From Compressed Sensing to Deep Learning: Tasks, Structures and Models

Поділитися
Вставка
  • Опубліковано 5 жов 2024

КОМЕНТАРІ • 3

  • @JuliusSmith
    @JuliusSmith 2 роки тому

    In the Xampling Hardware section ~17:00, I would say "multiply" instead of "mixing", and "heterodyning" or "frequency-shifting" instead of "aliasing". Aliasing only occurs on the downsampling n Tp itself. Mixing more commonly means scaling and summing.

  • @davidvincent1932
    @davidvincent1932 Рік тому

    Very cool, thanks for sharing.

  • @vtrandal
    @vtrandal 2 роки тому

    I recall looking at the proof for the Xampling “Union of Sub-Spaces” argument (mentioned at 16:30) and thought it to be deficient insofar as it seems to assume what it is trying to prove.