Analogues of mental simulation and imagination in deep learning
By Jessica B. Hamrick. 2019
79 citations recorded by the source corpus.
Mental simulation — the capacity to imagine what will or what could be — is a salient feature of human cognition, playing a key role in a wide range of cognitive abilities. In artificial intelligence, the last few years have seen the development of methods which are analogous to mental models and mental simulation. This paper outlines recent methods in deep learning for constructing such models from data and learning to use them via reinforcement learning, and compares such approaches to human mental simulation. Model-based methods in deep learning can serve as powerful tools for building and scaling cognitive models. However, a number of challenges remain in matching the capacity of human mental simulation for efficiency, compositionality, generalization, and creativity.
Subfields: Artificial Intelligence
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Sergey Levine
- J. Urgen Schmidhuber
- Yoshua Bengio
- Pieter Abbeel
- Nicolas Manfred Otto Heess
- Ruslan Salakhutdinov
- Josh Tenenbaum
- Chelsea Finn
Citations in the corpus
- On the Binding Problem in Artificial Neural Networks (2020, background): “Smolensky, 1990; Bader and Hitzler, 2005), especially in recent years (e.g. Lake et al., 2017; Battaglia et al., 2018; Hamrick, 2019; Garnelo and Shanahan, 2019).”
- Counterfactual Credit Assignment in Model-Free Reinforcement Learning (2020, background): “While this topic has recently generated a lot of interest (Heess et al., 2015; Ha & Schmidhuber, 2018; Hamrick, 2019; Kaiser et al., 2019; Schrittwieser et al., 2019), it remains difficult to model complex, partially observed environments.”
- The Tools Challenge: Rapid Trial-and-Error Learning in Physical Problem Solving (2019, background): “…progress on sample-efficient learning for complex tasks (Kaiser et al., 2019; Hafner et al., 2018; A. Xie, Ebert, Levine, & Finn, 2019; Holland, Talvitie, & Bowling, 2018; Hamrick, 2019) but the learned models are generally unstructured, and do not necessarily work well for long horizon predictions.”