Study investigates graph recovery in triangular random linear cocycles, indicating new extension classes support it.
We study triangular random linear cocycles whose invariant reducing graphs need not be measurable from the available observation filtration. Observation-adapted extension classes quantify this obstruction, while past and future Green operators, bounded cohomological resolvents, Markov compression, and partial Gaussian observations determine when and how the graph can be recovered.
No takes yet. Share an insight, caveat, or question.
Hiroyuki Shioiri (2026) studied this question.