Using matrices to model symbolic relationship

By Ilya Sutskever, Geoffrey E. Hinton. 2008

27 citations recorded by the source corpus.

Subfields: Representation & Self-Supervised Learning

How researchers see this paper

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

In Representation & Self-Supervised Learning

Citations in the corpus

Neural Execution Engines: Learning to Execute Subroutines (2020, background): “[23] expands on this idea, modeling objects as matrices so that relationships can equivalently be treated as objects, allowing the system to learn higher-order relationships.”
Neural Programmer-Interpreters (2015, background): “This idea appeared in (Sutskever & Hinton, 2009) in the context of learning higher order symbolic relations and in (Donnarumma et al., 2015) as the key ingredient of an architecture for prefrontal cognitive control.”
Interaction Networks for Learning about Objects, Relations and Physics (2016, background): “Others have explored deep 54 learning of logical and arithmetic relations [26], and relations suitable for visual question-answering 55 [1].”
Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoning (2019, background): “Technically, this has been done by searching for the answer to a query that gives the highest grounding score [4, 3, 47, 48].”
Flexible Neural Representation for Physics Prediction (2018, positive): “Our work draws inspiration from and extends that of Chang et al. [11] and Battaglia et al. [4], which in turn use ideas from graph-based neural networks [39, 44, 9, 30, 22, 14, 13, 24, 8, 40].”
Neurosymbolic AI: the 3rd wave (2020, unknown): “gration, so-called embedding techniques seek to transform symbolic representations into vector spaces where reasoning can take place through matrix computations over distance functions [7, 85, 87, 80, 75, 11, 26, 100, 24, 70].”