review: Starting Point of Research about associative_memory

[mewmew]

@Chilinot, @chrisking2020 and @slax32, please select the research literature you wish to review, and comment in this issue to tell the others which ones you took, so that we don't do too much redundant work. (I've marked the papers which I've printed and started skimming through with the subtle ROBIN: prefix above)

We should try to conduct a preliminary literature review (basically skimming the main research sources) before handing in the project plan on Monday next week. The Project Proposal Milestone tracks each task we should complete before Monday. Feel free to open new issues, tracking sub-tasks and assign them to the relevant milestone. This way I hope we may keep better track of time.

[Chilinot]

I have now selected the papers to read until monday in the original comment :)

[mewmew]

@Chilinot, perfect! Just finished reading the first paper ("Experimental demonstration of associative memory with memristive neural networks."), and it was very interesting! They managed to create artificial synapses using resistors with microprocessors to store and update the resistive properties over time. Using these they managed to show experimentally the formation of associative memory, as the firing of two previously uncorrelated synapses became correlated after firing together. They also go on the mention the underlying rule behind how associative memory is believed to be formed, namely the Hebbian rule stating

"Neurons that fire together, wire together"

[mewmew]

Lets try to find papers through other search engines than Google scholar. For instance the library may be a good place to start. Looking for other meta-study papers may help us get a good overview of the field, needed to identify the key models of associative memory (see issue #25).

[mewmew]

Closing this issue for now, as the literature review has been completed. Some research papers were skimmed, others read with great care. Even the papers which were not included in the final literature helped give an high-level picture and understanding of what's out there, and to get a grasp of the interdisciplinary field of models for associative memory.

In the future, it may be interesting to refer back to these research papers to gain a deeper understanding across the different fields of research.

[mewmew]

Retrospectively, our starting point of research was heavily based on Google scholar search results, in combination with Google and Wikipedia search results. It would be interesting to find what research literature that was censored by our filter bubbles.

Personally, I was sad to find out that the research search capabilities of the University of Portsmouth's library were not open to non-faculty members or students of UoP. This was a real shame, as those capabilities are astounding in their quality and manage to provide a good complementary service to be used along side of Google scholar. I asked the staff members working at the library of UoP to elaborate on why this was the case, and if it was intentional. Sadly, for the time being, it does not seem like they will update their policy in this regard, thus limiting access to an otherwise wonderful resource.

I have yet to fully discover the capabilities provided in this regard by UU. Will take a deeper look for future research projects.