Online Learning Algorithms#
Study the approximation carried by each algorithm before treating its output as a gradient estimate. The examples deliberately use different model structures: a matched task would be easier to compare visually, but it would not exercise the structural assumptions that distinguish these estimators.
Note
Read Core Concepts first. The chapters assume that you understand ETP selection and the role of hidden-state recurrences.
Choose an algorithm#
Follow diagonal recurrent traces across a sequence and examine when the approximation can differ from BPTT.
Follow input/output-factorized traces and the contraction that turns those factors into parameter gradients.
Separate local eligibility traces from symmetric or random-feedback learning signals in a recurrent spiking network.
Compare the with-H recurrent rule with the without-H feedforward rule.
Widen a trace over an explicitly sparse recurrent dependency graph.
Select by retained structure#
Chapter |
Trace structure |
Appropriate regime |
Important boundary |
|---|---|---|---|
Parameter-shaped trace with a diagonal hidden-Jacobian approximation |
General recurrent ETP models |
Cross-position recurrence is approximated |
|
Separate input and output factors |
Linear-memory recurrent SNN training |
Factorization error is model dependent |
|
Local trace times a broadcast learning signal |
Recurrent LIF/ALIF networks |
Random feedback is not symmetric feedback |
|
With-H or without-H temporal factor |
Recurrent or feedforward SNNs, respectively |
The two regimes are not interchangeable |
|
Sparse n-step recurrent neighborhood |
Structurally sparse recurrence |
Dense recurrence saturates too early to be informative |
For a first pass, read D-RTRL before pp-prop, then choose the SNN-specific or sparsity-specific chapter that matches the model. No chapter establishes universal equality with BPTT; each states the regime and checks appropriate to its approximation.