SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2603.03897 · 2026

IROSA: Interactive Robot Skill Adaptation Using Natural Language

Markus Knauer, Samuel Bustamante, Freek Stulp, Alin Albu-Schaffer, Joao Silverio, Thomas Eiband

arXiv · PDF · Open in the Atlas

Code that ran

5 of 11 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 2 failures of our instrument; 6 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/2603.03897.json · badge: /api/ran/2603.03897.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
DLR-RM/IROSA community
Papers with Code archive
5 ran · 6 no recorded run
FunctionStatusWhere it lives
Kmp Ran
ran on a synthesized input · constructed an object; no result was computed
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("b3c108d907d04326")
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 an instance of its own class, Kmp: the run constructed the object and did not call it, so no result was computed. 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 numpy as np
import scipy.linalg as sp
from scipy.stats import multivariate_normal
from sklearn.mixture import GaussianMixture

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

# Create small sample data
n_samples = 20
n_dim_in = 2
n_dim_out = 3

X = np.random.randn(n_samples, n_dim_in)
Y = np.random.randn(n_samples, n_dim_out)
x_in = np.random.randn(10, n_dim_in)

ARGS = {
    'gmm_n_components': 3,
    'N': 10,
    'length_scale': 0.1,
    'h': 1.0,
    'lambda1': 0.1,
    'lambda2': 1.0,
    'alpha': 1.0,
    'kernel_function': 'matern2',
    'l': None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Kmp — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Kmp 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
ToolArgsValidationError Ran
ran on a synthesized input · constructed an object; no result was computed
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("64c80be640b7ee6f")
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 an instance of its own class, ToolArgsValidationError: the run constructed the object and did not call it, so no result was computed. 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

torch.manual_seed(0)

class ToolArgsValidationError(Exception):
    """Raised when tool argument validation fails."""
    pass

ARGS = {}
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 ToolArgsValidationError — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ToolArgsValidationError 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
TrajectoryCorrector Ran
ran on a synthesized input
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("d85613fbb203d5fa")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned ndarray, ndim 1, 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
from scipy.spatial.transform import Rotation

# TrajectoryCorrector is a static-method-only class, so __init__ takes no arguments
INIT_ARGS = {}

# The class has only static methods, so we need to pick one to test
# Let's test scene_sdf_gradient_analytical which takes point and obstacles
METHOD = 'scene_sdf_gradient_analytical'

# Create a test point
point = np.array([1.0, 2.0, 3.0])

# Create obstacles list with different types
obstacles = [
    {
        'type': 'sphere',
        'center': np.array([0.0, 0.0, 0.0]),
        'radius': 1.5
    },
    {
        'type': 'obb',
        'center': np.array([5.0, 5.0, 5.0]),
        'half_extents': np.array([1.0, 2.0, 0.5]),
        'rotation': Rotation.from_euler('xyz', [30, 45, 60], degrees=True).as_matrix()
    },
    {
        'type': 'ellipsoid',
        'center': np.array([3.0, 1.0, 2.0]),
        'half_extents': np.array([0.8, 1.2, 0.6]),
        'rotation': Rotation.from_euler('xyz', [15, 30, 45], degrees=True).as_matrix()
    }
]

ARGS = {
    'point': point,
    'obstacles': obstacles
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
kernel_matrix Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("a6fe74e591f37547")
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 ndarray, ndim 2, dtype family float, shape (2, 3) 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 numpy as np

# Construct arguments for kernel_matrix function
x1 = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float64)

x2 = np.array([
    [1.5, 2.5, 3.5, 4.5],
    [5.5, 6.5, 7.5, 8.5],
    [0.5, 1.5, 2.5, 3.5]
], dtype=np.float64)

length_scale = np.array([1.0, 1.2, 0.8, 1.1])

h = 1.0

kernel_function = 'matern2'

ARGS = {
    'x1': x1,
    'x2': x2,
    'length_scale': length_scale,
    'h': h,
    'kernel_function': kernel_function
}
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 ndarray, ndim 2, dtype family float, shape (2, 3) — shape (2, 3) 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
matern_kernel_p2 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("ec8801fdb4f2042e")
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 ndarray, ndim 2, dtype family float, shape (2, 3) 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 numpy as np

# Construct arguments for matern_kernel_p2
x1 = np.random.randn(2, 4)  # shape (2, 4) - 2 points in 4D space
x2 = np.random.randn(3, 4)  # shape (3, 4) - 3 points in 4D space
length_scale = np.array([1.0, 1.5, 2.0, 0.5])  # shape (4,) - one per dimension
h = 1.0

ARGS = {
    "x1": x1,
    "x2": x2,
    "length_scale": length_scale,
    "h": h,
}
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 ndarray, ndim 2, dtype family float, shape (2, 3) — shape (2, 3) 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
DataValidationError Not yet run
unverified (harvested, no recorded run)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("80294cc0a2e14a2d")
How it ran

no invocation recorded.

GaussianMixtureModel Not yet run
unverified (harvested, no recorded run)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("16209b787b4c3a4c")
How it ran

no invocation recorded.

IROSAError Not yet run
unverified (harvested, no recorded run)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("c0bf5e64a8072f40")
How it ran

no invocation recorded.

KMPWrapper Not yet run
unverified (harvested, no recorded run)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("4f951cd744e4229f")
How it ran

no invocation recorded.

Robot Not yet run
unverified (harvested, no recorded run)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("15a607817856ec90")
How it ran

no invocation recorded.

RobotType Not yet run
unverified (harvested, no recorded run)
DLR-RM/IROSA/irosa/models/kmp.py
code served (permissive licence) · get_code("ef03ae48d1876264")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 11 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.

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

Foundation models have demonstrated impressive capabilities across diverse domains, while imitation learning provides principled methods for robot skill adaptation from limited data. Combining these approaches holds significant promise for direct application to robotics, yet this combination has received limited attention, particularly for industrial deployment. We present a novel framework that enables open-vocabulary skill adaptation through a tool-based architecture maintaining a protective abstraction layer between the language model and robot hardware. Our approach leverages pre-trained LLMs to select and parameterize specific tools for adapting robot skills without requiring fine-tuning or direct model-to-robot interaction. We demonstrate the framework on a 7-DoF torque-controlled robot performing an industrial bearing ring insertion task, showing successful skill adaptation through natural language commands for speed adjustment, trajectory correction, and obstacle avoidance while maintaining safety, transparency, and interpretability.

For agents

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

get_harvested_code_for_paper("2603.03897")
get_code_for_paper("2603.03897")
have("2603.03897")

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

curl https://syntology.ai/api/ran/2603.03897.json

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

[![Syntology run record](https://syntology.ai/api/ran/2603.03897.svg)](https://syntology.ai/paper/2603.03897)

Connect an agent — have() is free.