Holographic Recurrent Networks

By Tony A. Plate. 1992

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep learning in neural networks: An overview (2014, methodology): “10, 1991), RL economies in RNNs with WTA units and local learning rules (Schmidhuber, 1989b), and other methods (e.g., Bengio, Simard, & Frasconi, 1994; de Vries & Principe, 1991; Plate, 1993; Ring, 1993, 1994; Sun, Chen, & Lee, 1993).”
Composition in Distributional Models of Semantics (2010, methodology): “Circular convolution has been applied to memorising pen trajectories for handwritten digits (Plate, 1993) and complex semantic structures, consisting of agents playing particular roles in various actions (Plate, 1995).”
Supervised Sequence Labelling with Recurrent Neural Networks (2012, background): “, 1994), explicitly introduced time delays (Lang et al., 1990; Lin et al., 1996; Plate, 1993) or time constants (Mozer, 1992), and hierarchical sequence compression (Schmidhuber, 1992).”
Analyzing and Exploiting NARX Recurrent Neural Networks for Long-Term Dependencies (2017, methodology): “Notable approaches include Hessian-free optimization [29], operating explicitly at multiple time scales [9, 22], using associative or explicit memory [33, 7, 16, 39, 35], and initializing or restricting weight matrices to be orthogonal [1, 19].”
Bridging Long Time Lags by Weight Guessing and \Long Short Term Memory" (1996, background): “…tasks, though, are more di(cid:14)cult than the above benchmark problems: they cannot be solved at all in reasonable time by random search (we tried various architectures) nor any other recurrent net learning algorithm we are aware of, e.g., [25, 20, 7, 8, 33, 14, 21, 5, 15, 29, 1, 24, 26, 19].”
Long Short-Term Memory (1997, unknown): “Other methods that seem practical for short time lags only are time-delay neural networks (Lang, Waibel, & Hinton, 1990) and Plate’s method (Plate, 1993), which updates unit activations based on a weighted sum of old activations (see also de Vries & Principe, 1991).”