ATON-QAI/QAI-Photonic-Noise-and-Gram-Surrogate-v1
Tabular Regression • Updated • 20
Datasets:
Formats:
json
Languages:
English
Size:
< 1K
Tags:
quantum-photonics
linear-optics
hong-ou-mandel
boson-sampling
formal-verification
quantum-computing
License:
benchmark_id stringlengths 10 10 | family stringlengths 8 24 | title stringlengths 45 76 | modes int64 2 6 | photons int64 1 3 | qai_source stringlengths 180 228 | ideal_fock_distribution unknown | physical_noise_profile unknown | verified_invariant stringlengths 60 79 | doi_reference stringlengths 16 27 |
|---|---|---|---|---|---|---|---|---|---|
QAI-PB-001 | HOM_INTERFERENCE | Hong-Ou-Mandel Two-Photon Bunching on Balanced MZI | 2 | 2 | qai 1.0;
profile photonic.dv.linear;
experiment hom_dip(theta: angle = pi / 2) {
let m: modes<2, fock> = prepare fock [1, 1];
let (a, b) = sym_mzi(theta)(m[0], m[1]);
emit measure pnr (a, b);
} | {
"|2,0>": 0.5,
"|0,2>": 0.5,
"|1,1>": 0
} | {
"waveguide_platform": "TFLN_LiNbO3",
"transmittance_eta": 0.92,
"indistinguishability_V": 0.96,
"dark_count_p_dc": 0.001,
"g2_zero": 0.005,
"observed_coincidence_p11": 0.0171
} | Balanced beam-splitter zero-coincidence bosonic bunching identity | 10.1103/PhysRevLett.59.2044 |
QAI-PB-002 | CLEMENTS_RECK_MZI | Universal 4-Mode Clements Interferometer Decomposition | 4 | 2 | qai 1.0;
profile photonic.dv.linear;
experiment clements_4mode(theta: angle = pi / 3) {
let m: modes<4, fock> = prepare fock [1, 1, 0, 0];
let out = clements_mesh<4>(theta)(m);
emit measure pnr out;
} | {
"|2,0,0,0>": 0.1875,
"|0,2,0,0>": 0.1875,
"|1,1,0,0>": 0.25,
"|0,0,1,1>": 0.375
} | {
"waveguide_platform": "Si3N4_LowLoss",
"transmittance_eta": 0.9500000000000001,
"indistinguishability_V": 0.97,
"dark_count_p_dc": 0.0005,
"g2_zero": 0.003,
"observed_coincidence_p11": 0.23140000000000002
} | SU(M) interferometer unitary preservation and Fock simplex normalization | 10.1364/OPTICA.3.001460 |
QAI-PB-003 | BELL_CHSH | Dual-Rail Photonic Bell State & CHSH Correlation Witness | 4 | 2 | qai 1.0;
profile photonic.dv.linear;
experiment bell_phi_plus() {
let q: modes<4, fock> = prepare fock [1, 0, 1, 0];
let out = postselected_bell_pair(q);
emit measure pnr out;
} | {
"|1,0,1,0>": 0.5,
"|0,1,0,1>": 0.5
} | {
"waveguide_platform": "TFLN_LiNbO3",
"transmittance_eta": 0.91,
"indistinguishability_V": 0.9500000000000001,
"dark_count_p_dc": 0.001,
"g2_zero": 0.004,
"chsh_s_value": 2.684
} | Tsirelson CHSH bound (2 < S <= 2*sqrt(2)) under postselected dual-rail encoding | 10.1103/PhysRevLett.23.880 |
QAI-PB-004 | GHZ_ENTANGLEMENT | 6-Mode Herald-Postselected Photonic GHZ State | 6 | 3 | qai 1.0;
profile photonic.dv.linear;
experiment ghz_3photon() {
let m: modes<6, fock> = prepare fock [1, 0, 1, 0, 1, 0];
let g = linear_optical_ghz6(m);
emit measure pnr g;
} | {
"|1,0,1,0,1,0>": 0.5,
"|0,1,0,1,0,1>": 0.5
} | {
"waveguide_platform": "Si3N4_LowLoss",
"transmittance_eta": 0.9400000000000001,
"indistinguishability_V": 0.96,
"dark_count_p_dc": 0.0008,
"g2_zero": 0.005,
"ghz_fidelity": 0.912
} | Three-photon Fock-number conservation and parity coherence bound | 10.1103/PhysRevLett.82.1345 |
QAI-PB-005 | BB84_QKD | Photonic Polarization/Dual-Rail BB84 Quantum Key Distribution Sifting | 2 | 1 | qai 1.0;
profile photonic.dv.linear;
experiment bb84_basis_match(basis_delta: angle = 0.0) {
let m: modes<2, fock> = prepare fock [1, 0];
let out = sym_mzi(basis_delta)(m[0], m[1]);
emit measure pnr out;
} | {
"|1,0>": 1,
"|0,1>": 0
} | {
"waveguide_platform": "Telecom_Fiber_1550nm",
"transmittance_eta": 0.85,
"indistinguishability_V": 0.99,
