Yorick Hardy, Willi-Hans Steeb
8 min
In quantum computing, a fundamental task is identifying the unitary operator that transforms a specific initial state into a target state. Because unitary operators preserve the inner product and map orthonormal sets to orthonormal sets, they are essential for describing time evolution and state transformations. This paper introduces a clear, algorithmic approach to finding such an operator, , given two sets of normalized states in a finite-dimensional Hilbert space.
The core of the methodology is a modified Gram-Schmidt orthonormalization process. Given two sets of orthonormal vectors (representing initial and target states), the algorithm extends these into full orthonormal bases for the Hilbert space. By orthogonalizing these sets against an arbitrary third basis, the authors construct the required unitary transformation as a sum of outer products of the basis vectors. This approach is computationally efficient and provides a constructive way to map any state, including entangled ones like Bell or GHZ states, to a corresponding target state.
The authors demonstrate the versatility of this approach by applying it to various scenarios in quantum information science, such as mapping standard basis states to Bell states or transforming between different entangled configurations, including GHZ and W-states. The study also provides an implementation using Symbolic C++, enabling researchers to compute these transformations programmatically. The flexibility of the approach allows for the transformation of unentangled states into entangled states, a critical operation in protocols like quantum teleportation and distributed consensus.
Alex: Yes, and the order matters. The supplied state vectors come first, followed by a complete auxiliary basis. Because the supplied vectors are already orthonormal, they remain unchanged. The later vectors fill only the missing directions.
Sam: Could you walk through the whole recipe without the notation? I want to picture what gets matched to what.
Alex: Start with the input states and complete them into a basis. Separately, start with the target states and complete those into another basis. Then pair every input basis vector with its corresponding output basis vector. The matrix is built by adding the elementary maps for all those pairs.
Sam: The requested state mappings occupy the front of the pairing. The extra pairs define what happens everywhere else.
Alex: That extra information is essential. Mapping one vector tells you very little about the rest of the space. The completed bases specify a transformation on the whole space, while preserving the mappings you requested.
Sam: Why does that make the result unitary rather than simply a matrix that happens to hit the targets?
Alex: Because it carries a complete orthonormal basis to another complete orthonormal basis. The authors prove this directly by multiplying the constructed matrix by its conjugate transpose, the transpose with complex entries conjugated. Orthogonality eliminates the cross terms, and completeness leaves the identity.
Sam: That sounds like a proof for the whole permitted class of inputs, rather than a claim inferred from a few successful examples.
Alex: It is. They also show that applying the matrix to each specified input produces the corresponding target. Those are the two guarantees: the full transformation is unitary, and the requested mappings hold. The worked examples illustrate the proof rather than substitute for it.
Sam: Give us a quantum example where this construction does something concrete, not just changes coordinate labels.
Alex: In a four-dimensional space, the paper starts with a two-qubit product state. A qubit is a quantum system with two basis states; here, both qubits are in the second basis state. The target is a Bell state, an entangled state with equal amplitudes for both qubits being first or both being second.
Sam: The starting state separates into a state for each qubit. The target has to be described jointly.
Alex: The authors explicitly give the unitary matrix connecting them. Under the two-tangle, the entanglement measure they use for this example, the Bell state is completely entangled. So the construction provides a full reversible map from that product state to the Bell state.
Sam: That shows a mathematical transformation can create entanglement. Does the paper also tell us how to implement it efficiently?
Alex: It provides the matrix, not an optimized implementation in elementary quantum gates. There is no runtime comparison or physical experiment. Its contribution is a constructive algebraic method, so we should judge it by the proof and the stated input conditions.
Sam: The paper also connects two named three-qubit states. Does that example make the entanglement story less straightforward?
Alex: It does. In an eight-dimensional space, they construct a transformation from the Greenberger–Horne–Zeilinger state, usually called the GHZ state, to the W state. The first has equal amplitudes at the two opposite ends of the computational basis. The second distributes equal amplitudes across states with one qubit in the second basis state.
Sam: How should listeners interpret the paper calling one of those states unentangled? That sounds stronger than just reporting a measure.
Alex: Keep the qualifier attached. The authors describe the W state as unentangled using the three-tangle, their measure of three-qubit entanglement. They also explicitly say it cannot be written as a product state. So their statement is about that measure, not about the state separating into independent parts.
Sam: Let us return to the construction's freedom. If I specify the same input and target, will everyone obtain the same matrix?
Alex: Generally, no. The auxiliary bases can be chosen differently, changing how the transformation acts outside the specified input subspace. The authors say those choices can generate all unitary operators implementing the desired mapping. Uniqueness arrives only when you have specified complete bases.
Sam: That freedom could be useful, but this particular recipe does not rank the resulting matrices by convenience.
Alex: It selects a result by using the standard basis to complete both sets. The paper also supplies a Symbolic C++ implementation, meaning code for symbolic algebra. One implementation detail deserves care: the displayed code uses ordinary transposes, while the general complex-space proof requires conjugate transposes; the worked inputs are real.
Sam: With those boundaries in mind, who should read the full document, and where should they begin?
Alex: Researchers needing explicit finite-dimensional quantum transformations should start with Section 2, the modified orthonormalization and proof. Section 3 gives the concise construction algorithm. Read Section 5 if you want the implementation, keeping the distinction between real and complex vectors in mind.
Sam: And what should someone outside that immediate use case remember?
Alex: The proof gives an exact mapping under explicit conditions, not a guarantee of efficient physical implementation. The central idea is portable: to turn a partial state mapping into a full reversible transformation, complete both sides into bases.
Sam: Keep the completion step in mind.