Existence of product vectors and their partial conjugates in a pair of spaces

Young‐Hoon Kiem, Seung-Hyeok Kye, Jungseob LeeJournal of Mathematical Physics 2011 · DOI

TL;DR

This paper investigates the conditions under which a pair of subspaces in a tensor product space must contain a "product vector" whose partial conjugate also lies within one of the subspaces. The findings provide crucial upper bounds for the ranks of certain types of entangled quantum states, offering a deeper understanding of entanglement structure.

What Problem Does This Paper Address?

The paper delves into the intricate world of quantum entanglement, specifically focusing on "positive partial transpose (PPT) entangled states" (PPTES). A key challenge in understanding entanglement is characterizing states that are "separable" (can be expressed as a convex combination of product states) versus those that are "entangled." While the PPT criterion provides a necessary condition for separability (if a state has a positive partial transpose, it's likely separable), it's not sufficient. This means there exist entangled states that still satisfy the PPT criterion (PPTES).

A crucial aspect of PPTES is the concept of "edge states," which are PPTES that lie on the boundary of the set of separable states in a specific way. Understanding these edge states is vital for a complete picture of entanglement. The paper tackles a fundamental question related to edge states: for a given pair of subspaces $D$ and $E$ within a tensor product space, when is it guaranteed that there exists a product vector $x \otimes y$ such that $x \otimes y \in D$ and its partial conjugate $\bar{x} \otimes y \in E$? This question is directly linked to the existence and classification of edge states.

Methods

The authors employ a sophisticated blend of algebraic geometry and linear algebra. The core of their approach involves translating the existence of a product vector with a specific property into a geometric condition related to the intersection of subspaces in a projective space.

Specifically, they leverage the fact that the set of product vectors in a tensor product space can be represented as the image of a Segre map. The problem of finding a product vector $x \otimes y \in D$ with $\bar{x} \otimes y \in E$ is reformulated as finding a non-zero intersection between the projected subspace $PD$ and the projected subspace $PE$ in a relevant projective space.

They utilize intersection theory, particularly Bertini's theorem, to analyze the conditions under which such intersections are guaranteed. This involves working with cohomology rings of projective spaces and analyzing polynomial conditions derived from these intersections. The paper also introduces and proves several lemmas concerning coefficients of specific polynomials, which are crucial for establishing the precise conditions for the existence of these product vectors.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Bridging Subspaces and Entanglement: The paper establishes a direct link between the geometric properties of subspaces (their dimensions and relationships) and the existence of specific types of entangled quantum states (edge states).
  • The Power of Partial Conjugation: The concept of the "partial conjugate" of a product vector is central. The paper shows that the existence of a product vector whose partial conjugate lies in a specific subspace is a key determinant for classifying entanglement.
  • Dimension is Key: The dimensions of the subspaces ($k$ and $l$, representing co-dimensions) relative to the dimensions of the ambient spaces ($m$ and $n$) are the primary factors determining whether such product vectors exist. The condition $k + l < m + n - 2$ emerges as a critical threshold.
  • Geometric Interpretation of Entanglement: The paper elegantly translates abstract quantum information concepts into concrete geometric problems involving intersections of manifolds in projective spaces. This provides a powerful new lens for studying entanglement.
  • Polynomial Conditions for Existence: The existence of these special product vectors is tied to the non-vanishing of specific polynomials derived from intersection theory. This offers a computational avenue for determining entanglement properties.

Key Results

The paper presents a set of precise conditions for the existence of a product vector $x \otimes y \in D$ such that its partial conjugate $\bar{x} \otimes y \in E$, where $D$ and $E$ are subspaces of $\mathbb{C}^m \otimes \mathbb{C}^n$ with co-dimensions $k$ and $l$ respectively.

  • Guaranteed Existence: If $k + l < m + n - 2$, then such a product vector is guaranteed to exist. This implies that certain types of edge states are always possible under these dimension constraints.
  • Conditional Existence: If $k + l = m + n - 2$, the existence of such a product vector depends on specific coefficients ($C_{k,l}^{m-1}$) of a polynomial related to the intersection theory. This indicates a more nuanced relationship between dimensions and entanglement in this regime.
  • Non-Guaranteed Existence: If $k + l > m + n - 2$, such a product vector is not guaranteed to exist. This provides upper bounds on the complexity of certain entangled states.
  • Upper Bounds on Edge States: These results directly translate into upper bounds for the ranks of edge states. Specifically, if an $m \otimes n$ edge state has type $(p, q)$ (meaning its range and the range of its partial transpose have dimensions $p$ and $q$), then $p+q \le 2mn - m - n + 2$.

Limitations and Future Work

  • Exceptional Cases: The paper highlights "exceptional cases" where $k+l = m+n-2$ and a specific polynomial coefficient is zero. These cases require further investigation, and the paper provides specific conditions for when these exceptional cases arise in $m=2$ and $m=3$ scenarios.
  • Conjecture on Real Semialgebraic Sets: The authors propose a conjecture that the set of pairs of subspaces $(D, E)$ for which a product vector satisfying the condition exists is a "full dimensional real semialgebraic proper subset" when the exceptional conditions hold. Proving this conjecture for general $m, n$ is left as future work.
  • Explicit Examples: While the paper provides conditions for existence, constructing explicit examples of subspaces $(D, E)$ that do not contain such product vectors (especially in the exceptional cases) is a significant part of the work and is explored for specific low-dimensional scenarios.
  • Complete Classification of Edge States: The ultimate goal of understanding edge states is not fully achieved, but this paper provides a critical piece of the puzzle by clarifying the conditions under which certain geometric configurations must exist, which in turn constrains the types of edge states possible.

The paper concludes by suggesting that further work could focus on proving the conjecture and exploring the implications of these findings for a more complete classification of entanglement.

Summary updated

Abstract

Let D and E be subspaces of the tensor product of the m- and n-dimensional complex spaces, with co-dimensions k and ℓ, respectively. In order to give upper bounds for ranks of entangled edge states with positive partial transposes, we show that if k + ℓ &amp;lt; m + n − 2, then there must exist a product vector in D whose partial conjugate lies in E. If k + ℓ = m + n − 2, then such a product vector may or may not exist depending on k and ℓ.
BibTeX
@article{kiem2011existence,
  author = {Young‐Hoon Kiem and Seung-Hyeok Kye and Jungseob Lee},
  title = {{Existence of product vectors and their partial conjugates in a pair of spaces}},
  journal = {Journal of Mathematical Physics},
  year = {2011},
  doi = {10.1063/1.3663835},
}

Loading related content...

Figure