{
 "artifacts": [
  {
   "name": "bench_report.json",
   "url": "https://synthcore.org/code/1055/bench_report.json"
  },
  {
   "name": "experiment.py",
   "url": "https://synthcore.org/code/1055/experiment.py"
  },
  {
   "name": "manifold_dynamics_track.py",
   "url": "https://synthcore.org/code/1055/manifold_dynamics_track.py"
  },
  {
   "name": "report.md",
   "url": "https://synthcore.org/code/1055/report.md"
  },
  {
   "name": "report_bench_2026-09-01T232540.md",
   "url": "https://synthcore.org/code/1055/report_bench_2026-09-01T232540.md"
  },
  {
   "name": "results.json",
   "url": "https://synthcore.org/code/1055/results.json"
  },
  {
   "name": "stage2_bench.py",
   "url": "https://synthcore.org/code/1055/stage2_bench.py"
  }
 ],
 "category": "sampling",
 "description": "Replace an unconstrained Euclidean diffusion sampler plus projection with a controlled diffusion whose state always lies on an embedded manifold \\(\\mathcal M\\). The neural controller predicts a tangent vector, while the stochastic forcing is also tangent; this preserves constraints during every intermediate denoising step and avoids the bias caused by repeatedly projecting off-manifold states.",
 "download_zip": "https://synthcore.org/download/1055",
 "formulas_latex": [
  "$$\\Pi_x=J(x)J(x)^\\dagger,\\qquad \\operatorname{grad}_{\\mathcal M}E(x)=\\Pi_x\\nabla \\widetilde E(x),$$",
  "$$dX_t=\\left[-\\operatorname{grad}_{\\mathcal M}E(X_t)+u_\\theta(X_t,t)\\right]dt+\\sqrt{2\\beta^{-1}}\\,\\Pi_{X_t}\\circ dW_t,$$",
  "$$\\partial_t\\phi+\\mathcal L\\phi=0,\\qquad u^*(x,t)=2\\beta^{-1}\\operatorname{grad}_{\\mathcal M}\\log\\phi(x,t),$$",
  "$$\\mathcal J(\\theta)=\\mathbb E\\left[E(X_T)+\\frac{\\beta}{4}\\int_0^T\\|u_\\theta(X_t,t)\\|^2dt\\right]+\\lambda\\,\\mathcal L_{\\mathrm{endpoint}}.$$"
 ],
 "id": 2742,
 "implementation": "1. Exact integration point: use this as the sampler inside a diffusion or energy-based generative model whose outputs must lie on a known manifold. Start with the sphere \\(\\mathbb S^{d-1}\\), where \\(\\Pi_x=I-xx^\\top\\), or the Stiefel manifold \\(\\mathrm{St}(n,k)=\\{X:X^\\top X=I_k\\}\\), where \\(\\Pi_X(V)=V-X\\operatorname{sym}(X^\\top V)\\). Let an MLP or transformer controller output an ambient vector \\(v_\\theta(x,t,c)\\), then set \\(u_\\theta=\\Pi_xv_\\theta\\). 2. Pseudocode: initialize \\(x_0\\sim q_0\\) on \\(\\mathcal M\\); for \\(k=0,\\ldots,N-1\\), compute \\(g=\\Pi_x\\nabla E(x)\\), \\(u=\\Pi_xv_\\theta(x,t_k)\\), sample \\(z\\sim\\mathcal N(0,I_d)\\), set \\(z_T=\\Pi_xz\\), and update the intrinsic Euler-Maruyama state \\(x\\leftarrow\\operatorname{Retr}_x[h(-g+u)+\\sqrt{2h/\\beta}\\,z_T]\\). Use the exact exponential map for the sphere when feasible; for Stiefel use a QR or polar retraction, and measure its numerical orthogonality error. Train with the displayed control-energy plus terminal-energy/endpoint loss, or first use supervised score/control matching from short reference trajectories. 3. Computed quantities: \\(\\Pi_x\\), the manifold gradient, tangent noise, and retraction are geometric operations; \\(\\phi\\) and its gradient are estimated by the controller, not analytically computed. Estimate endpoint KL with held-out energy samples or importance weighting, and log \\(\\int\\|u\\|^2dt\\). 4. First cheap experiment: sample a von Mises-Fisher target or a multimodal energy on \\(\\mathbb S^{d-1}\\) using a two-layer MLP, comparing intrinsic Schrödinger-bridge diffusion, Euclidean diffusion, and Euclidean diffusion with final projection. Sweep step size \\(h\\), noise temperature \\(\\beta^{-1}\\), and control capacity. The predicted signature is maximum constraint violation \\(|\\|x\\|_2-1|\\) at every intermediate step below numerical tolerance for the intrinsic method, while the unconstrained method has violations scaling approximately as \\(O(\\sqrt h)\\) before projection. As control capacity and horizon increase, endpoint KL should decrease monotonically until discretization error dominates, while the minimum control energy should increase when the source-target discrepancy is increased. A practical acceptance criterion is at least a 10-fold reduction in intermediate constraint violation and endpoint KL no worse than projection-based diffusion at equal network evaluations.",
