SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2003.06417 · NeurIPS · 2020

Sparse Graphical Memory for Robust Planning

Pieter Abbeel, Scott Emmons, Ajay Jain, Michael Laskin, Thanard Kurutach, Deepak Pathak

arXiv · PDF · Open in the Atlas

Code that ran

1 of 13 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 1 has no contract verdict; 0 failures of our instrument; 12 no recorded run), measured 2026-11-21; 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/2003.06417.json · badge: /api/ran/2003.06417.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
scottemmons/sgm canonical (classifier label)
classifier basis: census
1 ran · 12 no recorded run
FunctionStatusWhere it lives
reset_with_difficulty Ran
ran on a synthesized input
scottemmons/sgm/sgm/utils/visualization.py
code served (permissive licence) · get_code("f2197aad1943ea5a")
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 torch
import numpy as np
from unittest.mock import Mock, MagicMock

torch.manual_seed(0)

# Create mock gym environment
mock_gym = Mock()
mock_gym.max_goal_dist = 10.0
mock_gym.set_sample_goal_args = Mock()

# Create mock pyenv wrapper
mock_pyenv = Mock()
mock_pyenv.envs = [Mock()]
mock_pyenv.envs[0].gym = mock_gym

# Create mock TimeStep with observation dict
mock_observation = {
    'observation': torch.tensor([[1.0, 2.0, 3.0, 4.0]], dtype=torch.float32),
    'goal': torch.tensor([[5.0, 6.0, 7.0, 8.0]], dtype=torch.float32)
}

mock_ts = Mock()
mock_ts.observation = mock_observation

# Create mock tf_env
tf_env = Mock()
tf_env.pyenv = mock_pyenv
tf_env.reset = Mock(return_value=mock_ts)

ARGS = {
    'tf_env': tf_env,
    'difficulty': 0.5
}
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 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
cleanup_and_eval Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/evaluation.py
code served (permissive licence) · get_code("08dda5d28096c7ca")
How it ran

no invocation recorded.

eval_search_policy Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/evaluation.py
code served (permissive licence) · get_code("0f50ab9ee95b7647")
How it ran

no invocation recorded.

get_checkpoint_and_writers Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/trainer.py
code served (permissive licence) · get_code("b6178f1dc67d907e")
How it ran

no invocation recorded.

get_combined_dataframe Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/plot.py
code served (permissive licence) · get_code("762860ef32385377")
How it ran

no invocation recorded.

get_plan_vec Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/utils/visualization.py
code served (permissive licence) · get_code("4f7680d8fee085ba")
How it ran

no invocation recorded.

merge_obs_goal Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/actor_critic.py
code served (permissive licence) · get_code("9d4365f23ca27285")
How it ran

no invocation recorded.

resize_walls Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/envs/simple_navigation.py
code served (permissive licence) · get_code("5b6108632887fcaf")
How it ran

no invocation recorded.

set_goal Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/actor_critic.py
code served (permissive licence) · get_code("95d6f61c0f98b7cd")
How it ran

no invocation recorded.

take_cleanup_steps Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/evaluation.py
code served (permissive licence) · get_code("df430b2152f4f3f4")
How it ran

no invocation recorded.

thin_wall_width Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/envs/simple_navigation.py
code served (permissive licence) · get_code("e26cc322f709751d")
How it ran

no invocation recorded.

thin_walls Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/envs/simple_navigation.py
code served (permissive licence) · get_code("7bbcb49f6e110c64")
How it ran

no invocation recorded.

train_eval Not yet run
unverified (harvested, no recorded run)
scottemmons/sgm/sgm/trainer.py
code served (permissive licence) · get_code("3b1432462bcadde8")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 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 13 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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

To operate effectively in the real world, agents should be able to act from highdimensional raw sensory input such as images and achieve diverse goals across long time-horizons. Current deep reinforcement and imitation learning methods can learn directly from high-dimensional inputs but do not scale well to long-horizon tasks. In contrast, classical graphical methods like A* search are able to solve long-horizon tasks, but assume that the state space is abstracted away from raw sensory input. Recent works have attempted to combine the strengths of deep learning and classical planning; however, dominant methods in this domain are still quite brittle and scale poorly with the size of the environment. We introduce Sparse Graphical Memory (SGM), a new data structure that stores states and feasible transitions in a sparse memory. SGM aggregates states according to a novel two-way consistency objective, adapting classic state aggregation criteria to goal-conditioned RL: two states are redundant when they are interchangeable both as goals and as starting states. Theoretically, we prove that merging nodes according to two-way consistency leads to an increase in shortest path lengths that scales only linearly with the merging threshold. Experimentally, we show that SGM significantly outperforms current state of the art methods on long horizon, sparsereward visual navigation tasks. Project video and code are available at https: //mishalaskin.github.io/sgm/.

For agents

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

get_harvested_code_for_paper("2003.06417")
get_code_for_paper("2003.06417")
have("2003.06417")

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

curl https://syntology.ai/api/ran/2003.06417.json

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

[![Syntology run record](https://syntology.ai/api/ran/2003.06417.svg)](https://syntology.ai/paper/2003.06417)

Connect an agent — have() is free.