The Utility Driven Dynamic Error Propagation Network

How researchers see this paper

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

In Optimization & Training

Citations in the corpus

Deep learning in neural networks: An overview (2014, neutral): “…for sequenceprocessing recurrent NNs (e.g., Atiya & Parlos, 2000; Baldi, 1995; Gherrity, 1989; Kremer & Kolen, 2001; Pearlmutter, 1989, 1995; Robinson & Fallside, 1987; Rohwer, 1989; Schmidhuber, 1992a; Werbos, 1988;Williams, 1989;Williams & Peng, 1990;Williams & Zipser, 1988, 1989a,…”
Forecasting Across Time Series Databases using Long Short-Term Memory Networks on Groups of Similar Series (2017, methodology): “RNNs often use backpropagation through time (BPTT; Williams and Zipser, 1995) and real time recurrent learning (RTRL; Robinson and Fallside, 1987) as the learning algorithms.”
Sequence-to-sequence neural net models for grapheme-to-phoneme conversion (2015, methodology): “To train these encoder and decoder networks, we used back-propagation through time (BPTT) [26,27], with the error signal originating in the decoder network.”
PowerPlay: Training an Increasingly General Problem Solver by Continually Searching for the Simplest Still Unsolvable Problem (2011, background): “, [59, 61, 27, 32, 12]) are general c omputers that allow for both sequential and parallel computations, unlike the strictly sequential FORTH-like language of Section 4.”
Learning Precise Timing with LSTM Recurrent Networks (2003, methodology): “Like standard LSTM, but unlike BPTT and RTRL, LSTM with forget gates and peephole connections is local in space and time.”
On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models (2015, neutral): “Now a gradient descent technique [299, 317, 208](see Sec.”
One Big Net For Everything (2018, background): “Nevertheless, this is often much harder than imitating teachers through gradient-based supervised learning [99, 105, 47][76] for LSTM [24, 12, 19].”
Self-Delimiting Neural Networks (2012, methodology): “Supervised RNNs can be trained to map sequences of input patterns to desired output sequences by gradient descent and other methods [78, 81, 47, 52, 44, 31, 75].”
Driven by Compression Progress: A Simple Principle Explains Essential Aspects of Subjective Beauty, Novelty, Surprise, Interestingness, Attention, Curiosity, Creativity, Art, Science, Music, Jokes (2008, methodology): “Early work [57, 58, 61] described a predictor based on a recur nt neural network [115, 120, 55, 62, 47, 78] (in principle a rather powerful com putational device, even by today’s machine learning standards), predicting sensor y inputs including reward signals from the entire history of previous inpu”
LSTM can Solve Hard Long Time Lag Problems (1996, background): “, [15]), error signals \ owing backwards in time" tend to vanish: the temporal evolution of the backpropagated”
Learning Unambiguous Reduced Sequence Descriptions (1991, background): “Since only unpredictable inputs convey new information, ignore all predictable inputs but let all unexpected inputs (plus information about the time step at which they occurred) become inputs to a higher-level network of the same kind (working on a slower, self-adjusting time scale).”
Optimal Artiﬁcial Curiosity, Creativity, Music, and the Fine Arts (2005, methodology): “Our ﬁrst publications on artiﬁcial curiosity [21, 24] (1990, 1991) described a predictor based on a recurrent neural network [53, 54,20, 26, 17,37] (in principle a rather powerful computational device, even by today’s machine learning standards), predicting inputs x ( t ) and r ( t ) from the…”
Making Working Memory Work: A Computational Model of Learning in the Prefrontal Cortex and Basal Ganglia (2006, background): “A real-time recurrent backpropagation learning network (RBP; Robinson & Fallside, 1987; Schmidhuber, 1992; Williams & Zipser, 1992), with the same basic parameters as the SRN, and a time constant for integrating activations and backpropagated errors of 1, and the gap between backpropagations and…”
Binding via Reconstruction Clustering (2015, background): “In principle, Recurrent Neural Networks (RNNs; e.g. Robinson & Fallside, 1987; Werbos, 1988) can solve the binding problem by learning a mechanism to avoid it.”
LSTM: A Search Space Odyssey (2015, methodology): “Training was done using a mixture of real-time recurrent learning [23], [24] and backpropagation through time (BPTT) [24], [25].”
A Clockwork RNN (2014, background): “Recurrent Neural Networks (RNNs; Robinson & Fallside, 1987; Werbos, 1988; Williams, 1989) are a class of connectionist models that possess internal state or short term memory due to recurrent feed-back connections, that make them suitable for dealing with sequential problems, such as speech…”
New results on recurrent network training: unifying the algorithms and accelerating convergence (2000, background): “A N OVERVIEW OF THE ALGORITHMS The recurrent network training algorithms can be grouped into five major categories, which are described below (for details of the algorithms refer to the cited references and to the reviews [32]–[34]).”
Causal Analysis of Agent Behavior for AI Safety (2021, methodology): “The gradients of the recurrent network were computed with backpropagation through time (Robinson & Fallside, 1987; Werbos, 1988), and we used Adam for optimization (Kingma & Ba, 2014).”
A Fixed Size Storage O(n3) Time Complexity Learning Algorithm for Fully Recurrent Continually Running Networks (1992, background): “Such an algorithm is the RTRL-algorithm (Robinson and Fallside, 1987)(Williams and Zipser, 1989).”