Near-optimal coherent state discrimination via continuously labelled non-Gaussian measurements

This paper demonstrates that near-optimal discrimination of coherent states, surpassing the traditional Gaussian limit and approaching the Helstrom bound, can be achieved using continuously labelled non-Gaussian measurements via non-Gaussian unitary operations with homodyne detection or orthogonal polynomials, thereby eliminating the necessity for discrete photon detection.

The Gist

Imagine you are trying to send a secret message using light. In the quantum world, you can't just turn a light switch on or off; instead, you send "coherent states" of light, which are like very specific, delicate ripples in a pond. Your goal is to tell the difference between two slightly different ripples: one representing a "0" and one representing a "1."

The problem? These ripples are so similar that they overlap. In the quantum world, you can never be 100% sure which one you received without making a mistake. There is a theoretical "perfect score" called the Helstrom bound—the absolute best anyone could possibly do.

The Old Way: The "Pixel" vs. The "Wave"

For a long time, scientists had two main ways to try to read these messages:

1. The "Pixel" Approach (Photon Detection): This is like taking a photo and counting the individual pixels (photons). It's very good at distinguishing the ripples, but it's hard to build the cameras needed to do it perfectly. It's the "gold standard" but expensive and finicky.
2. The "Wave" Approach (Homodyne Detection): This is like listening to the sound of the water. It's easy to do and robust, but it's not very precise. It hits a "ceiling" of accuracy called the Gaussian limit. Think of this as trying to guess the difference between two whispers by listening to the general noise; you'll get it wrong often.

For years, the consensus was: "If you want to beat the 'Wave' ceiling and get close to the 'Perfect Score,' you must use the 'Pixel' approach (counting photons)."

The New Discovery: A New Kind of "Wave"

This paper says: "Not so fast!"

The authors, James Moran, Spiros Kechrimparis, and Hyukjoon Kwon, discovered that you don't need to count individual photons to get a near-perfect score. You can still use the "Wave" approach (continuous measurements), but you have to twist the wave in a very specific, non-standard way before you listen to it.

They call these Continuously Labelled Non-Gaussian Measurements. That's a mouthful, so let's break it down with an analogy.

The Analogy: The Shapeshifting Mirror

Imagine you are trying to tell the difference between two very similar-looking twins (the two light states).

• The Old Way (Homodyne): You look at them through a standard, flat mirror. They look almost identical. You guess wrong a lot.
• The "Pixel" Way: You put them under a microscope that counts their atoms. You can tell them apart perfectly, but the microscope is huge and expensive.
• The New Way (This Paper): You put the twins in front of a shapeshifting mirror (a non-Gaussian unitary operation) before you look at them. This mirror doesn't count atoms; it just warps their shapes.
  • Twin A gets stretched into a long, thin noodle.
  • Twin B gets squashed into a fat, round ball.
  • Now, when you look at them with your standard mirror (homodyne detection), the difference is huge! You can tell them apart easily without needing the expensive microscope.

Two Magic Tricks They Invented

The paper proposes two specific ways to build these "shapeshifting mirrors":

1. The "Cat State" Rotation: They use a mathematical trick involving "cat states" (quantum states that are like Schrödinger's cat being both alive and dead at once). By rotating the light through this weird state, they stretch the differences between the two messages so much that a simple detector can see them clearly.
2. The "Polynomial" Filter: They use a mathematical concept called Orthogonal Polynomials (specifically Legendre and Laguerre polynomials). Imagine these as a special set of sieves or filters. Instead of just looking at the light, they pass it through a filter that rearranges the light's pattern based on these complex math rules. This rearrangement makes the two messages look completely different to the detector.

Why This Matters

• No "Pixel" Counting Needed: You don't need the difficult, expensive photon-counting technology to get high performance. You can stick with the easier, continuous wave measurements.
• Better than the Old "Pixel" Methods: In some energy ranges (when the light isn't too bright or too dim), their new methods actually perform better than the famous "Kennedy Receiver," which is the current champion of photon-counting methods.
• The "Non-Gaussian" Secret: The paper also investigates why this works. They found that the "magic" comes from non-Gaussianity.
  • Analogy: A "Gaussian" shape is a perfect bell curve (like a normal distribution of heights). A "Non-Gaussian" shape is weird, spiky, or has holes in it.
  • They discovered that you need these "weird, spiky" shapes to break the ceiling. However, they also found that just being "weird" isn't enough. If you make the shape too weird (like using a Cubic Phase Gate), it actually makes things worse. It's a Goldilocks situation: you need the right kind of weirdness.

