Christopher J. Cueva, Xue-Xin Wei
2018 · arXiv
Decades of research on the neural code underlying spatial navigation have revealed a diverse set of neural response properties. The Entorhinal Cortex (EC) of the mammalian brain contains a rich set of spatial correlates, including grid cells which encode space using tessellating patterns. However, the mechanisms and functional significance of these spatial representations remain largely mysterious. As a new way to understand these neural representations, we trained recurrent neural networks (RNNs) to perform navigation tasks in 2D arenas based on velocity inputs. Surprisingly, we find that grid-like spatial response patterns emerge in trained networks, along with units that exhibit other spatial correlates, including border cells and band-like cells. All these different functional types of neurons have been observed experimentally. The order of the emergence of grid-like and border cells is also consistent with observations from developmental studies. Together, our results suggest that grid cells, border cells and others as observed in EC may be a natural solution for representing space efficiently given the predominant recurrent connections in the neural circuits.
Metadata: arXiv. Abstracts retain their original authors’ rights.
Original HN comments, with labels from simple text rules. A label describes the whole comment and may miss context or irony. Read the discussion to make up your own mind. About the labels
5 selected excerpts · 5 total mentions
You might enjoy this paper[0] which shows that recurrent position encodings recover grid cell representations and maps to path integration found in a popular model of the hippocampus. This isn't terribly surprising since RNNs have shown this before[1, 2] but its an interesting connection. [0] https://arxiv.org/abs/2112.04035 [1] https://arxiv.org/abs/1803.07770 [2] https://www.nature.com/articles/s41586-018-0102-6
Neuroscience is not my specialty, I only understand the ML side of it. My understanding is that when our machine learning architectures are specially crafted to model specific types of data, we observe similar neurons as those in the brain because they converge to the same optimal solution to handle this data. For example, taking into account the translational invariance of image data with convolutional layers leads to similar neurons as the V1 cortex. Training recurrent layers on navigation tasks lead to grid-cells similar to those in the brain. Some references: https://arxiv.org/abs/1803.07770 https://hal.inria.fr/hal-01389809/document https://www.biorxiv.org/content/10.1101/296301v1.full
In addition, multiple works[0][1] have discovered that grid cell representations arise from regularized recurrent networks when provided relative inputs to predict absolute outputs. [0] https://arxiv.org/abs/1803.07770 [1] https://deepmind.com/blog/grid-cells/
This DeepMind work has been gathering a lot of hype, but it is very similar to this work by Cueva and Wei: Published on Arxiv on March 31: https://arxiv.org/abs/1803.07770 Published at ICLR 2018, a well-known machine learning conference: https://openreview.net/forum?id=B17JTOe0- It is rather disheartening that they didn't acknowledge the -- seems like a deliberate oversight. The Cueva and Wei paper is well written, explains concepts clearly, without the hype, and published in an open access venue.
I think you're referring to this paper: "Emergence of grid-like representations by training recurrent neural networks to perform spatial localization". https://arxiv.org/abs/1803.07770 It appears to be from Columbia vs DeepMind with different authors.
More from the same topics.
Tom B. Brown, Benjamin Mann, Nick Ryder +28
Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task.
Chiyuan Zhang, Samy Bengio, Moritz Hardt +2
Despite their massive size, successful deep artificial neural networks can exhibit a remarkably small difference between training and test performance.
Scott Aaronson
In honor of Alan Turing's hundredth birthday, I unwisely set out some thoughts about one of Turing's obsessions throughout his life, the question of physics and free will.
Mariusz Bojarski, Davide Del Testa, Daniel Dworakowski +10
We trained a convolutional neural network (CNN) to map raw pixels from a single front-facing camera directly to steering commands.