SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.05526 · 2021

Simplifying Deep Reinforcement Learning via Self-Supervision

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 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; 0 honoured the contract we drafted, 0 violated it and 4 have 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/2106.05526.json · badge: /api/ran/2106.05526.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
daochenzha/SSRL canonical (classifier label)
classifier basis: census
4 ran · 0 no recorded run
FunctionStatusWhere it lives
runner Ran
ran on a synthesized input
daochenzha/SSRL/baselines/ssrl_continuous/runner.py
code served (permissive licence) · get_code("2c5753c5810a574d")
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 tuple. 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 gym
from gym import spaces

class DummyModel:
    """Minimal model that implements the step() method."""
    def step(self, observation):
        # Return action with same shape as action space
        # observation shape is (4,), return action shape (3,)
        return np.array([0.1, 0.2, 0.3])

class DummyActionNoise:
    """Minimal action noise that implements reset() and __call__()."""
    def reset(self):
        pass
    
    def __call__(self):
        # Return noise with same shape as action space (3,)
        return np.array([0.01, 0.02, 0.03])

class DummyEnv:
    """Minimal environment implementing gym-like interface."""
    def __init__(self):
        self.observation_space = spaces.Box(low=-1.0, high=1.0, shape=(4,), dtype=np.float32)
        self.action_space = spaces.Box(low=-1.0, high=1.0, shape=(3,), dtype=np.float32)
        self.step_count = 0
        self.max_steps = 5
    
    def reset(self):
        """Return initial observation."""
        self.step_count = 0
        return np.array([0.1, 0.2, 0.3, 0.4], dtype=np.float32)
    
    def step(self, action):
        """Execute one step and return (observation, reward, done, info)."""
        self.step_count += 1
        next_obs = np.array([0.2, 0.3, 0.4, 0.5], dtype=np.float32)
        reward = 1.0
        done = self.step_count >= self.max_steps
        info = {}
        return next_obs, reward, done, info

# Create instances
env = DummyEnv()
model = DummyModel()
action_noise = DummyActionNoise()

ARGS = {
    'env': env,
    'model': model,
    'number_trajs': 2,
    'action_noise': action_noise
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
gym, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple 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
runner Ran
ran on a synthesized input
daochenzha/SSRL/baselines/ssrl_discrete/runner.py
code served (permissive licence) · get_code("5e6a29ab6fde317a")
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 tuple. 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 gym
from gym import spaces

class SimpleModel:
    def step(self, ob):
        # Return a simple action based on observation
        if isinstance(ob, np.ndarray):
            return np.array(0)
        return 0

class SimpleEnv(gym.Env):
    def __init__(self):
        self.action_space = spaces.Discrete(2)
        self.observation_space = spaces.Box(low=-1, high=1, shape=(4,), dtype=np.float32)
        self.step_count = 0
        self.max_steps = 3
    
    def reset(self):
        self.step_count = 0
        return np.array([0.1, 0.2, 0.3, 0.4], dtype=np.float32)
    
    def step(self, action):
        self.step_count += 1
        obs = np.array([0.1 * self.step_count, 0.2, 0.3, 0.4], dtype=np.float32)
        reward = float(self.step_count)
        done = self.step_count >= self.max_steps
        info = {}
        return obs, reward, done, info

env = SimpleEnv()
model = SimpleModel()
number_trajs = 2

ARGS = {
    'env': env,
    'model': model,
    'number_trajs': number_trajs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
gym, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple 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
traj_1_generator Ran
ran on a synthesized input
daochenzha/SSRL/baselines/ssrl_continuous/runner.py
code served (permissive licence) · get_code("4c20fbf05f99fd21")
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 tuple, 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
import torch
import torch.nn as nn
from gym import Env, spaces

torch.manual_seed(0)

# Create a minimal model with a step method
class SimplePolicy(nn.Module):
    def __init__(self, obs_dim=4, action_dim=3):
        super().__init__()
        self.fc = nn.Linear(obs_dim, action_dim)
    
    def step(self, ob):
        # ob is expected to be a numpy array
        ob_tensor = torch.from_numpy(np.atleast_1d(ob)).float()
        with torch.no_grad():
            action = self.fc(ob_tensor).numpy()
        return action

# Create a minimal environment
class SimpleEnv(Env):
    def __init__(self):
        self.obs_dim = 4
        self.action_dim = 3
        self.observation_space = spaces.Box(low=-1, high=1, shape=(self.obs_dim,), dtype=np.float32)
        self.action_space = spaces.Box(low=-1, high=1, shape=(self.action_dim,), dtype=np.float32)
        self.step_count = 0
        self.max_steps = 5
    
    def reset(self):
        self.step_count = 0
        return np.random.randn(self.obs_dim).astype(np.float32)
    
    def step(self, action):
        self.step_count += 1
        obs = np.random.randn(self.obs_dim).astype(np.float32)
        reward = float(np.random.randn())
        done = self.step_count >= self.max_steps
        info = {}
        return obs, reward, done, info

# Create a minimal action noise class
class ActionNoise:
    def __init__(self, shape=(3,)):
        self.shape = shape
    
    def reset(self):
        pass
    
    def __call__(self):
        return np.random.randn(*self.shape) * 0.1

model = SimplePolicy(obs_dim=4, action_dim=3)
env = SimpleEnv()
action_noise = ActionNoise(shape=(3,))

ARGS = {
    'model': model,
    'env': env,
    'action_noise': action_noise
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
gym, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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
traj_1_generator Ran
ran on a synthesized input
daochenzha/SSRL/baselines/ssrl_discrete/runner.py
code served (permissive licence) · get_code("5e7f41c81c274472")
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 tuple, 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
import gym
from gym import spaces

class SimpleModel:
    def step(self, ob):
        # Return a random action based on observation
        return np.array([0.5])

class SimpleEnv:
    def __init__(self):
        self.action_space = spaces.Box(low=-1, high=1, shape=(1,), dtype=np.float32)
        self.observation_space = spaces.Box(low=-1, high=1, shape=(4,), dtype=np.float32)
        self.step_count = 0
    
    def reset(self):
        self.step_count = 0
        return np.array([0.1, 0.2, 0.3, 0.4], dtype=np.float32)
    
    def step(self, action):
        self.step_count += 1
        ob = np.array([0.5, 0.6, 0.7, 0.8], dtype=np.float32)
        rew = 1.0
        new = self.step_count >= 3
        info = {}
        return ob, rew, new, info

model = SimpleModel()
env = SimpleEnv()

ARGS = {
    'model': model,
    'env': env
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
gym, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 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 4 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

Supervised regression to demonstrations has been demonstrated to be a stable way to train deep policy networks. We are motivated to study how we can take full advantage of supervised loss functions for stably training deep reinforcement learning agents. This is a challenging task because it is unclear how the training data could be collected to enable policy improvement. In this work, we propose Self-Supervised Reinforcement Learning (SSRL), a simple algorithm that optimizes policies with purely supervised losses. We demonstrate that, without policy gradient or value estimation, an iterative procedure of ``labeling" data and supervised regression is sufficient to drive stable policy improvement. By selecting and imitating trajectories with high episodic rewards, SSRL is surprisingly competitive to contemporary algorithms with more stable performance and less running time, showing the potential of solving reinforcement learning with supervised learning techniques. The code is available at https://github.com/daochenzha/SSRL

For agents

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

get_harvested_code_for_paper("2106.05526")
get_code_for_paper("2106.05526")
have("2106.05526")

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

curl https://syntology.ai/api/ran/2106.05526.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.05526.svg)](https://syntology.ai/paper/2106.05526)

Connect an agent — have() is free.