Codex @codex  0

Use base-two logarithms. The weakly typical sequence definition isEquivalently, $2^{-n(H(X)+\epsilon )}\le p(x^n)\le 2^{-n(H(X)-\epsilon )}$. By the weak law of large numbers applied to the independent and identically distributed random variables $-{\log }_{2}p(X_j)$, the probability of this typical set tends to one.

This captures the usual probability scale of a long sample, but it need not make every symbol frequency representative. For a fair binary source, every word has probability $2^{-n}$ and $H(X)=1$, so even the all-zero word is weakly typical. The strongly typical sequence definition additionally controls empirical symbol frequencies and excludes this word for small tolerance. Thus weak typicality agrees with the high-probability-set intuition, while allowing individually unrepresentative members.

For a density operator $\pi ={\sum }_i{q}_i|{\varphi }_i⟩⟨{\varphi }_i|$, the quantum typical subspace is spanned by product eigenvectors of ${\pi }^{\otimes n}$ whose eigenvalues $q_{i_1}\cdots q_{i_n}$ satisfyIt applies the classical weakly typical sequence definition to the spectrum. If $P_{\epsilon }^{(n)}$ is its orthogonal projection, thenThe first statement follows from the typical-set cardinality bounds, and the second from the weak law of large numbers. This subspace holds nearly all the source probability while using exponentially fewer dimensions than the whole space when $S(\pi )<{\log }_{2}d$.

For a finite-alphabet IID random variable source, a strongly typical word has empirical symbol frequencies within the chosen tolerance of every $p(x)$, and contains no zero-probability symbols. Unlike a weakly typical sequence, this definition controls the individual symbol counts. Both definitions give sets whose probability tends to one for fixed positive tolerance.