Time for retrieval in recurrent associative memories
نویسندگان
چکیده
Experimental evidence shows that certain types of visual information processing, such as face recognition, may be extremely rapid: in a few tens of milliseconds a neuron may yield most of the information about a visual stimulus that can ever be extracted from the response of that neuron. These data might be taken to indicate that there is insufficient time for recurrent or feedback processing. However, a novel analytical method allows the analysis of the dynamics of an associative memory network of integrate-and-fire neurons, laterally connected through realistically modelled synapses. The analysis, supported by preliminary simulations, indicates that the network may retrieve information from memory over a time determined mainly by excitatory synaptic conductance time constants, that is in times in the order of tens of milliseconds. The results thus show that the dynamics of recurrent processing is sufficiently rapid for it to contribute to processing within the short times observed in neurophysiological experiments.
منابع مشابه
On the Recall Capability of Recurrent Exponential Fuzzy Associative Memories Based on Similarity Measures
Recurrent exponential fuzzy associative memories (RE-FAMs) are non-distributive memory models derived from the multivalued exponential recurrent associative memory (MERAM) of Chiueh and Tsai. A RE-FAM defines recursively a sequence of fuzzy sets obtained by a weighted average of the fundamental memories. In this paper, we show that the output of a single-step RE-FAM can be made as close as desi...
متن کاملGated Fast Weights for Associative Retrieval
We improve previous end-to-end differentiable neural networks (NNs) with fast weight memories. A gate mechanism updates fast weights at every time step of a sequence through two separate outer-product-based matrices generated by slow parts of the net. The system is trained on a complex sequence to sequence variation of the Associative Retrieval Problem with roughly 70 times more temporal memory...
متن کاملA neural network with a single recurrent unit for associative memories based on linear optimization
Recently, some continuous-time recurrent neural networks have been proposed for associative memories based on optimizing linear or quadratic programming problems. In this paper, a simple and efficient neural network with a single recurrent unit is proposed for realizing associative memories. Compared with the existing neural networks for associative memories, the main advantage of the proposed ...
متن کاملFast Weight Long Short-term Memory
Associative memory using fast weights is a short-term memory mechanism that substantially improves the memory capacity and time scale of recurrent neural networks (RNNs). As recent studies introduced fast weights only to regular RNNs, it is unknown whether fast weight memory is beneficial to gated RNNs. In this work, we report a significant synergy between long short-term memory (LSTM) networks...
متن کاملNeurogenesis and multiple plasticity mechanisms enhance associative memory retrieval in a spiking network model of the hippocampus
Hippocampal CA3 is crucial for the formation of long-term associative memory. It has a heavily recurrent connectivity, and memories are thought to be stored as memory engrams in the CA3. However, despite its importance for memory storage and retrieval, spiking network models of the CA3 to date are relatively small-scale, and exist as only proof-of-concept models. Specifically, how neurogenesis ...
متن کاملBipolar spectral associative memories
Nonlinear spectral associative memories are proposed as quantized frequency domain formulations of nonlinear, recurrent associative memories in which volatile network attractors are instantiated by attractor waves. In contrast to conventional associative memories, attractors encoded in the frequency domain by convolution may be viewed as volatile online inputs, rather than nonvolatile, off-line...
متن کامل