ASR are tech that helps make transcripts from speech

After a brief survey of current literature, it appears that no standardized benchmarks for ASR on clinical data exists that are widely used. Given the vast resources available from the TalkBank corpus, it is not infeasible to build such a corpus and evaluate the performance of a few commercial ASR systems in its ability to perform such a task.

Despite there not being a single baseline that works to benchmark ASR on clinical datasets, a few different subsets of efforts exists on each component of this front.

associative means that operations can be grouped in any way as long as order is preserved.

That is:

\begin{equation} (AB)C = A(BC) \end{equation}

Intuition:

• \(O\): \(\leq\)
• \(\theta\): \(=\)
• \(\Omega\): \(\geq\)

Definitions:

• \(f(n) = O(g(n)) \implies \exists n_{0}: \forall n > n_0, f(n) \leq c (g(n))\)
• \(f(n) = \Omega(g(n)) \implies \exists n_{0}: \forall n > n_0, f(n) \geq c (g(n))\)
• \(f(n) = \theta(g(n)) \implies \exists n_{0}: \forall n > n_0, f(n) \geq 1 (g(n)), f(n) \leq c (g(n))\)

~

Given functions \(f(n)\) and \(g(n)\), if:

\begin{equation} \lim_{n\to \infty} \left(\frac{f(n)}{g(n)}\right) = 1 \end{equation}

we say that \(f \sim g\).