SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.04657 · ICML · 2021

Meta-Learning Bidirectional Update Rules

Mark Sandler, Max Vladymyrov, Andrey Zhmoginov, Nolan Miller, Andrew Jackson, Tom Madams, Blaise Ag Üera Y Arcas

arXiv · PDF · Open in the Atlas

Code that ran

2 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 1 has no contract verdict; 1 failure of our instrument; 2 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/2104.04657.json · badge: /api/ran/2104.04657.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/google-research unrelated (classifier label)
classifier basis: classifier_2025
2 ran · 2 no recorded run
FunctionStatusWhere it lives
encoder_w Ran
ran on a synthesized input
google-research/google-research/meta_learning_without_memorization/pose_code/np_bbb.py
code served (permissive licence) · get_code("f67d7efae25ed9e5")
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 SymbolicTensor. 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.v1 as tf
from absl import flags

# Disable v2 behavior for TF1 compatibility
tf.disable_v2_behavior()

# Set up FLAGS
FLAGS = flags.FLAGS
flags.DEFINE_integer('dim_w', 5, 'Dimension of w')

# Create a minimal encoder model
class MockEncoderW0(tf.keras.Model):
    def __init__(self, dim_w):
        super(MockEncoderW0, self).__init__()
        self.dim_w = dim_w
        self.dense = tf.keras.layers.Dense(dim_w)
    
    def call(self, xs):
        # xs shape: [batch_size, 128, 128, 1]
        # Flatten spatial dimensions
        batch_size = tf.shape(xs)[0]
        flattened = tf.reshape(xs, [batch_size, -1])
        # Output shape: [batch_size, dim_w]
        ws = self.dense(flattened)
        return ws

# Create encoder instance
encoder_w0 = MockEncoderW0(dim_w=5)

# Create xs tensor: [n_task=2, n_im=3, dim_x=128*128*1]
# Reshaped to [2, 3, 16384] for the flattened image representation
xs = tf.constant([
    [[i*0.1 + j*0.01 for _ in range(16384)] for j in range(3)]
    for i in range(2)
], dtype=tf.float32)

ARGS = {
    'xs': xs,
    'encoder_w0': encoder_w0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
absl, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
sampling Ran
ran; our contract draft was wrong, not the code
google-research/google-research/meta_learning_without_memorization/pose_code/np_bbb.py
code served (permissive licence) · get_code("503db301e1432cad")
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 tensorflow.compat.v1 as tf

# Disable eager execution to use TF 1.x style
tf.disable_eager_execution()

# Create a small tensor output with shape (batch_size=2, features=8)
# The function splits this into mu and logstd, each with shape (2, 4)
output = tf.constant([
    [0.1, 0.2, 0.3, 0.4, -0.1, -0.2, -0.3, -0.4],
    [0.5, 0.6, 0.7, 0.8, -0.5, -0.6, -0.7, -0.8]
], dtype=tf.float32)

ARGS = {
    "output": output
}
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 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
encoder_r Not yet run
unverified (harvested, no recorded run)
google-research/google-research/meta_learning_without_memorization/pose_code/np_bbb.py
code served (permissive licence) · get_code("01f21b7037fcaf57")
How it ran

no invocation recorded.

xy_to_z Not yet run
unverified (harvested, no recorded run)
google-research/google-research/meta_learning_without_memorization/pose_code/np_bbb.py
code served (permissive licence) · get_code("cd8cea3350fad622")
How it ran

no invocation recorded.

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

In this paper, we introduce a new type of generalized neural network where neurons and synapses maintain multiple states. We show that classical gradient-based backpropagation in neural networks can be seen as a special case of a twostate network where one state is used for activations and another for gradients, with update rules derived from the chain rule. In our generalized framework, networks have neither explicit notion of nor ever receive gradients. The synapses and neurons are updated using a bidirectional Hebb-style update rule parameterized by a shared low-dimensional "genome". We show that such genomes can be meta-learned from scratch, using either conventional optimization techniques, or evolutionary strategies, such as CMA-ES. Resulting update rules generalize to unseen tasks and train faster than gradient descent based optimizers for several standard computer vision and synthetic tasks.

For agents

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

get_harvested_code_for_paper("2104.04657")
get_code_for_paper("2104.04657")
have("2104.04657")

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

curl https://syntology.ai/api/ran/2104.04657.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.04657.svg)](https://syntology.ai/paper/2104.04657)

Connect an agent — have() is free.