SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2605.12704 · 2026

FePySR: A Neural Feature Extraction Framework for Efficient and Scalable Symbolic Regression

Xin Lai, Zhiming Yu, Wangtao Lu

arXiv · PDF · Open in the Atlas

Code that ran

5 of 7 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 2 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2605.12704.json · badge: /api/ran/2605.12704.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
laixn/FePySR canonical (classifier label)
classifier basis: census
5 ran · 2 no recorded run
FunctionStatusWhere it lives
R_scor Ran
ran, honoured the contract we drafted
laixn/FePySR/fepysr/pysr_train.py
code served (permissive licence) · get_code("cafccbb9b5d1ca86")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create small ground truth values
y_true = np.array([1.0, 2.0, 3.0, 4.0, 5.0])

# Create predicted values (close to ground truth for a good R^2 score)
y_pred = np.array([1.1, 2.0, 2.9, 4.1, 4.9])

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
count_all_expressions Ran
ran; our contract draft was wrong, not the code
laixn/FePySR/fepysr/feature_maker.py
code served (permissive licence) · get_code("9222c8f49936f557")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Counter. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from collections import Counter

# Set random seed for reproducibility
import torch
torch.manual_seed(0)

# Define ARGS for count_all_expressions
ARGS = {
    "num_layers": 3,
    "extracted_symbols": [
        ["sin(x)", "cos(x)", "sin(x)"],
        ["x**2", "sin(x)", "exp(x)"],
        ["cos(x)", "log(x)", "x**2"]
    ]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Counter — a Counter output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Counter output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
merge_experiment_results Ran
ran; our contract draft was wrong, not the code
laixn/FePySR/fepysr/feature_maker.py
code served (permissive licence) · get_code("93bfd80df373cb59")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# The function merge_experiment_results takes all_results which is a list of lists
# of shape (num_experiments, num_layers, ...)
# Each element at all_results[experiment_idx][layer_idx] is expected to be a list
# that can be extended into merged_results[layer_idx]

# Create a simple nested list structure representing results from multiple experiments
# Let's say we have 3 experiments, 2 layers, and each layer has symbolic expressions (strings as stand-ins)

experiment_1_results = [
    ["feature_1a", "feature_1b"],  # layer 0
    ["feature_2a"]                  # layer 1
]

experiment_2_results = [
    ["feature_1c", "feature_1d"],  # layer 0
    ["feature_2b", "feature_2c"]   # layer 1
]

experiment_3_results = [
    ["feature_1e"],                # layer 0
    ["feature_2d"]                 # layer 1
]

all_results = [experiment_1_results, experiment_2_results, experiment_3_results]

ARGS = {
    "all_results": all_results
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
squared_loss Ran
ran, honoured the contract we drafted
laixn/FePySR/fepysr/optimization.py
code served (permissive licence) · get_code("79ac7c4d082c8f49")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create small tensors for y_hat and y
# y_hat: predicted values, shape (2, 4) - batch size 2, 4 features
y_hat = torch.randn(2, 4)

# y: target values, shape (2, 4) - batch size 2, 4 features
y = torch.randn(2, 4)

ARGS = {
    "y_hat": y_hat,
    "y": y
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
svd_ridge_regression Ran
ran; our contract draft was wrong, not the code
laixn/FePySR/fepysr/optimization.py
code served (permissive licence) · get_code("90cc682ff8b861b2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (4, 1) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small tensors for ridge regression
# X: feature matrix (n_samples, n_features)
# y: label vector (n_samples,)
X = torch.randn(2, 4, dtype=torch.float32)
y = torch.randn(2, dtype=torch.float32)

# L2 regularization coefficient
alpha = 0.1

ARGS = {
    "X": X,
    "y": y,
    "alpha": alpha,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (4, 1) — shape (4, 1) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_layer_expressions Not yet run
unverified (harvested, no recorded run)
laixn/FePySR/fepysr/feature_maker.py
code served (permissive licence) · get_code("fba5d4f0097cf27d")
How it ran

no invocation recorded.

run_experiments Not yet run
unverified (harvested, no recorded run)
laixn/FePySR/fepysr/fmn_train.py
code served (permissive licence) · get_code("7c272f2a1acfcd4c")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 6 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
squared_loss 1 other paper
1 dated earlier
2006.11316 (2020-06)
SqueezeBERT: What can computer vision teach NLP about efficient neural networks?

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

A fundamental challenge in symbolic regression (SR) is efficiently recovering complex mathematical expressions from observational data. Although this problem is NP-hard, many expressions of practical interest decompose naturally into combinations of nonlinear feature modules, concentrating structural complexity into a small number of reusable components. Here, we introduce FePySR, a two-stage framework that reduces the SR search space by extracting valid features prior to equation search. FePySR first employs a heterogeneous neural network to constrain observational data to a set of candidate expressions, then performs structural optimization within this refined expression space using PySR. Across five standard benchmarks, FePySR outperforms state-of-the-art methods by achieving higher equation recovery rates. On a set of 75 highly complex synthesized equations, FePySR recovers 36 equations, while producing substantially smaller mean squared errors on the remaining unrecovered cases, with reduced computation time compared to PySR. FePySR's first stage also maintains consistent performance under varying numbers of selected top features and increasing levels of noise in the observational data. Applied to ordinary differential equations governing biological systems, FePySR successfully identifies governing equations in 24 out of 100 tests where PySR recovers none. Taken together, FePySR is a generalizable framework that can enhance the SR solvers, enabling the efficient and reliable recovery of symbolic expressions across scientific domains.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2605.12704")
get_code_for_paper("2605.12704")
have("2605.12704")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2605.12704.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2605.12704.svg)](https://syntology.ai/paper/2605.12704)

Connect an agent — have() is free.