Learning Cognitive Map Representations for Navigation by Sensory-Motor Integration

Thoughts this paper might be of some interest

How to transform a mixed flow of sensory and motor information into memory state of self-location and to build map representations of the environment are central questions in the navigation research. Studies in neuroscience have shown that place cells in the hippocampus of the rodent brains form dynamic cognitive representations of locations in the environment. We propose a neural-network model called sensory-motor integration network model (SeMINet) to learn cognitive map representations by integrating sensory and motor information while an agent is exploring a virtual environment. This biologically inspired model consists of a deep neural network representing visual features of the environment, a recurrent network of place units encoding spatial information by sensorimotor integration, and a secondary network to decode the locations of the agent from spatial representations. The recurrent connections between the place units sustain an activity bump in the network without the need of sensory inputs, and the asymmetry in the connections propagates the activity bump in the network, forming a dynamic memory state which matches the motion of the agent. A competitive learning process establishes the association between the sensory representations and the memory state of the place units, and is able to correct the cumulative path-integration errors. The simulation results demonstrate that the network forms neural codes that convey location information of the agent independent of its head direction. The decoding network reliably predicts the location even when the movement is subject to noise. The proposed SeMINet thus provides a brain-inspired neural-network model for cognitive map updated by both self-motion cues and visual cues

http://ir.sia.cn/handle/173321/30329

2 Likes

Link is not working … Can u please attach the pdf itself ?
Thanks

1 Like

Here’s the source link.

1 Like

Unfortunately it’s paywalled …

2 Likes

I’d post it here, but then I would go to jail.

2 Likes

Leaving aside the questionable process of restricting access to scientific publications, be aware the original authors are the actual copyright holders of their published papers so it is perfectly legal for them to share their work with anybody they want and for everyone else to contact authors directly and ask for a copy of their work.

1 Like

Not to drag this out, but that is not entirely true. There are all sorts of complexities in copyright law as well as distribution methods and modalities. This is why on some authors websites (the first place I go when I hit a paywall) they simply provide all of the PDFs of their papers while others, maddeningly, include a link to the paywall. The most useful and informative pubs are published in the most prestigious journals and these are typically the most aggressive, read expensive, to access. This always strikes me as counterproductive because an author (and the greatest percentage of those are professors) wants their work to be widely disseminated and read, which is what leads to their ‘payment’, which is, of course, citations.

The only way for a professor to make money with their research (aside from being an expert witness or getting a patent) is to write a book. I find those the worst sort of paywall, but that is just me.

1 Like