![]() ![]() The scales from human behavior to synaptic connectivity. The neural computations underlying a class of visual decoding tasks, bridging Together, these results offer an understanding of Inductive biases in RNNs are important for uncovering how the human brain Ordinal relationship between the two orientations. Noise and reliance on the more stable and behaviorally relevant memory of the Test and support the hypothesis that human behavior is a product of both neural By varying the training conditions of the RNNs, we LSTM-SNP model is a recently developed long short-term memory (LSTM) network, which is inspired from the mechanisms of spiking neural P (SNP) systems. Our framework to a human psychophysics experiment in which subjects reported Multiple inputs enter a memory system via shared connections. We argue that such dynamic codes are generally required whenever Their tuning to prevent new sensory input from overwriting the previously Persistent activity, our networks also use a dynamic code where units change Moreover, in addition to attractors that store information via We find and characterize the connectivity patterns that support theĬlifford torus. Them in orthogonal subspaces, as demanded by the task, whereas a standard torusĭoes not. ![]() A Clifford torus treats the two orientations equally and keeps Of a donut) are topologically equivalent, they have important functionalĭifferences. Although a Clifford and standard torus (the surface Our model will be a feed forward neural network that takes in the difference between the current and previous screen patches. We find the activity manifold for the two orientations Neural circuits by training recurrent networks to report two previously shown ![]() We investigate how this structure is represented in When items take continuous values (e.g., orientation,Ĭontrast, length, loudness) they must be stored in a continuous structure ofĪppropriate dimensions. Cueva and 3 other authors Download PDF Abstract: Many daily activities and psychophysical experiments involve keeping multiple Download a PDF of the paper titled Recurrent neural network models for working memory of continuous variables: activity manifolds, connectivity patterns, and dynamic codes, by Christopher J. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |