LearningDense with non-decaying traces

[gkarray]

User story

As a user, I want to be able to run a connection Process with learning enabled (i.e LearningDense) with non-decaying traces.

Conditions of satisfaction

• Works in simulation (floating-pt and bit-approximate-loihi) and hardware (hard_coded).
• Lava interface should accept a 0 decay value and treat it as non-decaying on CPU and Loihi.

Acceptance tests

• Test that we can run a LearningDense Process with non-decaying traces x1, x2, y1, y2 and y3.

Tasks

• [ ] Implement the feature on all backends (4h)
• [ ] Write unit tests (4h)
• [ ] Update documentation (2h)
• [ ] Get 2 reviews (PW, EH), merge into lava-nc/main (4h)
• [ ] Get 2 reviews (PW, EH), merge into frameworks.ai.lava.lava-loihi/main (4h)

[EternalGoldenBraid]

Hey is this implemented? I cannot find examples or tests on this.