"dark_count_p_dc": 0.0005,
"g2_zero": 0.002,
"qber_observed": 0.0034000000000000002
} | Matched-basis zero intrinsic QBER and dark-count bounded error rate | 10.1016/j.tcs.2014.05.025 |
QAI-PB-006 | GROVER_OPTICAL | 4-Mode Linear-Optical Single-Step Grover Search Oracle | 4 | 1 | qai 1.0;
profile photonic.dv.linear;
experiment grover_4mode(marked_mode: count = 2) {
let m: modes<4, fock> = prepare fock [1, 0, 0, 0];
let out = optical_grover4(marked_mode)(m);
emit measure pnr out;
} | {
"|0,0,1,0>": 1
} | {
"waveguide_platform": "Si3N4_LowLoss",
"transmittance_eta": 0.93,
"indistinguishability_V": 0.98,
"dark_count_p_dc": 0.0005,
"g2_zero": 0.002,
"target_mode_selectivity": 0.928
} | N=4 single-step exact amplitude amplification to marked spatial mode | 10.1103/PhysRevLett.79.325 |
QAI-PB-007 | KLM_NS_GATE | KLM Nonlinear Sign Shift (NS-Gate) with Ancilla Herald Conditioning | 3 | 2 | qai 1.0;
profile photonic.dv.linear;
experiment klm_ns_gate() {
let m: modes<3, fock> = prepare fock [2, 1, 0];
let out = klm_nonlinear_sign(m);
emit measure pnr out;
} | {
"herald_success_probability": 0.25,
"conditional_phase_rad": 3.141593
} | {
"waveguide_platform": "TFLN_LiNbO3",
"transmittance_eta": 0.92,
"indistinguishability_V": 0.96,
"dark_count_p_dc": 0.0005,
"g2_zero": 0.003,
"herald_fidelity": 0.9410000000000001
} | Conditional pi phase shift on |2> Fock component under single-ancilla herald | 10.1038/35051009 |
QAI-PB-008 | BOSON_SAMPLING | 6-Mode 3-Photon Aaronson-Arkhipov Boson Sampling Permanent Benchmark | 6 | 3 | qai 1.0;
profile photonic.dv.linear;
experiment boson_sampling_6x3() {
let m: modes<6, fock> = prepare fock [1, 1, 1, 0, 0, 0];
let out = haar_clements_6(m);
emit measure pnr out;
} | {
"fock_hilbert_dimension": 56,
"sum_of_permanents_l1": 1
} | {
"waveguide_platform": "Integrated_6Mode_Mesh",
"transmittance_eta": 0.9,
"indistinguishability_V": 0.9450000000000001,
"dark_count_p_dc": 0.001,
"g2_zero": 0.006,
"tvd_ideal_vs_noisy": 0.048
} | Submatrix permanent L1 normalization across 56-state 3-photon Fock subspace | 10.4086/toc.2013.v009a004 |
QAI-PB-009 | VIBRONIC_DUSCHINSKY | Molecular Vibronic Franck-Condon Spectrum via Duschinsky Rotation | 4 | 2 | qai 1.0;
profile photonic.dv.linear;
experiment vibronic_so2() {
let m: modes<4, fock> = prepare fock [1, 1, 0, 0];
let out = duschinsky_interferometer<4>(m);
emit measure pnr out;
} | {
"|1,1,0,0>": 0.41200000000000003,
"|2,0,0,0>": 0.218,
"|0,1,1,0>": 0.245,
"|0,0,2,0>": 0.125
} | {
"waveguide_platform": "Si3N4_LowLoss",
"transmittance_eta": 0.93,
"indistinguishability_V": 0.96,
"dark_count_p_dc": 0.0005,
"g2_zero": 0.004,
"spectral_overlap_fidelity": 0.964
} | Orthogonal Duschinsky normal-mode rotation and Franck-Condon sum rule | 10.1038/nphoton.2015.153 |
QAI-PB-010 | PHOTONIC_VQE_H2 | Variational Quantum Eigensolver for H2 Bond Dissociation on 4 Photonic Modes | 4 | 2 | qai 1.0;
profile photonic.dv.linear;
experiment vqe_h2_ansatz(theta: angle = 0.226) {
let m: modes<4, fock> = prepare fock [1, 0, 1, 0];
let out = uccsd_photonic_ansatz(theta)(m);
emit measure pnr out;
} | {
"equilibrium_bond_angstrom": 0.7414000000000001,
"exact_ground_energy_hartree": -1.13727
} | {
"waveguide_platform": "TFLN_LiNbO3",
"transmittance_eta": 0.9400000000000001,
"indistinguishability_V": 0.97,
"dark_count_p_dc": 0.0005,
"g2_zero": 0.003,
"noisy_energy_hartree": -1.1348
} | Rayleigh-Ritz variational energy lower bound E(theta) >= E_0 | 10.1038/ncomms5213 |
QAI-PB-011 | QUANTUM_KERNEL_QML | Photonic Fock-Space Quantum Kernel Gram Matrix Estimator | 4 | 2 | qai 1.0;
profile photonic.dv.linear;
