Definition
Hilbert–Schmidt operator
A Hilbert-space operator whose squared norms on an orthonormal basis have finite sum.
Definition
Let and be Hilbert spaces. A bounded operator is Hilbert–Schmidt if, for one—and hence every— orthonormal basis of ,
The basis-independent quantity
is the Hilbert–Schmidt norm. The Hilbert–Schmidt operators form , the Schatten class. Every Hilbert–Schmidt operator is compact, but a compact operator need not be Hilbert–Schmidt. The condition measures square-summability across orthogonal input directions and is independent of all choices of basis.
Equivalent characterizations
For , the following conditions are equivalent:
- is Hilbert–Schmidt;
- its singular values satisfy ;
- the positive operator has finite trace.
With the trace understood as the basis-independent sum of diagonal coefficients,
These formulas also show that is Hilbert–Schmidt exactly when is.
Hilbert-space and ideal structure
For , the formula
makes a Hilbert space. If and are bounded operators of compatible sizes, then
Consequently, is a two-sided operator ideal. The product of two Hilbert–Schmidt operators is trace class, a fact behind many operator-valued integral formulas.
Examples and near misses
On , the diagonal operator is Hilbert–Schmidt exactly when , and then . Every finite-rank operator is Hilbert–Schmidt.
The diagonal operator with is compact but not Hilbert–Schmidt. Thus convergence of singular values to zero is insufficient; their squares must be summable.
Canonical scope and legacy entry
This general definition is the canonical corpus entry for Hilbert–Schmidt operators. The older Shale-paper Hilbert–Schmidt entry records the same notion only in the specialized setting and is retained for compatibility with links in that project. New general-purpose content should link to the present functional-analysis entry.
References
- Barry Simon, Trace Ideals and Their Applications, 2nd ed., American Mathematical Society, 2005. DOI record. Relevant: Chapters 1–2 on Hilbert–Schmidt operators, trace ideals, and ideal inequalities.
- Rajendra Bhatia, Matrix Analysis, Springer, 1997. DOI record. Relevant: Chapter IV on unitarily invariant norms and Schatten classes.