
1991 Machine Learning Proceedings 1991  
Partial problem solving in connectionist networks is likely to occur under the following conditions: hidden unit herding, over-generalization, training pattern bias, and hidden unit recruitment.  ...  All of these findings were found to be consistent with new connectionist models of cognitive development.  ...  HIDDEN UNIT RECRUITMENT The recruitment of hidden units as needed during learning is a technique used by Cascade-Correlation (Fahlman & Lebiere, 1990 ) and a few other connectionist algorithms.  ... 
doi:10.1016/b978-1-55860-200-7.50025-8 dblp:conf/icml/Shultz91 fatcat:cs4ugxrqvzh4hcxv7okkj7qsm4

An extended local connectionist manifesto: Embracing relational and procedural knowledge

Lokendra Shastri
2000 Behavioral and Brain Sciences  
In the recruitment learning framework, learning occurs within a network of quasi-randomly connected nodes.  ...  The learning framework presented by the author has a strong overlap with work on recruitment learning Wickelgren, 1979; Feldman, 1982; Shastri, 1988; 1999b; 1999c; Diederich, 1989; Valiant, 1994 .  ... 
doi:10.1017/s0140525x00493350 fatcat:ddqsspgdpbcb5om5hogvlgdfne

The Implications of Connectionism and Neural Networks for Simultaneous Interpreting

2014 US-China Foreign Language  
Neural networks, in connectionist models, acquire and learn diverse cognitive patterns through the strengthening and weakening of nodes associations.  ...  The aim of this paper is to approach simultaneous interpreting adopting some connectionist models and conceptions and, thus, nurture interpreting modeling.  ...  Similarly, representing the recruitment learning framework, Shastri (1999, p. 1) referred to the process of learning to be based on the 'quasi-randomly' connected networks of neural nodes.  ... 
doi:10.17265/1539-8080/2014.02.005 fatcat:iaug5dqqpnfjzj2bs4fdrrvfee

Learning structured representations

Lokendra Shastri, Carter Wendelken
2003 Neurocomputing  
shruti is a connectionist model that demonstrates how a network of neuron-like elements can encode a large body of semantic, episodic, and causal knowledge, and rapidly make decisions and perform explanatory  ...  Previous work has already demonstrated the rapid learning of episodic facts via cortico-hippocampal interactions.  ...  Learning in structured connectionist models Two forms of learning, both related to hebbian learning, form the basis of learning structured representations in shruti.  ... 
doi:10.1016/s0925-2312(02)00840-8 fatcat:s3455dv2izhsxmrjt3mj2thcey

Generation, Local Receptive Fields and Global Convergence Improve Perceptual Learning in Connectionist Networks

Vasant G. Honavar, Leonard Uhr
1989 International Joint Conference on Artificial Intelligence  
This paper presents and compares results for three types of connectionist networks on perceptual learning tasks: [A] Multi-layered converging networks of neuron-like units, with each unit connected to  ...  generation-discovery, which involves the growth of links and recruiting of units in addition to reweighting of links.  ...  Connectionist Networks That Learn By Generation and Discovery As Well As Re-weighting Connectionist network structures that learn by generation and re-weighting of links and recruiting of new nodes from  ... 
dblp:conf/ijcai/HonavarU89 fatcat:si7hoff7ubg6nnlnsakrfo5hqi

Connectionist Models of Reinforcement, Imitation, and Instruction in Learning to Solve Complex Problems

F. Dandurand, T.R. Shultz
2009 IEEE Transactions on Autonomous Mental Development  
We modeled learning by reinforcement (rewards) using SARSA, a softmax selection criterion and a neural network function approximator; learning by imitation using supervised learning in a neural network  ...  ; and learning by instructions using a knowledge-based neural network.  ...  KBCC allows networks to recruit existing knowledge in the service of new learning.  ... 
doi:10.1109/tamd.2009.2031234 fatcat:uqzm4qowq5hm3g7uiecmyok4lq

Page 2364 of Psychological Abstracts Vol. 90, Issue 7 [page]

2003 Psychological Abstracts  
A modi- fied recruitment rule is introduced that creates new conceptual clusters in response to surprising events during learning.  ...  The implications of us- ing a unified recruitment method for both supervised and unsupervised learning are discussed. 20978. Hummel, John E. & Holyoak, Keith J.  ... 

