SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_16474 · AAAI · 2021

Optimizing Information Theory Based Bitwise Bottlenecks for Efficient Mixed-Precision Activation Quantization

Ji Liu, Xichuan Zhou, Kui Liu, Cong Shi, Haijun Liu

Open in the Atlas

Code that ran

7 of 12 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 6 have no contract verdict; 1 failure of our instrument; 5 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/aaai_16474.json · badge: /api/ran/aaai_16474.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
CQUlearningsystemgroup/BitwiseBottleneck canonical (classifier label)
classifier basis: census
7 ran · 5 no recorded run
FunctionStatusWhere it lives
fixed_padding Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CQUlearningsystemgroup/BitwiseBottleneck/resnet_model.py
code served (permissive licence) · get_code("7fa95a4ad6afe798")
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 4, dtype family float, shape (2, 6, 5, 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 tensorflow as tf
import numpy as np

# Create a small input tensor for testing
# Shape: [batch=2, height=4, width=3, channels=5] for channels_last
inputs_channels_last = tf.constant(
    np.random.randn(2, 4, 3, 5).astype(np.float32)
)

# Alternative: [batch=2, channels=5, height=4, width=3] for channels_first
inputs_channels_first = tf.constant(
    np.random.randn(2, 5, 4, 3).astype(np.float32)
)

# Kernel size
kernel_size = 3

# Data format
data_format = 'channels_last'

# Set up arguments dictionary with channels_last format
ARGS = {
    'inputs': inputs_channels_last,
    'kernel_size': kernel_size,
    'data_format': data_format
}
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 4, dtype family float, shape (2, 6, 5, 5) — shape (2, 6, 5, 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
generate_synthetic_data Ran
ran on a synthesized input
CQUlearningsystemgroup/BitwiseBottleneck/utils/misc/model_helpers.py
code served (permissive licence) · get_code("fff245c120236253")
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 _RepeatDataset. 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 as tf

ARGS = {
    "input_shape": tf.TensorShape([2, 4]),
    "input_value": 1.0,
    "input_dtype": tf.float32,
    "label_shape": tf.TensorShape([2, 3]),
    "label_value": 0.5,
    "label_dtype": tf.float32,
}
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 _RepeatDataset — a _RepeatDataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _RepeatDataset 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_filenames Ran
ran on a synthesized input
CQUlearningsystemgroup/BitwiseBottleneck/imagenet_main.py
code served (permissive licence) · get_code("46be7bbcb9840ec2")
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 list, 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 os

_NUM_TRAIN_FILES = 1024

def get_filenames(is_training, data_dir):
    """Return filenames for dataset."""
    if is_training:
        return [os.path.join(data_dir, 'train-%05d-of-01024' % i) for i in range(_NUM_TRAIN_FILES)]
    else:
        return [os.path.join(data_dir, 'validation-%05d-of-00128' % i) for i in range(128)]


ARGS = {
    'is_training': True,
    'data_dir': '/tmp/dataset'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list 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
get_synth_input_fn Ran
ran on a synthesized input
CQUlearningsystemgroup/BitwiseBottleneck/resnet_run_loop.py
code served (permissive licence) · get_code("caf9e0e8da31477b")
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 as tf

ARGS = {
    "height": 4,
    "width": 3,
    "num_channels": 2,
    "num_classes": 5,
    "dtype": tf.float32,
}
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
past_stop_threshold Ran
ran on a synthesized input
CQUlearningsystemgroup/BitwiseBottleneck/utils/misc/model_helpers.py
code served (permissive licence) · get_code("30d4972b9e8abe2b")
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 bool, 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 numbers
import tensorflow as tf

# Test case 1: stop_threshold is None - should return False
ARGS_1 = {
    'stop_threshold': None,
    'eval_metric': 0.5
}

# Test case 2: eval_metric below threshold - should return False
ARGS_2 = {
    'stop_threshold': 0.9,
    'eval_metric': 0.5
}

# Test case 3: eval_metric at threshold - should return True
ARGS_3 = {
    'stop_threshold': 0.5,
    'eval_metric': 0.5
}

# Test case 4: eval_metric above threshold - should return True
ARGS_4 = {
    'stop_threshold': 0.5,
    'eval_metric': 0.75
}

# Use the most complete test case
ARGS = {
    'stop_threshold': 0.8,
    'eval_metric': 0.85
}
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 bool, 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
process_record_dataset Ran
ran on a synthesized input
CQUlearningsystemgroup/BitwiseBottleneck/resnet_run_loop.py
code served (permissive licence) · get_code("62e9089a02dd3f50")
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 _PrefetchDataset. 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 as tf
import numpy as np

def parse_record_fn(record, is_training, dtype):
    """Stand-in parse function that returns (image, label) pair."""
    # Simulate parsing a record into an image and label
    image = tf.cast(tf.ones((28, 28, 3)), dtype)
    label = tf.constant(0, dtype=tf.int32)
    return image, label

