SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2001.06588 · 2020

FlexiBO: A Decoupled Cost-Aware Multi-Objective Optimization Approach for Deep Neural Networks

arXiv · PDF · Open in the Atlas

Code that ran

5 of 11 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 5 have no contract verdict; 0 failures of our instrument; 6 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/2001.06588.json · badge: /api/ran/2001.06588.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
softsys4ai/FlexiBO canonical (classifier label)
classifier basis: census
5 ran · 6 no recorded run
FunctionStatusWhere it lives
exitFlow Ran
ran on a synthesized input
softsys4ai/FlexiBO/networks/xception.py
code served (permissive licence) · get_code("6688f2b9ca92908c")
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, 10) 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 with shape (batch_size, height, width, channels)
# The function expects a 4D tensor suitable for 2D convolutions
x = tf.random.normal((2, 8, 8, 3))

# Number of output classes
n_classes = 10

# Filter size for convolutional operations
filter_size = 3

ARGS = {
    "x": x,
    "n_classes": n_classes,
    "filter_size": filter_size
}
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, 10) — shape (2, 10) 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
group Ran
ran on a synthesized input
softsys4ai/FlexiBO/networks/inceptionv1.py
code served (permissive licence) · get_code("eb7abab5b4404fe3")
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 as tf
from tensorflow.keras import Input
from tensorflow.keras.layers import Conv2D, ReLU, ZeroPadding2D, Flatten, Dropout
from tensorflow.keras.layers import MaxPooling2D, Dense, Concatenate, AveragePooling2D

# Create a small input tensor to simulate the stem output
# The stem function reduces 224x224 to approximately 28x28 with 192 channels
inputs = Input(shape=(224, 224, 3))

# Define a minimal stem-like tensor output
# After stem: approximately 28x28x192
x = tf.random.normal((2, 28, 28, 192))

# Define blocks for the group function
# Each block is a tuple of (f1x1, f3x3, f5x5, fpool)
# where each is a list/tuple of filter counts
blocks = [
    ([64], [96, 128], [16, 32], [32]),      # Inception block 1
    ([128], [128, 192], [32, 96], [64]),    # Inception block 2
    None,                                    # Auxiliary classifier
    ([192], [96, 208], [16, 48], [64]),     # Inception block 3
]

n_classes = 1000

ARGS = {
    'x': x,
    'blocks': blocks,
    'pooling': True,
    'n_classes': n_classes
}
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
group Ran
ran on a synthesized input
softsys4ai/FlexiBO/networks/squeezenet.py
code served (permissive licence) · get_code("09486147ac624af7")
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, 1, 1, 128) 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
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Concatenate, Dropout
import numpy as np

# Create a small input tensor (batch_size=2, height=4, width=4, channels=3)
x = tf.random.normal((2, 4, 4, 3))

# Define filters list for the group function
filters = [8, 16]

# Construct ARGS dict
ARGS = {
    'x': x,
    'filters': filters
}
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, 1, 1, 128) — shape (2, 1, 1, 128) 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
projection_block Ran
ran on a synthesized input
softsys4ai/FlexiBO/networks/resnet50.py
code served (permissive licence) · get_code("c67895e93916524d")
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, 2, 2, 32) 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 the projection_block
# The function expects a tensor that can be passed through Conv2D layers
# Shape: (batch_size, height, width, channels)
batch_size = 2
height = 4
width = 4
input_channels = 3

x = tf.random.normal((batch_size, height, width, input_channels))

# Define arguments
n_filters = 8
filter_size = 3
strides = (2, 2)

ARGS = {
    'n_filters': n_filters,
    'filter_size': filter_size,
    'x': x,
    'strides': strides
}
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, 2, 2, 32) — shape (2, 2, 2, 32) 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
stem Ran
ran on a synthesized input
softsys4ai/FlexiBO/networks/resnet50.py
code served (permissive licence) · get_code("6bde0b827d9c7e67")
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 KerasTensor. 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

# Create a small input tensor with shape (batch_size, height, width, channels)
# batch=2, height=4, width=3, channels=1 (distinct dimensions for probing)
inputs = tf.keras.Input(shape=(4, 3, 1))

# Define n_filters parameter
n_filters = 8

# Construct the ARGS dictionary
ARGS = {
    'inputs': inputs,
    'n_filters': n_filters
}
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 KerasTensor — a KerasTensor 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
bottleneck_block Not yet run
unverified (harvested, no recorded run)
softsys4ai/FlexiBO/networks/resnet50.py
code served (permissive licence) · get_code("e04dc9dc4a87450d")
How it ran

no invocation recorded.

entryFlow Not yet run
unverified (harvested, no recorded run)
softsys4ai/FlexiBO/networks/xception.py
code served (permissive licence) · get_code("26e993102ae84b89")
How it ran

no invocation recorded.

learner Not yet run
unverified (harvested, no recorded run)
softsys4ai/FlexiBO/networks/inceptionv1.py
code served (permissive licence) · get_code("03d2b991fd144b8f")
How it ran

no invocation recorded.

learner Not yet run
unverified (harvested, no recorded run)
softsys4ai/FlexiBO/networks/squeezenet.py
code served (permissive licence) · get_code("b4aa07555c351c35")
How it ran

no invocation recorded.

middleFlow Not yet run
unverified (harvested, no recorded run)
softsys4ai/FlexiBO/networks/xception.py
code served (permissive licence) · get_code("cc698640c4496b75")
How it ran

no invocation recorded.

stem Not yet run
unverified (harvested, no recorded run)
softsys4ai/FlexiBO/networks/squeezenet.py
code served (permissive licence) · get_code("af56cd5321ea5452")
How it ran

no invocation recorded.

Also present in other papers' code

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

The design of machine learning systems often requires trading off different objectives, for example, prediction error and energy consumption for deep neural networks (DNNs). Typically, no single design performs well in all objectives; therefore, finding Pareto-optimal designs is of interest. The search for Pareto-optimal designs involves evaluating designs in an iterative process, and the measurements are used to evaluate an acquisition function that guides the search process. However, measuring different objectives incurs different costs. For example, the cost of measuring the prediction error of DNNs is orders of magnitude higher than that of measuring the energy consumption of a pre-trained DNN, as it requires re-training the DNN. Current state-of-the-art methods do not consider this difference in objective evaluation cost, potentially incurring expensive evaluations of objective functions in the optimization process. In this paper, we develop a novel decoupled and cost-aware multi-objective optimization algorithm, we call Flexible Multi-Objective Bayesian Optimization (FlexiBO) to address this issue. FlexiBO weights the improvement of the hypervolume of the Pareto region by the measurement cost of each objective to balance the expense of collecting new information with the knowledge gained through objective evaluations, preventing us from performing expensive measurements for little to no gain. We evaluate FlexiBO on seven state-of-the-art DNNs for image recognition, natural language processing (NLP), and speech-to-text translation. Our results indicate that, given the same total experimental budget, FlexiBO discovers designs with 4.8$\%$ to 12.4$\%$ lower hypervolume error than the best method in state-of-the-art multi-objective optimization.

For agents

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

get_harvested_code_for_paper("2001.06588")
get_code_for_paper("2001.06588")
have("2001.06588")

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

curl https://syntology.ai/api/ran/2001.06588.json

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

[![Syntology run record](https://syntology.ai/api/ran/2001.06588.svg)](https://syntology.ai/paper/2001.06588)

Connect an agent — have() is free.