Are cognitive maps something which is new?



In a paper [1], a robot control system was introduced which is able to navigate through obstacles. It is working by reasoning about the input sensor and then a decision is made, according to the rules stored in the cognitive map. An unkind picture of the input-output relations should make the reader believe, that the system is working great. But it seems, that the idea has a more serious bug as well. The idea of using a concept map for storing the behavior is nothing which is new, but it was researched in quantum physics before. So called “quantum baker maps” were introduced in the past.[2] They are simple geometrical functions to store relationships between chaotic signals.

My question is: why was the paper [1] published in the year 2010, if it doesn't contain a new idea, but it's repeating existing scholarly knowledge? Is this equal to a low quality publication?

[1] MENDONÇA, MÁRCIO, LÚCIA VALÉRIA RAMOS DE ARRUDA, and FLÁVIO NEVES JÚNIOR. "Qualitative Autonomous Navigation System Employing Event Driven-Fuzzy Cognitive Maps and Fuzzy Logic." CONGRESSO BRASILEIRO DE AUTOMÁTICA. 2010.

[2] Scholarpedia: Quantized baker map,

Manuel Rodriguez

Posted 2019-10-02T17:40:44.640

Reputation: 1



There are at least these reasons behind (without going too much in detail what is new knowledge and which not in this particular case):

  1. Science in level of papers and publications does not always (nor any time?) go with giant leaps. There are hundreds or thousands of people researching one big topic and when everyone takes their own little steps, the entire thing takes in long run those big steps, if any. There are a few Einstein's existing but many genious individuals who in crowd can advance high.

  2. That said, the paper is not always or never a complete new reveal or invention (requisite for patent, not for science). To be credible it has to have a solid base where on top you put your little extra. To put that little extra to any context, well, you write down the under lying theory as reference. In short, but still that paper remains as an entity, not just blurry bunch of ideas related to something.

  3. These said, I would say the paper 1 follows normal strategy. To rehearse, their main point is somewhere where they state:

This paper developed an autonomous navigation system based on a new type of fuzzy cognitive maps, named event driven-fuzzy cognitive map, ED-FCM. The developed ED-FCM approach adds new types of relationships and concepts into a classical FCM that allows modeling the human ability of to take decision in presence of random events.

  1. So, a lot talk about Fuzzy this and that, the research point was that little enhancement. You stated the algorithm had a bug. Well, that happens in normal science, too. Imagine how many times some diets have been stated healthy and later again unhealthy and so forth, to get an perspective.

  2. Last point is cross scientific approach. There they did not read quantum physics, so would it not came to my mind neither, but about the theory on conceptual maps they did know their existence and thus provided that knowledge as preliminar part for their work, for reasons stated above.


Posted 2019-10-02T17:40:44.640

Reputation: 1 019

1Patents do not require "complete new reveal" - many are improvements to existing inventions and do not e en need to even be shown to literally be imprisonments in any particular dimension, just useful, new, and not obvious in light of everything that came before. – George White – 2019-10-07T22:25:47.907

OK, that was my mis-conception. – mico – 2019-10-08T09:09:09.457