2
$\begingroup$

According to "A Large-Scale Model of the Functioning Brain" by Eliasmith et al., Spaun 1.0 was able to perform 8 distinct cognitive tasks:

  1. Copy Drawing: Given a randomly chosen handwritten digit, Spaun should produce the same digit written in the same style as the handwriting.

  2. Digit recognition: Given a randomly chosen handwritten digit, Spaun should produce the same digit written in its default writing.

  3. N-arm bandit task: Spaun should perform a three-armed bandit task, in which it must determine which of three possible choices generates the greatest stochastically generated reward. Reward contingencies can change from trial to trial.

  4. List memory: Given a list of any length, Spaun should reproduce it.

  5. Counting: Counting. Given a starting value and a count value, Spaun should write the final value (that is, the sum of the two values)

  6. Question answering: Given a list of numbers, Spaun should answer either one of two possible questions:

    i. What is in a given position in the list?

    ii. Given a kind of number, at what position is this number in the list?

  7. Rapid variable creation Given example syntactic input/output patterns (e.g., 0 0 7 4 → 7 4; 0 0 2 4 → 2 4; etc.), Spaun should complete a novel pattern given only the input (e.g., 0 0 1 4 → ?)

  8. Fluid Induction Spaun should perform a syntactic or semantic reasoning task that is isomorphic to the induction problems from the Raven’s Progressive Matrices (RPM) test for fluid intelligence

In "Spaun 2.0: Extending the World’s Largest Functional Brain Model" by Xuan Choo, it is written:

The final Spaun 2.0 model consists of approximately 6.6 million neurons, can perform 12 cognitive tasks, and has been demonstrated to reproduce behavioural and neurological data observed in natural cognitive agents.

What novel cognitive tasks does Spaun 2.0 add? Reading the thesis, I've only been able to identify:

  1. Adaptive arm control: Given a force field applied to Spaun's arm, Spaun should adapt to the force and continue to be able to draw in spite of the changing environment dynamics.

  2. Stimulus matching task: Given an example image from ImageNet, match other images of the same categories.

Hypothetically, instruction following could be considered as a task, but it seems like more of a meta-task, since it defines how Spaun can combine tasks to create new tasks without any architecture changes.

$\endgroup$
  • $\begingroup$ I understand if this question looks like shameless self-promotion, as this research comes from my lab. However, I keep honestly having to look this up and this seemed like a good place as any to post it publicly. $\endgroup$ – Seanny123 Aug 2 '18 at 20:21
1
$\begingroup$

Counting task numbers in Spaun becomes somewhat meaningless due to it's instruction following capabilities. However, the missing task was the stimulus response task, wherein given an image from ImageNet, classify it according to it's given identifier.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.