SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2206.05255 · ICML · 2022

Interactively Learning Preference Constraints in Linear Bandits

Andreas Krause, Katja Hofmann, David Lindner, Sebastian Tschiatschek

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 2 have no contract verdict; 0 failures of our instrument; 0 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/2206.05255.json · badge: /api/ran/2206.05255.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
lasgroup/adaptive-constraint-learning canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
CLBAdaptiveGreedyAlgorithm Ran
ran on a synthesized input · constructed an object; no result was computed
lasgroup/adaptive-constraint-learning/src/constrained_bai/algorithms/adaptive_greedy.py
code served (permissive licence) · get_code("8b23d2f771d8f5ac")
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, CLBAdaptiveGreedyAlgorithm: 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
from abc import abstractmethod

class MockBandit:
    def __init__(self):
        self.d = 4  # dimension
        self.n_arms = 5  # number of arms
        self.threshold = 0.5
        
        # constraint_features: list of feature vectors for each arm
        self.constraint_features = [
            np.array([0.1, 0.2, 0.3, 0.4]),
            np.array([0.15, 0.25, 0.35, 0.45]),
            np.array([0.2, 0.3, 0.4, 0.5]),
            np.array([0.25, 0.35, 0.45, 0.55]),
            np.array([0.3, 0.4, 0.5, 0.6]),
        ]
        
        # arms_X_constraint: matrix where each row is a constraint feature
        self.arms_X_constraint = np.array(self.constraint_features)
        
        # reward values for each arm
        self.rewards = np.array([0.8, 0.7, 0.9, 0.6, 0.75])
        
        # constraint observations
        self.constraint_observations = np.array([0.3, 0.4, 0.35, 0.5, 0.45])
        self.obs_idx = 0
    
    def observe_constraint(self, arm_i):
        """Return constraint observation for arm"""
        obs = self.constraint_observations[self.obs_idx % len(self.constraint_observations)]
        self.obs_idx += 1
        return obs
    
    def get_reward(self, arm_i):
        """Return reward for arm"""
        return self.rewards[arm_i]

ARGS = {
    'delta': 0.01,
    'tuned_ci': False,
    'selection': 'adaptive',
    'sqbeta_tuned': 0.5,
    'debug': False,
}
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 CLBAdaptiveGreedyAlgorithm — 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 CLBAdaptiveGreedyAlgorithm 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
CLBAlgorithm Ran
ran on a synthesized input
lasgroup/adaptive-constraint-learning/src/constrained_bai/algorithms/adaptive_greedy.py
code served (permissive licence) · get_code("e8de59c1a0d9d1ce")
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 NoneType. 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

# CLBAlgorithm is abstract, so we need a concrete implementation
class ConcreteCLBAlgorithm:
    def __init__(self):
        pass
    
    def run(self, bandit):
        # Simple implementation that returns a result
        return {"actions": bandit.get_actions(), "rewards": bandit.get_rewards()}

# Mock bandit class
class MockBandit:
    def __init__(self, n_arms=3, n_rounds=5):
        self.n_arms = n_arms
        self.n_rounds = n_rounds
        self.actions = np.random.randint(0, n_arms, size=n_rounds)
        self.rewards = np.random.randn(n_rounds)
    
    def get_actions(self):
        return self.actions
    
    def get_rewards(self):
        return self.rewards

# Create a mock bandit instance
mock_bandit = MockBandit(n_arms=3, n_rounds=5)

INIT_ARGS = {}

ARGS = {
    "bandit": mock_bandit
}

METHOD = "run"
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type NoneType — a NoneType output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a NoneType 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
compute_xnorm Ran
ran, violated the contract we drafted
lasgroup/adaptive-constraint-learning/src/constrained_bai/algorithms/adaptive_greedy.py
code served (permissive licence) · get_code("497a1f936b172919")
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 1, dtype family float, shape (2,) 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 compute_xnorm
# The function expects:
# - arms_X: a list of arm feature vectors
# - A_inv: the inverse of a matrix A (used in the A-norm computation)

# Create a small positive definite matrix and its inverse
# Using a 4x4 matrix for dimension 4
A = np.array([
    [4.0, 1.0, 0.5, 0.2],
    [1.0, 3.0, 0.3, 0.1],
    [0.5, 0.3, 2.0, 0.2],
    [0.2, 0.1, 0.2, 1.5]
], dtype=np.float64)

A_inv = np.linalg.inv(A)

# Create a list of arm feature vectors (each is a 1D array of dimension 4)
# Using 2 arms as per the "batch 2" requirement
arms_X = [
    np.array([1.0, 0.5, 0.2, 0.1]),
    np.array([0.5, 1.0, 0.3, 0.2])
]

ARGS = {
    "arms_X": arms_X,
    "A_inv": A_inv
}
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 1, dtype family float, shape (2,) — shape (2,) 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 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 3 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

We study sequential decision-making with known rewards and unknown constraints, motivated by situations where the constraints represent expensive-to-evaluate human preferences, such as safe and comfortable driving behavior. We formalize the challenge of interactively learning about these constraints as a novel linear bandit problem which we call constrained linear best-arm identification. To solve this problem, we propose the Adaptive Constraint Learning (ACOL) algorithm. We provide an instancedependent lower bound for constrained linear best-arm identification and show that ACOL's sample complexity matches the lower bound in the worst-case. In the average case, ACOL's sample complexity bound is still significantly tighter than bounds of simpler approaches. In synthetic experiments, ACOL performs on par with an oracle solution and outperforms a range of baselines. As an application, we consider learning constraints to represent human preferences in a driving simulation. ACOL is significantly more sample efficient than alternatives for this application. Further, we find that learning preferences as constraints is more robust to changes in the driving scenario than encoding the preferences directly in the reward function.

For agents

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

get_harvested_code_for_paper("2206.05255")
get_code_for_paper("2206.05255")
have("2206.05255")

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

curl https://syntology.ai/api/ran/2206.05255.json

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

[![Syntology run record](https://syntology.ai/api/ran/2206.05255.svg)](https://syntology.ai/paper/2206.05255)

Connect an agent — have() is free.