SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2102.09225 · NeurIPS · 2021

Continuous Doubly Constrained Batch Reinforcement Learning

Pratik Chaudhari, Jonas Mueller, Amazon Services, Rasool Fakoor, Kavosh Asadi, Alexander Smola

arXiv · PDF · Open in the Atlas

Code that ran

1 of 1 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, 0 violated it and 1 has no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; 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/2102.09225.json · badge: /api/ran/2102.09225.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
amazon-research/cdc-batch-rl marked official in Papers with Code
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
CDC Ran
ran on a synthesized input · constructed an object; no result was computed
amazon-research/cdc-batch-rl/algs/CDC/cdc.py
code served (permissive licence) · get_code("a9ca22247437846a")
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, CDC: 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
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

# Define a minimal actor network
class Actor(nn.Module):
    def __init__(self, obs_dim, action_dim):
        super().__init__()
        self.net = nn.Linear(obs_dim, action_dim)
    
    def forward(self, obs, gt_actions=None, with_log_mle=False):
        action = self.net(obs)
        if with_log_mle:
            # Return action and log probabilities
            log_probs = torch.zeros(action.shape[0], 1)
            return action, log_probs
        return action

# Define a minimal critic network with multiple Q-functions
class Critic(nn.Module):
    def __init__(self, obs_dim, action_dim, num_qs=2):
        super().__init__()
        self.number_of_qs = num_qs
        # Create multiple Q-networks
        self.q_nets = nn.ModuleList([
            nn.Sequential(
                nn.Linear(obs_dim + action_dim, 32),
                nn.ReLU(),
                nn.Linear(32, 1)
            )
            for _ in range(num_qs)
        ])
    
    def forward(self, obs, action):
        # Concatenate obs and action
        x = torch.cat([obs, action], dim=-1)
        # Pass through all Q-networks and stack results
        qs = torch.stack([q_net(x) for q_net in self.q_nets], dim=0)
        return qs

# Create network instances
obs_dim = 4
action_dim = 3
actor = Actor(obs_dim, action_dim)
critic = Critic(obs_dim, action_dim, num_qs=2)

# Define arguments dictionary
ARGS = {
    'actor': actor,
    'critic': critic,
    'p_lr': 1e-3,
    'q_lr': 1e-3,
    'gamma': 0.99,
    'ptau': 0.005,
    'batch_size': 2,
    'max_action': 1.0,
    'num_samples': 15,
    'eta_coef': 1.0,
    'action_dims': (action_dim,),
    'nu': 0.5,
    'max_timesteps': 1000000.0,
    'lambda_coef': 1.0,
    'device': 'cpu'
}
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 CDC — 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 CDC 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

Also present in other papers' code

Of 1 samples lifted from this paper's linked repositories, 1 carry a function identity and 0 do not. Of the 1 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 1 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

Reliant on too many experiments to learn good actions, current Reinforcement Learning (RL) algorithms have limited applicability in real-world settings, which can be too expensive to allow exploration. We propose an algorithm for batch RL, where effective policies are learned using only a fixed offline dataset instead of online interactions with the environment. The limited data in batch RL produces inherent uncertainty in value estimates of states/actions that were insufficiently represented in the training data. This leads to particularly severe extrapolation when our candidate policies diverge from one that generated the data. We propose to mitigate this issue via two straightforward penalties: a policy-constraint to reduce this divergence and a value-constraint that discourages overly optimistic estimates. Over a comprehensive set of 32 continuous-action batch RL benchmarks, our approach compares favorably to state-of-the-art methods, regardless of how the offline data were collected.

For agents

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

get_harvested_code_for_paper("2102.09225")
get_code_for_paper("2102.09225")
have("2102.09225")

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

curl https://syntology.ai/api/ran/2102.09225.json

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

[![Syntology run record](https://syntology.ai/api/ran/2102.09225.svg)](https://syntology.ai/paper/2102.09225)

Connect an agent — have() is free.