SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2306.17833 · NeurIPS · 2023

Resetting the Optimizer in Deep RL: An Empirical Study

Rasool Fakoor, Kavosh Asadi, Shoham Sabach

arXiv · PDF · Open in the Atlas

Code that ran

4 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; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 1 failure of our instrument; 3 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/2306.17833.json · badge: /api/ran/2306.17833.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
devsisters/DQN-tensorflow canonical (classifier label)
classifier basis: census
4 ran · 3 no recorded run
FunctionStatusWhere it lives
class_vars Ran
ran; our contract draft was wrong, not the code
devsisters/DQN-tensorflow/dqn/base.py
pointer only (licence: MIT); the stored copy get_code serves, from hadijomaa/HypRL, is pointer only (licence: NONE) · get_code("3a0e13b504c8be57")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from hadijomaa/HypRL, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

clipped_error Ran
ran on a synthesized input
devsisters/DQN-tensorflow/dqn/ops.py
code served (permissive licence) · get_code("4cb9e720620b6eee")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

def clipped_error(x):
    try:
        return tf.select(tf.abs(x) < 1.0, 0.5 * tf.square(x), tf.abs(x) - 0.5)
    except:
        return tf.where(tf.abs(x) < 1.0, 0.5 * tf.square(x), tf.abs(x) - 0.5)

# Create a small tensor with mixed values to test the clipped error function
# Values include some < 1.0 and some >= 1.0 to exercise both branches
x_data = np.array([[-0.5, 0.2, 1.5, 2.0],
                    [0.0, 0.8, -1.2, 0.3]], dtype=np.float32)

x = tf.constant(x_data)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
rgb2gray Ran
ran on a synthesized input
devsisters/DQN-tensorflow/dqn/utils.py
code served (permissive licence) · get_code("6f76cce2b6054e5f")
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 (4, 5) 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

# Create a small RGB image
# Shape: (height, width, channels) where channels = 3 for RGB
# Using small dimensions: 4 height, 5 width, 3 channels
image = np.random.rand(4, 5, 3).astype(np.float32)

ARGS = {
    "image": image
}
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 (4, 5) — shape (4, 5) 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
timeit Ran
ran on a synthesized input
devsisters/DQN-tensorflow/dqn/utils.py
code served (permissive licence) · get_code("db4c70416d7901da")
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 function. 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 time
import torch

def timeit(f):

    def timed(*args, **kwargs):
        start_time = time.time()
        result = f(*args, **kwargs)
        end_time = time.time()
        print('   [-] %s : %2.5f sec' % (f.__name__, end_time - start_time))
        return result
    return timed

# Create a simple function to be decorated
def sample_function(x, y):
    """A simple function that performs some computation"""
    total = 0
    for i in range(1000):
        total += x + y
    return total

# Apply the decorator
decorated_func = timeit(sample_function)

# Define arguments for calling the decorated function
ARGS = {
    'f': sample_function,
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
get_config Not yet run
unverified (harvested, no recorded run)
devsisters/DQN-tensorflow/config.py
code served (permissive licence) · get_code("d2447e0a77e70092")
How it ran

no invocation recorded.

linear Not yet run
unverified (harvested, no recorded run)
devsisters/DQN-tensorflow/dqn/ops.py
code served (permissive licence) · get_code("5622fc4882566cdc")
How it ran

no invocation recorded.

load_pkl Not yet run
unverified (harvested, no recorded run)
devsisters/DQN-tensorflow/dqn/utils.py
code served (permissive licence) · get_code("d49a89bcb28274d4")
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
class_vars 1 other paper
1 dated earlier
1906.11527 (2019-06)
Hyp-RL : Hyperparameter Optimization by Reinforcement Learning

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 focus on the task of approximating the optimal value function in deep reinforcement learning. This iterative process is comprised of solving a sequence of optimization problems where the loss function changes per iteration. The common approach to solving this sequence of problems is to employ modern variants of the stochastic gradient descent algorithm such as Adam. These optimizers maintain their own internal parameters such as estimates of the first-order and the secondorder moments of the gradient, and update them over time. Therefore, information obtained in previous iterations is used to solve the optimization problem in the current iteration. We demonstrate that this can contaminate the moment estimates because the optimization landscape can change arbitrarily from one iteration to the next one. To hedge against this negative effect, a simple idea is to reset the internal parameters of the optimizer when starting a new iteration. We empirically investigate this resetting idea by employing various optimizers in conjunction with the Rainbow algorithm. We demonstrate that this simple modification significantly improves the performance of deep RL on the Atari benchmark.

For agents

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

get_harvested_code_for_paper("2306.17833")
get_code_for_paper("2306.17833")
have("2306.17833")

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

curl https://syntology.ai/api/ran/2306.17833.json

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

[![Syntology run record](https://syntology.ai/api/ran/2306.17833.svg)](https://syntology.ai/paper/2306.17833)

Connect an agent — have() is free.