Skip to content
erivon
Scientific ML infrastructure

Closed-form derivatives for scientific AI

Derivon computes Laplacians, Hessians, and high-order operators in closed form - orders of magnitude faster than nested autodiff, bit-stable to machine precision, and certified across PyTorch and JAX.

Prefer email? info@derivon.ai

Open-core · PyTorch & JAX · float64 bit-identical · reproducible benchmarks

199x
faster Laplacian
GPU · vs torch.func.hessian, D=240
108x
less memory
GPU · vs dense Hessian, D=240
480x
faster Δ³
GPU · vs nested forward-Laplacian
1e-15
agreement
float64, bit-for-bit

Headline speedups are single data-center GPU measurements. See benchmark methodology.

Built for derivative-heavy scientific computing

42
Python distributions
509
cross-backend parity tests
PyTorch + JAX
bit-identical float64 path
Apache + AGPL
two-tier open licensing
The problem

Autodiff is the wrong tool for iterated high-order operators

Automatic differentiation is exactly right for first-order gradients. But scientific models are dominated by operators that are not first-order gradients - Laplacians, Hessians, trace-of-Jacobians, high-order Stein operators.

For those, nested AD grows exponentially in cost as you iterate the operator, materializes quadratic-memory Hessians, and accumulates round-off across an ever-deeper graph.

  • Dense Hessians scale as O(D²) in memory and time
  • Iterated Laplacians Δᵏ re-pay the whole pass each order
  • High-order finite differences are numerically hopeless
The solution

A closed-form derivative tower from a single forward pass

Derivon evaluates σⁿ(z) as a polynomial recurrence on one activation call. The Laplacian becomes a contraction whose only dimension-dependent term is computed once and reused across the batch - so the overhead is O(1) in D.

quickstart.py
import jax.numpy as jnp
from omnibias.jax import neural_field_value_grad_laplacian

# Closed-form value, gradient, and Laplacian of a one-layer field on R^D.
# One activation-tower evaluation - no nested autodiff, O(1) overhead in D.
val, grad, lap = neural_field_value_grad_laplacian(
    x, W, beta, c, b, activation="tanh"
)

# Iterated Laplacian for high-order PDEs stays flat in k:
from omnibias.jax import neural_field_polylaplacian
d4 = neural_field_polylaplacian(x, W, beta, c, b, "tanh", k=2)  # biharmonic
Platform capabilities

Everything you need to make derivatives cheap and trustworthy

One primitive, exposed as a full operator surface, with reproducible numerics and certified bounds where it matters.

One forward pass, any order

Evaluate the derivative tower σⁿ(z) in closed form for arbitrary n. A single activation call yields the whole tower - gradient, Laplacian, biharmonic, and beyond.

O(1) in input dimension

The Laplacian overhead is independent of D. Grow the dimension 80x and the wall-clock barely moves, because the D-dependent term is computed once and reused.

Bit-identical backends

PyTorch, JAX, and Keras 3 import the same pure-Python coefficients, so every (activation, order) pair is float64-ULP-equal across frameworks - verified each release.

Certified, not just fast

Outward-rounded interval and Taylor-model arithmetic produce enclosures that provably contain the true value, sealed in tamper-evident certificates with a formal-kernel gate.

Drop-in integration

A folx-compatible Laplacian adapter is bit-identical to FermiNet's default. Swap the kernel, keep the physics - the energy curve does not move.

Operators, not just numbers

Gradient, divergence, curl, Hessian, Jacobian, Laplace-Beltrami, exterior derivative, and multivariate jets - all closed-form contractions of the same exact tower.

Architecture

One source of truth, every backend, end to end

The polynomial coefficients live in one pure-Python core. Every backend imports them, so the closed-form derivative math is bit-identical across backends by construction.

omnibias-core
Pure-Python polynomial recurrences (Eulerian / Legendre / Hermite). One source of truth.
PyTorch
bit-identical
JAX
bit-identical
Keras 3
bit-identical
Gradient
Laplacian
Hessian
Δᵏ poly-Laplacian
Jacobian
Divergence / curl
Jets (Faà di Bruno)
Curvature / Fisher
Neural VMC / FermiNet
Physics-informed NNs
Certified numerics + Lean
Benchmarks

Identical answers. Dramatically less compute.

float64 on a single data-center GPU. Every method agrees to ≤ 1e-15 - the speedups are bit-for-bit, never an accuracy trade.

Laplacian wall-clock at D = 240 (relative to omnibias)

float64, GPU, H = 256, B = 4096. Lower is faster. All methods agree to <= 1e-15.

omnibias (closed form)
1.0x
folx
1.13x
jax.hessian
67.7x
torch func.hessian
198.7x
Peak device memory at D = 240

MiB, process-isolated per method. Lower is better.

omnibias
86
folx
86
jax.hessian
5,424
torch
9,305
Derivative overhead vs input dimension D

Per-sample overhead over the forward pass. omnibias is O(1) in D; dense-Hessian autodiff is O(D²).

3123060120240input dimension Ddense Hessian O(D²)omnibias O(1)
Iterated Laplacian Δ³ (D = 60)

milliseconds. omnibias is flat in k; folx-nested OOMs at Δ⁴.

omnibias
0.12 ms
folx-nested
63.5 ms

Conditions: H = 256, B = 4096, float64, data-center GPU. Full derivation and reproduction scripts in the documentation.

Why teams adopt Derivon

A kernel your reviewers can inspect

Ship faster simulations

Cut wall-clock and memory in the inner loop without touching your model's physics or its converged answers.

De-risk the numerics

Bit-identical cross-backend results and certified enclosures replace 'trust me' with reproducible, falsifiable evidence.

Spend less on compute

63-108x memory reductions on high-dimensional Laplacians mean smaller instances and larger batches per GPU.

Put closed-form derivatives in your inner loop

Send us one derivative bottleneck. We will benchmark a closed-form, bit-stable replacement against your current autodiff path - on your problem sizes.

Or email info@derivon.ai