The Bottom Line

This paper is like finding a new way to tune a radio. Everyone thought you needed a super-expensive, digital tuner (photon counting) to get crystal clear sound. These researchers showed that you can get almost the same clarity with a standard analog tuner (homodyne detection), as long as you tweak the knobs (apply non-Gaussian operations) in a very specific, clever way.

It opens the door to building better, cheaper, and more robust quantum communication systems without needing the most cutting-edge, difficult-to-build hardware.

Technical Summary

Here is a detailed technical summary of the paper "Near-optimal coherent state discrimination via continuously labelled non-Gaussian measurements."

1. Problem Statement

Quantum state discrimination is a fundamental task in quantum information, aiming to distinguish between non-orthogonal quantum states with the minimum possible error rate. For optical coherent states ($|\alpha\rangle$ and $|-\alpha\rangle$), the theoretical minimum error rate is defined by the Helstrom bound.

However, achieving this bound experimentally is challenging.

• Gaussian Limit: Standard measurements like homodyne detection (measuring quadrature operators) are limited by the "Gaussian limit," which results in error rates significantly higher than the Helstrom bound, especially at low to moderate energies.
• Current Solutions: Known near-optimal and optimal receivers (e.g., Dolinar, Sasaki-Hirota, Kennedy receivers) typically rely on photon detection (on-off detection), which is a discretely labelled measurement.
• The Gap: It was an open question whether continuously labelled measurements (which produce continuous outcomes, like homodyne detection) could surpass the Gaussian limit without resorting to photon detection. While a specific scheme involving a Kerr gate and homodyne detection was known to exceed the Gaussian limit, a general framework for continuously labelled non-Gaussian discrimination was lacking.

2. Methodology

The authors propose and analyze two distinct types of continuously labelled non-Gaussian measurement schemes that do not require photon counting. They categorize these into Type A and Type B receivers:

Type A: Non-Gaussian Unitary Preprocessing + Gaussian Detection

This approach involves applying a non-Gaussian unitary operation ($\hat{U}_{NG}$) to the input coherent states before performing a standard Gaussian measurement (specifically homodyne detection).

• Mechanism: The coherent states $|\pm\alpha\rangle$ are transformed into non-Gaussian states via $\hat{U}_{NG} |\pm\alpha\rangle$, which are then measured via homodyne detection.
• Specific Implementations: The authors investigate unitary rotations generated by projectors onto specific states:
  1. Fock State Rotations: Generalized SNAP (Selective Number-dependent Arbitrary Phase) gates.
  2. Cat State Rotations: Rotations based on even/odd Schrödinger cat states.
  3. Coherent State Rotations: Rotations based on displaced coherent states.
• Optimization: The rotation angles and parameters (e.g., displacement $\beta$) are numerically optimized for specific signal intensities $|\alpha|^2$ to maximize the total variation distance between the resulting probability distributions.

Type B: Direct Continuously Labelled Non-Gaussian Measurements

This approach constructs measurement operators directly from families of orthogonal polynomials, without relying on a preceding unitary transformation of a Gaussian state.

• Mechanism: The measurement basis is defined by states $|s\rangle$ constructed from the coefficients of orthogonal polynomials in the Fock basis.
• Specific Implementations:
  1. Legendre Polynomial States: Defined on the interval $s \in [-1, 1]$.
  2. Laguerre Polynomial States: Defined on the interval $r \in [0, \infty)$.
• Theoretical Basis: These states satisfy completeness relations derived from the orthogonality of the respective polynomials, allowing them to form valid Positive Operator-Valued Measures (POVMs).

Quantification of Non-Gaussianity

To characterize the resources required, the authors utilize the Stellar Rank ($r^*$), a measure of non-Gaussianity based on the number of zeros in the stellar function of a state.

• Gaussian measurements have $r^* = 0$.
• Discretely labelled photon detection has $r^* = \infty$.
• The authors analyze the stellar rank of their proposed schemes to determine the "cost" of non-Gaussianity.

