Learning to Use Selective Attention and Short-Term Memory in Sequential Tasks

By Andrew McCallum. 1996

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, neutral): “…Jaakkola, Singh, & Jordan, 1995; Kaelbling et al., 1995; Kimura, Miyazaki, & Kobayashi, 1997; Lin, 1993; Littman, Cassandra, & Kaelbling, 1995; McCallum, 1996; Otsuka, Yoshimoto, & Doya, 2010; Ring, 1991, 1993, 1994; Schmidhuber, 1990d, 1991c; Teller, 1994; Wiering & Schmidhuber, 1996, 1998a;…”
Toward Interpretable Deep Reinforcement Learning with Linear Model U-Trees (2018, methodology): “U-tree [13] learning was developed as an online reinforcement learning algorithm with a tree structure representation.”
On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models (2015, background): “This triggered work on partially observable Markov decision problems (POMDPs) [223, 222, 227, 204, 205, 206, 316, 148, 278, 122, 152, 25, 114, 160, 126, 308, 309, 183].”
Policy Tree: Adaptive Representation for Policy Gradient (2015, methodology): “The U-Tree algorithm (McCallum 1996) is an example of this.”
Learning and Using Models (2012, methodology): “The U-TREE algorithm (McCallum, 1996) is one model-based algorithm that addresses this issue.”
A robot that reinforcement-learns to identify and memorize important previous observations (2003, methodology): “In contrast to memory bits and Utree , LSTM explicitly computes gradients with respect to past events, allowing more effective credit assignment, especially with respect to inputs from long ago and with many intermediate irrelevant inputs [1].”
CHILD: A First Step Towards Continual Learning (1997, methodology): “Very recently, McCallum has introduced the U-Tree algorithm (McCallum, 1996), which also takes an action-oriented approach.”
Imitative Reinforcement Learning for Soccer Playing Robots (2006, methodology): “Algorithms like U-tree [17] or HQ-learning [18] integrate past information into the current state.”