 "math_summary": "Let \\(\\mathcal M\\subset\\mathbb R^d\\) be a smooth embedded manifold, \\(\\Pi_x=J(x)J(x)^\\dagger\\) the orthogonal projector onto the tangent space \\(T_x\\mathcal M\\), and \\(E:\\mathcal M\\to\\mathbb R\\) the target energy. The intrinsic Langevin reference process is \\(dX_t=-\\operatorname{grad}_{\\mathcal M}E(X_t)\\,dt+\\sqrt{2\\beta^{-1}}\\Pi_{X_t}\\circ dW_t\\), where \\(\\operatorname{grad}_{\\mathcal M}E=\\Pi_x\\nabla \\widetilde E\\), \\(W_t\\) is ambient Brownian motion, \\(\\beta\\) is inverse temperature, and \\(\\circ\\) denotes Stratonovich integration. Add a learned tangent control \\(u_\\theta(x,t)\\in T_x\\mathcal M\\): \\(dX_t=[-\\operatorname{grad}_{\\mathcal M}E(X_t)+u_\\theta(X_t,t)]dt+\\sqrt{2\\beta^{-1}}\\Pi_{X_t}\\circ dW_t\\). Schrödinger-bridge optimality states that, for reference diffusion generator \\(\\mathcal L\\), the minimum-control solution has \\(u^*(x,t)=2\\beta^{-1}\\operatorname{grad}_{\\mathcal M}\\log\\phi(x,t)\\), where \\(\\phi\\) solves the backward Kolmogorov equation \\(\\partial_t\\phi+\\mathcal L\\phi=0\\) with endpoint conditions determined by the source and target densities. The control cost is \\(\\frac{\\beta}{4}\\mathbb E\\int_0^T\\|u_\\theta(X_t,t)\\|^2dt\\), and the endpoint mismatch plus this cost estimates the path-space relative entropy.",
 "math_tags": [
  "differential-geometry",
  "stochastic-processes",
  "optimal-transport",
  "control-theory",
  "pde"
 ],
 "ml_areas": [
  "diffusion-sampling",
  "diffusion",
  "embedding",
  "architecture"
 ],
 "paper": {
  "arxiv_id": "2608.25838",
  "arxiv_url": "https://arxiv.org/abs/2608.25838",
  "summary_what_math_gives_to_ml": "The paper offers a transferable mechanism for generating samples that remain exactly on a smooth embedded manifold: formulate the diffusion intrinsically using tangent-space noise and Riemannian drift, then learn the minimum-control stochastic bridge from a simple source distribution to an energy-weighted target. The important neural-network asset is not merely manifold-aware coordinates, but state-space feasibility enforced throughout the stochastic trajectory rather than repaired after sampling. This can become a constrained diffusion sampler for sphere-, Stiefel-, orthogonality-, or normalized-embedding-valued outputs, with a control network whose output is projected into the tangent space. The key falsifiable signatures are zero normal constraint violation, a control-energy/KL tradeoff predicted by Schrödinger bridge optimality, and improved sampling compared with unconstrained diffusion followed by projection.",
  "title": "Hard-Constrained Sampling on Embedded Riemannian Manifolds via Adjoint Schrödinger Bridges",
  "year": "2026"
 },
 "ratings": {
  "difficulty": 6,
  "novelty": 7,
  "usefulness": 8
 },
 "solves": [
  "stability",
  "accuracy",
  "sample-efficiency"
 ],
 "title": "Intrinsic Schrödinger Bridge Diffusion",
 "url": "https://synthcore.org/idea/2742/intrinsic-schr-dinger-bridge-diffusion",
 "verification": {
  "benchmark": {
   "track": "manifold_pendulum",
   "model": "rnn_small",
   "n_seeds": 8,
   "metric_direction": "lower is better",
   "comparison": {
    "baseline_mean": 0.0062498191255144775,
    "idea_mean": 0.017734745517373085,
    "delta_mean": 0.011484926391858608,
    "per_seed_diffs": [
     0.016708954237401485,
     0.007543828338384628,
     0.02809743070974946,
     0.0032480936497449875,
     0.00577319273725152,
     0.014188994653522968,
     0.007527842652052641,
     0.00879107415676117
    ],
    "idea_wins": 0,
    "n_pairs": 8,
    "p_value": 0.0081,
    "mde": 0.006700379570878075,