3. Key Contributions

1. Demonstration of Near-Optimal Performance without Photon Detection:
   The paper proves that photon detection is not strictly necessary to approach the Helstrom bound. Both Type A and Type B schemes achieve error rates significantly lower than the Gaussian limit (homodyne detection) and, in certain energy regimes, outperform the Kennedy receiver (a standard photon-detection-based near-optimal scheme).

2. Two Novel Receiver Architectures:

   • Unitary Rotations (Type A): Showed that rotating coherent states by cat states or coherent states (using SNAP gates) followed by homodyne detection yields near-optimal discrimination.
   • Orthogonal Polynomials (Type B): Introduced a new class of receivers based on Legendre and Laguerre polynomials, demonstrating that mathematical constructs from orthogonal polynomial theory can be directly mapped to quantum measurement bases for state discrimination.

3. Quantification of Non-Gaussianity:
   The authors provide a nuanced view of non-Gaussianity. They show that while non-Gaussianity is necessary to beat the Gaussian limit, it is not sufficient.

   • They demonstrate that some non-Gaussian measurements (e.g., displaced Fock states with finite stellar rank, or a single Cubic Phase Gate) actually perform worse than the Gaussian limit.
   • They establish that all near-optimal schemes identified in this work (and known optimal ones) possess an infinite stellar rank, suggesting a high degree of non-Gaussian complexity is required for optimality.

4. Experimental Feasibility:
   The paper highlights that the Coherent State Rotation scheme can be decomposed into a single SNAP gate and displacement operations. Since SNAP gates have been experimentally realized in cavity QED systems, the authors argue that their proposed Type A receiver is experimentally viable with current technology.

4. Results

• Error Rates:
  • In the low-energy regime ($0 < |\alpha|^2 < 0.47$), the Cat State Rotation receiver performs best among the continuously labelled schemes, closely approaching the Helstrom bound and outperforming the Kennedy receiver.
  • In the moderate-to-high energy regime, the Coherent State Rotation and Legendre Polynomial receivers maintain a significant advantage over homodyne detection.
  • Specifically, the Legendre-based receiver yields the second-largest improvement over the Gaussian limit in the high-energy limit, outperforming coherent state rotations in the range $2.2 < |\alpha|^2 < 2.7$.
• Comparison with Photon Detection:
  The continuously labelled schemes maintain an advantage over the Kennedy receiver (photon detection based) for a moderate range of coherent state amplitudes, particularly in the low-energy limit where photon detection schemes often struggle with higher error rates.
• Non-Gaussianity Analysis:
  • Displaced Fock States (Finite $r^*$): Perform worse than homodyne detection.
  • Cubic Phase Gate (Infinite $r^*$): A single CPG followed by homodyne detection also degrades performance compared to homodyne alone.
  • Conclusion: Simply having non-Gaussianity (even infinite stellar rank) does not guarantee improved discrimination; the specific form of the non-Gaussian operation is critical.

5. Significance

• Theoretical Breakthrough: The work challenges the prevailing assumption that photon counting is essential for optimal coherent state discrimination. It expands the landscape of viable quantum receivers to include purely continuous-variable, non-Gaussian strategies.
• Experimental Pathway: By linking near-optimal performance to SNAP gates (which are experimentally realizable), the paper provides a concrete roadmap for building high-performance optical receivers without the complexity of real-time feedback loops required by the Dolinar receiver or the inefficiencies of on-off photon detectors.
• Resource Theory: The application of Stellar Rank to measurement schemes offers a new framework for classifying quantum measurements. It highlights that "more non-Gaussian" is not always "better," and that specific structural properties (like those found in cat states or Legendre polynomials) are required to unlock the Helstrom bound.
• Future Directions: The paper opens avenues for exploring other families of orthogonal polynomials (e.g., Askey scheme) and investigating whether purely continuously labelled measurements can ever exactly saturate the Helstrom bound, a question that remains open.

In summary, this paper establishes that continuously labelled non-Gaussian measurements are a powerful and experimentally feasible alternative to photon detection for near-optimal quantum state discrimination, offering a new paradigm for optical quantum communication protocols.