SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2007.12401 · NeurIPS · 2020

Predictive Information Accelerates Learning in RL

Honglak Lee, Ian Fischer, Kuang-Huei Lee, Anthony Liu, Yijie Guo, Sergio Guadarrama

arXiv · PDF · Open in the Atlas

Code that ran

3 of 5 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; 0 failures of our instrument; 2 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/2007.12401.json · badge: /api/ran/2007.12401.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
google-research/pisac canonical (classifier label)
classifier basis: census
3 ran · 2 no recorded run
FunctionStatusWhere it lives
expd Ran
ran on a synthesized input
google-research/pisac/pisac/schedule_utils.py
code served (permissive licence) · get_code("31b1ba8f3801d92a")
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 0, dtype family float, shape () 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.compat.v2 as tf

# Enable eager execution for TensorFlow 2.x compatibility
tf.enable_v2_behavior()

# Define arguments for the expd function
ARGS = {
    "global_step": tf.constant(5, dtype=tf.int32),
    "start_step": tf.constant(0, dtype=tf.int32),
    "end_step": tf.constant(10, dtype=tf.int32),
    "start_val": tf.constant(0.1, dtype=tf.float32),
    "end_val": tf.constant(0.01, dtype=tf.float32),
    "warmup": True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 0, dtype family float, shape () — shape () 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
filter_invalid_transition Ran
ran on a synthesized input
google-research/pisac/pisac/utils.py
code served (permissive licence) · get_code("6084facad2835cd4")
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 0, dtype family float, shape () 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

class MockTrajectory:
    """Mock trajectory object that has is_boundary() method."""
    def __init__(self, batch_size=2, sequence_length=5):
        self.batch_size = batch_size
        self.sequence_length = sequence_length
        # Create boundary flags: shape (sequence_length, batch_size)
        # is_boundary() should return a tensor indicating which steps are boundaries
        self._boundary_flags = tf.constant([
            [False, False],
            [False, True],
            [True, False],
            [False, False],
            [True, True]
        ], dtype=tf.bool)
    
    def is_boundary(self):
        """Return boundary flags tensor of shape (sequence_length, batch_size)."""
        return self._boundary_flags

trajectories = MockTrajectory(batch_size=2, sequence_length=5)
unused_arg1 = None

ARGS = {
    "trajectories": trajectories,
    "unused_arg1": unused_arg1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 0, dtype family float, shape () — shape () 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
get_schedule_fn Ran
ran on a synthesized input
google-research/pisac/pisac/schedule_utils.py
code served (permissive licence) · get_code("b8a75ea16d98165d")
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 tensorflow.compat.v2 as tf

# Define interpolation functions that are referenced in the code
def lerp(current_step, start_step, end_step, start_value, end_value):
    """Linear interpolation."""
    progress = tf.cast(current_step - start_step, tf.float32) / tf.cast(end_step - start_step, tf.float32)
    progress = tf.maximum(0.0, tf.minimum(1.0, progress))
    return start_value + progress * (end_value - start_value)

def berp(current_step, start_step, end_step, start_value, end_value):
    """Beta distribution interpolation."""
    progress = tf.cast(current_step - start_step, tf.float32) / tf.cast(end_step - start_step, tf.float32)
    progress = tf.maximum(0.0, tf.minimum(1.0, progress))
    return start_value + progress * (end_value - start_value)

def expd(current_step, start_step, end_step, start_value, end_value):
    """Exponential decay interpolation."""
    progress = tf.cast(current_step - start_step, tf.float32) / tf.cast(end_step - start_step, tf.float32)
    progress = tf.maximum(0.0, tf.minimum(1.0, progress))
    return start_value * tf.exp(progress * tf.math.log(end_value / (start_value + 1e-8)))

# Create a simple step tensor
step = tf.constant(100, dtype=tf.int32)

# Create test arguments
ARGS = {
    'base': 0.01,
    'sched': '0:1:1000_100:0.5:1000',
    'step': step
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — 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
image_strip_summary Not yet run
unverified (harvested, no recorded run)
google-research/pisac/pisac/utils.py
code served (permissive licence) · get_code("01b60cf975a07343")
How it ran

no invocation recorded.

schedule_from_str Not yet run
unverified (harvested, no recorded run)
google-research/pisac/pisac/schedule_utils.py
code served (permissive licence) · get_code("a42ce42da2f7b6c5")
How it ran

no invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 5 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

The Predictive Information is the mutual information between the past and the future, I(X past ; X future ). We hypothesize that capturing the predictive information is useful in RL, since the ability to model what will happen next is necessary for success on many tasks. To test our hypothesis, we train Soft Actor-Critic (SAC) agents from pixels with an auxiliary task that learns a compressed representation of the predictive information of the RL environment dynamics using a contrastive version of the Conditional Entropy Bottleneck (CEB) objective. We refer to these as Predictive Information SAC (PI-SAC) agents. We show that PI-SAC agents can substantially improve sample efficiency over challenging baselines on tasks from the DM Control suite of continuous control environments. We evaluate PI-SAC agents by comparing against uncompressed PI-SAC agents, other compressed and uncompressed agents, and SAC agents directly trained from pixels. Our implementation is given on GitHub. 1

For agents

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

get_harvested_code_for_paper("2007.12401")
get_code_for_paper("2007.12401")
have("2007.12401")

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

curl https://syntology.ai/api/ran/2007.12401.json

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

[![Syntology run record](https://syntology.ai/api/ran/2007.12401.svg)](https://syntology.ai/paper/2007.12401)

Connect an agent — have() is free.