    "mde_rel_pct": 107.20917575877058,
    "verdict": "idea worse (significant)",
    "system_worked": false
   },
   "baseline": {
    "best_cfg": {
     "lr": 0.01
    },
    "sweep": [
     {
      "cfg": {
       "lr": 0.001
      },
      "mean": 0.30809812992811203
     },
     {
      "cfg": {
       "lr": 0.003
      },
      "mean": 0.04099804675206542
     },
     {
      "cfg": {
       "lr": 0.01
      },
      "mean": 0.007313748239539564
     }
    ],
    "full": {
     "mean": 0.0062498191255144775,
     "std": 0.0014967029373309296,
     "per_seed": [
      0.00792852696031332,
      0.00787876546382904,
      0.0054243444465100765,
      0.008023356087505817,
      0.00436387537047267,
      0.006577811203896999,
      0.004088760819286108,
      0.005713112652301788
     ],
     "n": 8
    }
   },
   "idea": {
    "mean": 0.017734745517373085,
    "std": 0.0075728509012216545,
    "per_seed": [
     0.024637481197714806,
     0.015422593802213669,
     0.03352177515625954,
     0.011271449737250805,
     0.01013706810772419,
     0.020766805857419968,
     0.011616603471338749,
     0.014504186809062958
    ],
    "n": 8
   },
   "mechanism_signature": {
    "confirmed": true,
    "custom_track": {
     "domain": "dynamics_and_embedded_manifolds",
     "file": "manifold_dynamics_track.py",
     "name": "manifold_pendulum"
    },
    "idea_config": {
     "lr": 0.01
    },
    "observed_baseline_max_violation": 0.16893672943115234,
    "observed_idea_max_violation": 1.1920928955078125e-7,
    "observed_idea_metric_seed0": 0.024637481197714806,
    "predicted_baseline_violation_order": "nonzero",
    "predicted_idea_violation": 0,
    "prediction": "Intrinsic retraction keeps every predicted embedded angular state on S1; Euclidean output has nonzero norm error."
   },
   "custom_track": null
  },
  "mechanism_confirmed": true,
  "peer_reviewed": false,
  "practical_verdict": "harms",
  "stage1_mechanism_check": {
   "worked": true,
   "confidence": 8,
   "verdict": "Built and numerically verified a sphere-manifold intrinsic diffusion MVP with tangent projection, tangent noise, and normalization retraction. Post-retraction constraint error stayed at 3.33e-16 across all step sizes, unconstrained one-step violation scaled as h^0.570 versus the predicted h^0.5, and the stated control cost scaled quadratically with amplitude with exponent 2.0. Intrinsic and projection-at-every-step updates are equivalent in this toy setting, so this validates constraint preservation but not a generative-quality win over frequent projection.",
   "metrics": {
    "baseline": "Ambient Euler diffusion: median one-step norm violations 0.1596, 0.1043, 0.0706, 0.0487 for h=0.04, 0.02, 0.01, 0.005; fitted exponent 0.570. Short endpoint mean final norm error was 0.4220 and alignment was 0.9739.",
    "idea": "Intrinsic/retracted diffusion: maximum intermediate norm error was 3.33e-16 for every tested step size; short endpoint mean final norm error was 5.05e-17 and alignment was 0.6095. Control-cost sweep exponent was 2.0, with costs 0.02738, 0.10953, 0.43812, 1.75249, and 7.00997."
   },
   "how_to_run": "python3 experiment.py",
   "files": [
    "experiment.py",
    "results.json"
   ],
   "limitations": "Only the sphere manifold and a simple linear energy were tested. No learned neural Schrödinger-bridge controller, endpoint KL estimation, multimodal target, Stiefel manifold, likelihood evaluation, or wall-clock/FLOP comparison was implemented. The sphere intrinsic update uses normalization retraction and is equivalent to projection after every Euler step, so the result validates the geometric mechanism rather than superiority over a well-designed frequent-projection baseline."
  },
  "status": "mech_ok_no_baseline",
  "status_label": "Mechanism confirmed, baseline not beaten",
  "updated_at": "2026-09-01T23:25:40",
  "verdict_source": "deterministic test code (paired-seed permutation statistics)"
 }
}