Associative memory of connectivity patterns
نویسندگان
چکیده
The goal of the visual correspondence problem is to establish a connectivity pattern (a mapping) between two images such that features projected from the same scene point are connected. Dynamic link matching (DLM) is a self-organizing dynamical system to establish such connectivity patterns for object recognition, but with rather naturally given simple interactions between pattern elements, its organizing process is slow. Here we propose to stabilize (store) established mappings so that they can be recovered efficiently and reliably in the future. This is implemented by modifying the underlying system of interactions using the established mappings as learning examples, where the Hebbian rule makes the adapted interactions proportional to the weights of an associative memory of these mappings. It is shown in simulation that the adapted interactions lead to faster and more robust DLM. r 2006 Published by Elsevier B.V.
منابع مشابه
Associative Memory Models with Structured Connectivity
High capacity associative memory models with dilute structured connectivity are trained using naturalistic bitmap patterns. The connectivity of the model is chosen to reflect the local spatial continuity of the data. The results show that the structured connectivity gives the networks a higher effective capacity than equivalent randomly diluted networks. Moreover the locally connected networks ...
متن کاملRecovery of Performance in a Partially Connected Associative Memory Network Through Coding
Introducing partial connectivity to an associative memory network increases the variance of the dendritic sum distributions, reducing the performance. A coding scheme to compensate for this effect is considered, in which output patterns are self-organised by the network. It is shown using signal-to-noise ratio analysis that when the output patterns are self-organised the performance is greater ...
متن کاملThe performance of associative memory models with biologically inspired connectivity
This thesis is concerned with one important question in artificial neural networks, that is, how biologically inspired connectivity of a network affects its associative memory performance. In recent years, research on the mammalian cerebral cortex, which has the main responsibility for the associative memory function in the brains, suggests that the connectivity of this cortical network is far ...
متن کاملAn analysis of connectivity and function in hippocampal associative memory
We discuss eeects of realistically dilute connectivity on a neural network previously proposed as a model of hippocampal associative memory. Several criteria for setting neuronal activation thresholds are studied. Even with optimal thresholding, dilution induces a substantial information loss in stored patterns compared to presented patterns. Consequently, a stricter constraint than previous on...
متن کاملConnection Strategies in Associative Memory Models with Spiking and Non-spiking Neurons
The problem we address in this paper is that of finding effective and parsimonious patterns of connectivity in sparse associative memories. This problem must be addressed in real neuronal systems, so results in artificial systems could throw light on real systems. We show that there are efficient patterns of connectivity and that these patterns are effective in models with either spiking or non...
متن کاملFunctional Connectivity Relationships Predict Similarities in Task Activation and Pattern Information during Associative Memory Encoding
Neural systems may be characterized by measuring functional interactions in the healthy brain, but it is unclear whether components of systems defined in this way share functional properties. For instance, within the medial temporal lobes (MTL), different subregions show different patterns of cortical connectivity. It is unknown, however, whether these intrinsic connections predict similarities...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Neurocomputing
دوره 69 شماره
صفحات -
تاریخ انتشار 2006