experiment photonic_kernel(x1: angle = 0.4, x2: angle = 0.9) {
let m: modes<4, fock> = prepare fock [1, 1, 0, 0];
let out = feature_map_adjoint(x1, x2)(m);
emit measure pnr out;
} | {
"kernel_self_overlap": 1,
"psd_gram_eigenvalue_min": 0
} | {
"waveguide_platform": "Integrated_4Mode_Mesh",
"transmittance_eta": 0.91,
"indistinguishability_V": 0.9500000000000001,
"dark_count_p_dc": 0.001,
"g2_zero": 0.005,
"psd_preserved": true
} | Positive semi-definiteness and unit diagonal of multiphoton Fock kernel matrix | 10.1038/s41586-019-0980-2 |
QAI-PB-012 | PARAMETER_SHIFT_GRADIENT | Exact Photonic Parameter-Shift Gradient Rule (+/- pi/2 Phase Shift) | 2 | 1 | qai 1.0;
profile photonic.dv.linear;
experiment param_shift_grad(theta: angle = 0.785398) {
let m: modes<2, fock> = prepare fock [1, 0];
let out = sym_mzi(theta)(m[0], m[1]);
emit measure pnr out;
} | {
"analytic_gradient_dz_dtheta": -1,
"parameter_shift_estimate": -1
} | {
"waveguide_platform": "TFLN_LiNbO3",
"transmittance_eta": 0.9500000000000001,
"indistinguishability_V": 0.98,
"dark_count_p_dc": 0.0005,
"g2_zero": 0.002,
"noisy_gradient_estimate": -0.9490000000000001
} | Exact trigonometric identity for +/- pi/2 phase-shifter derivative | 10.1103/PhysRevA.98.032309 |
ATON QAI — Photonic Benchmark & Fock-Space Reference Corpus
Organization: ATON QAI (huggingface.co/ATON-QAI)
Website: https://aton-qai.ai
License: Apache-2.0
Dataset Summary
QAI-Photonic-Benchmark is an open reference dataset by ATON QAI, pairing canonical .qai linear-optical quantum circuits across 12 foundational photonic algorithm families with:
- Exact lossless Fock-space probability distributions derived from permanent amplitudes;
- Parameterized physical waveguide noise profiles for thin-film lithium niobate (
TFLN LiNbO3), low-loss silicon nitride (Si3N4), and telecom fiber platforms ($\eta, V_{\text{HOM}}, p_{\text{dc}}, g^{(2)}(0)$); - Formal mathematical invariant specifications and peer-reviewed literature DOIs.
Quick Load with Hugging Face datasets
from datasets import load_dataset
ds = load_dataset("ATON-QAI/QAI-Photonic-Benchmark", split="test")
for row in ds:
print(row["benchmark_id"], row["family"], row["verified_invariant"])
Schema Fields
| Field | Type | Description |
|---|---|---|
benchmark_id |
string |
Public benchmark identifier (QAI-PB-001 .. QAI-PB-012). |
family |
string |
Photonic algorithm family (HOM_INTERFERENCE, CLEMENTS_RECK_MZI, BELL_CHSH, GHZ_ENTANGLEMENT, BB84_QKD, GROVER_OPTICAL, KLM_NS_GATE, BOSON_SAMPLING, VIBRONIC_DUSCHINSKY, PHOTONIC_VQE_H2, QUANTUM_KERNEL_QML, PARAMETER_SHIFT_GRADIENT). |
modes |
int |
Number of optical waveguide modes $M$. |
photons |
int |
Input Fock photon number $N$. |
qai_source |
string |
Reference .qai circuit specification. |
ideal_fock_distribution |
dict |
Exact lossless Fock-state probability distribution. |
physical_noise_profile |
dict |
Waveguide & detector parameters ($\eta, V, p_{\text{dc}}, g^{(2)}(0)$) and resulting physical observables. |
verified_invariant |
string |
Mathematical property verified for the circuit family. |
doi_reference |
string |
Primary peer-reviewed physical reference DOI. |
Citation (BibTeX)
@dataset{aton_qai_photonic_benchmark_2026,
author = {{ATON QAI}},
title = {{ATON QAI Photonic Benchmark: Linear-Optical Circuits, Exact Fock-Space Distributions, and Physical Noise Profiles}},
year = {2026},
publisher = {Hugging Face / ATON QAI},
url = {https://huggingface.co/datasets/ATON-QAI/QAI-Photonic-Benchmark}
}
- Downloads last month
- -
Models trained or fine-tuned on ATON-QAI/QAI-Photonic-Benchmark