Development: it's about time

Jeff Elman
2003 Developmental Science  
Rather, these brain areas have intrinsic capabilities that lend themselves to being recruited -as a result of learned expertise -to serve the specific needs of chess.  ...  Indeed, the title of one of the most influential papers on connectionist learning was 'Learning internal representations by error propagation' (Rumelhart, Hinton & Williams, 1986) .  ... 
doi:10.1111/1467-7687.00297 fatcat:fuiulshjenf47mog4h5fg4wpj4

Learning structured representations from experience [chapter]

Leonidas A.A. Doumas, Andrea E. Martin
2018 The psychology of learning and motivation  
Newly recruited P units in the recipient learn connections to active recipient RB units, and newly recruited RB units learn connections to active PO units (i.e., DORA learns connections between the new  ...  (E) DORA recruits a P unit in the recipient. (F-G) DORA learns a connection between the recruited P unit and the active RB unit (the RB coding for larger + bowl).  ... 
doi:10.1016/bs.plm.2018.10.002 fatcat:s25msu3l7ra7dhhsbdg55rtsni

Connectionist models of development

Yuko Munakata, James L. McClelland
2003 Developmental Science  
Connectionist and dynamic systems approaches to development have differed, with connectionist approaches focused on learning processes and representations in cognitive tasks, and dynamic systems approaches  ...  Why are there critical periods for some types of learning? Why do children sometimes show U-shaped performance curves as they learn, initially getting worse at a skill before ultimately succeeding?  ...  For example, 'generative' connectionist architectures have been investigated (Mareschal & Shultz, 1996) , in which new units (along with connections to other units) are recruited into a network as a function  ... 
doi:10.1111/1467-7687.00296 fatcat:xwchto5rxfbsbg7qlxdxl6ljti

Modeling cognitive development on balance scale phenomena

Thomas R. Shultz, Denis Mareschal, William C. Schmidt
1994 Machine Learning  
Cascade-correlation is a generative connectionist algorithm that constructs its own network topology as it learns.  ...  Cascade-correlation networks provided better fits to these human data than did previous models, whether rule-based or connectionist.  ...  Of the 16 nets, nine of them recruited a single hidden unit, six recruited two hidden units, and one recruited three hidden units.  ... 
doi:10.1007/bf00993174 fatcat:7puasfmi7fgj5kjiagpvisnwbu

Thinking about the Self from a Social Cognitive Neuroscience Perspective

Lian Rameson, Matthew D. Lieberman
2007 Psychological Inquiry  
Simple associative and declarative models of self-representations are not well equipped to handle these nonlinear dynamics, but connectionist models excel at capturing such dynamics (Hopfield, 1982)  ...  More specifically, the authors use a connectionist framework to model temporal, cultural and social aspects of self which may give rise to the enculturated stream of consciousness which is seen the essential  ...  ideal candidate for connectionist models.  ... 
doi:10.1080/10478400701416228 fatcat:g7km4cbfgzdcncwlm2x4kihy7u

On the Integration of Knowledge Graphs into Deep Learning Models for a More Comprehensible AI—Three Challenges for Future Research

Giuseppe Futia, Antonio Vetrò
2020 Information  
In contexts where the impact of AI on human life is relevant (e.g., recruitment tools, medical diagnoses, etc.), explainability is not only a desirable property, but it is -or, in some cases, it will be  ...  Deep learning models contributed to reaching unprecedented results in prediction and classification tasks of Artificial Intelligence (AI) systems.  ...  Background: Learning Approach in Connectionist and Symbolic AI In this Section we describe the main learning principles behind deep learning models (connectionist AI) and KGs and ontologies (symbolic AI  ... 
doi:10.3390/info11020122 fatcat:77ni2i6tdrhqxopw25vbybghi4
« Previous Showing results 1 — 15 out of 2,905 results