# Create a small dataset with raw record strings
raw_records = [b"record_1", b"record_2", b"record_3", b"record_4"]
dataset = tf.data.Dataset.from_tensor_slices(raw_records)

ARGS = {
    "dataset": dataset,
    "is_training": True,
    "batch_size": 2,
    "shuffle_buffer": 10,
    "parse_record_fn": parse_record_fn,
    "num_epochs": 1,
    "dtype": tf.float32,
    "datasets_num_private_threads": None,
    "drop_remainder": False,
    "tf_data_experimental_slack": False,
}
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 _PrefetchDataset — a _PrefetchDataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _PrefetchDataset 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
register_key_flags_in_core Ran
ran on a synthesized input
CQUlearningsystemgroup/BitwiseBottleneck/utils/flags/core.py
code served (permissive licence) · get_code("cc349b67cb589ec0")
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.

from absl import flags
import inspect

def create_test_function():
    """Create a simple function that returns a list of flag objects."""
    # Create some test flags
    test_flags = [
        flags.DEFINE_string('test_flag_1', 'default_value_1', 'Test flag 1'),
        flags.DEFINE_integer('test_flag_2', 42, 'Test flag 2'),
        flags.DEFINE_boolean('test_flag_3', False, 'Test flag 3'),
    ]
    
    def example_function(*args, **kwargs):
        # This function returns a list of flags (key_flags)
        return test_flags
    
    return example_function

test_function = create_test_function()

ARGS = {
    'f': test_function
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
absl — 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
alpha_update Not yet run
unverified (harvested, no recorded run)
CQUlearningsystemgroup/BitwiseBottleneck/bitwise_information_bottleneck.py
code served (permissive licence) · get_code("30b90c7f1f180724")
How it ran

no invocation recorded.

batch_norm Not yet run
unverified (harvested, no recorded run)
CQUlearningsystemgroup/BitwiseBottleneck/resnet_model.py
code served (permissive licence) · get_code("cfbf106c4ab1fc93")
How it ran

no invocation recorded.

conv2d_fixed_padding Not yet run
unverified (harvested, no recorded run)
CQUlearningsystemgroup/BitwiseBottleneck/resnet_model.py
code served (permissive licence) · get_code("00d606fe5c48bd93")
How it ran

no invocation recorded.

dorefa_quantization Not yet run
unverified (harvested, no recorded run)
CQUlearningsystemgroup/BitwiseBottleneck/bitwise_information_bottleneck.py
code served (permissive licence) · get_code("ed2ac1a063d656ef")
How it ran

no invocation recorded.

preprocess_image Not yet run
unverified (harvested, no recorded run)
CQUlearningsystemgroup/BitwiseBottleneck/imagenet_preprocessing.py
code served (permissive licence) · get_code("efa92eb3f8f87769")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 0 later) and 8 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
fixed_padding 4 other papers
3 dated earlier, 1 same year, order not determinable
1811.06964 (2018-11)
Grasp2Vec: Learning Object Representations from Self-Supervised Grasping
1812.00442 (2018-12)
Deep Cosine Metric Learning for Person Re-Identification
1906.01444 (2019-06)
Heterogeneous Gaussian Mechanism: Preserving Differential Privacy in Deep Learning with Provable Robustness
and 1 more not named here
preprocess_image 2 other papers
2 dated earlier
1904.01569 (2019-04)
Exploring Randomly Wired Neural Networks for Image Recognition
1908.07919 (2019-08)
Deep High-Resolution Representation Learning for Visual Recognition
batch_norm 1 other paper
1 dated earlier
1812.00442 (2018-12)
Deep Cosine Metric Learning for Person Re-Identification
conv2d_fixed_padding 1 other paper
1 dated earlier
1812.00442 (2018-12)
Deep Cosine Metric Learning for Person Re-Identification

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

Recent researches on information theory shed new light on the continuous attempts to open the black box of neural signal encoding. Inspired by the problem of lossy signal compression for wireless communication, this paper presents a Bitwise Bottleneck approach for quantizing and encoding neural network activations. Based on the rate-distortion theory, the Bitwise Bottleneck attempts to determine the most significant bits in activation representation by assigning and approximating the sparse coefficients associated with different bits. Given the constraint of a limited average code rate, the bottleneck minimizes the distortion for optimal activation quantization in a flexible layer-by-layer manner. Experiments over ImageNet and other datasets show that, by minimizing the quantization distortion of each layer, the neural network with bottlenecks achieves the state-of-the-art accuracy with low-precision activation. Meanwhile, by reducing the code rate, the proposed method can improve the memory and computational efficiency by over six times compared with the deep neural network with standard single-precision representation. The source code is available on GitHub: https: //github.com/CQUlearningsystemgroup/BitwiseBottleneck.

For agents

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

get_harvested_code_for_paper("aaai_16474")
get_code_for_paper("aaai_16474")
have("aaai_16474")

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

curl https://syntology.ai/api/ran/aaai_16474.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_16474.svg)](https://syntology.ai/paper/aaai_16474)

Connect an agent — have() is free.