The Tolman-Eichenbaum Machine: Unifying Space and Relational Memory through Generalization in the Hippocampal Formation (Whittington et al., 2020)

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• The spatial memory pipeline: A model of egocentric to allocentric understanding in mammalian brains (Uria et al., 2020)concept0.796
  Alternative neuroscience model analogous to TEM; the mathematical relationship to transformers also holds for this model.
• Tolman-Eichenbaum Machine (TEM)framework0.787
  Neuroscience model of hippocampal formation that the paper shows is mathematically equivalent to a transformer with recurrent position encodings.
• The hippocampus as a spatial map (O'Keefe & Dostrovsky, 1971)concept0.779
  Original discovery of place cells; the neural phenomenon TEM-t memory neurons resemble.
• Hippocampal representation of related and opposing memories (McKenzie et al., 2014)concept0.776
  Experimental evidence that hippocampal neurons respond to more than two task variables; motivates multi-input extension of TEM-t.
• Many apparently different object-oriented, functional, and relational mechanisms are closely related as variations within a general parameterizable n-way associative memory.claim0.773
• The Hippocampal Indexing Theory and Episodic Memory (Teyler & Rudy, 2007)concept0.768
  Original paper on hippocampal indexing theory; TEM-t is shown to instantiate this theory.
• Rahwan et al. 2019 called for a new field 'machine behavior' in which best explanations of machines combine holistic methods from ethology, social sciences, and cognitive science.finding0.763
  Cited as the key precedent motivating the paper's framework
• The hippocampus as a predictive map (Stachenfeld et al., 2017)concept0.760
  Graph representation model of hippocampus; related work category 2.