SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2202.09492 · AAAI · 2022

Highlighting Object Category Immunity for the Generalization of Human-Object Interaction Detection

Yong-Lu Li, Xinpeng Liu, Cewu Lu

arXiv · PDF · Open in the Atlas

Code that ran

4 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 4 have no contract verdict; 0 failures 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/2202.09492.json · badge: /api/ran/2202.09492.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
Foruck/OC-Immunity community
Papers with Code archive
4 ran · 2 no recorded run
FunctionStatusWhere it lives
Swish Ran
ran on a synthesized input
Foruck/OC-Immunity/model.py
code served (permissive licence) · get_code("92359874c5d1a6bc")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 2, dtype family float, shape (2, 4) 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 torch
import torch.nn as nn

class Swish(nn.Module):

    def __init__(self):
        super(Swish, self).__init__()

    def forward(self, x):
        return x * torch.sigmoid(x)


INIT_ARGS = {}

ARGS = {
    'x': torch.randn(2, 4)
}

if __name__ == '__main__':
    module = Swish(**INIT_ARGS)
    output = module(**ARGS)
    print(output)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
fusion_module Ran
ran on a synthesized input · constructed an object; no result was computed
Foruck/OC-Immunity/model.py
code served (permissive licence) · get_code("522132f1ee4a0c82")
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 an instance of its own class, fusion_module: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

class Config:
    def __init__(self, num_classes):
        self.NUM_CLASSES = num_classes

config = Config(num_classes=4)

batch = {
    'sH': torch.randn(2, 4),
    'eH': torch.randn(2, 4),
    'sO': torch.randn(2, 4),
    'eO': torch.randn(2, 4),
    'ssp': torch.randn(2, 4),
    'esp': torch.randn(2, 4),
}

ARGS = {
    'config': config
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type fusion_module — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a fusion_module 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
human_classifier Ran
ran on a synthesized input · constructed an object; no result was computed
Foruck/OC-Immunity/model.py
code served (permissive licence) · get_code("738c12b84f0a4a79")
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 an instance of its own class, human_classifier: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn
from collections import OrderedDict

torch.manual_seed(0)

# Define a minimal config class
class Config:
    def __init__(self):
        self.ENCODER = EncoderConfig()
        self.LAYER_SIZE = [512, 256, 128]
        self.NUM_CLASSES = 2
        self.KEY = ['image', 'label']
        self.BN = True
        self.ACT = 'ReLU'
        self.UNSUP_FAC = 1.0
        self.CHECKPOINT = None

class EncoderConfig:
    def __init__(self):
        self.RESNET = True
        self.LAYER_SIZE = [3, 16, 32]
        self.KERNEL_SIZE = [0, 3, 3]
        self.PADDING = [0, 1, 1]
        self.BN = True
        self.ACT = 'ReLU'

# Create config instance
config = Config()

# Create loss_weight tensor - BCEWithLogitsLoss expects pos_weight for the label key
loss_weight = {
    'label': torch.tensor([1.5])
}

ARGS = {
    'config': config,
    'loss_weight': loss_weight
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type human_classifier — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a human_classifier 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
spatial_classifier Ran
ran on a synthesized input · constructed an object; no result was computed
Foruck/OC-Immunity/model.py
code served (permissive licence) · get_code("6012c37ba6de1a36")
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 an instance of its own class, spatial_classifier: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn
from collections import OrderedDict

torch.manual_seed(0)

# Define a minimal config class
class Config:
    def __init__(self):
        self.KEY = ('input_feat', 'label')
        self.LAYER_SIZE = [4, 8, 6]  # input_dim, hidden_dim, output_dim
        self.BN = True
        self.ACT = 'ReLU'
        self.NUM_CLASSES = 3
        self.UNSUP_FAC = 0.1
        self.CHECKPOINT = None

config = Config()

# Define loss_weight parameter
loss_weight = {
    'label': torch.tensor([1.0, 1.0, 1.0])
}

ARGS = {
    'config': config,
    'loss_weight': loss_weight
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type spatial_classifier — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a spatial_classifier 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
object_classifier Not yet run
unverified (harvested, no recorded run)
Foruck/OC-Immunity/model.py
code served (permissive licence) · get_code("596caa8796a528bb")
How it ran

no invocation recorded.

unified_classifier Not yet run
unverified (harvested, no recorded run)
Foruck/OC-Immunity/model.py
code served (permissive licence) · get_code("d0bd78b11d0f3825")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 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.

FunctionAlso present inEarliest-datedOthers named
Swish
brick of this paper's assembly: helper
1 other paper
1 later · 1 only as an assembly brick
2512.02486 (2025-12)
Dual-Robust Cross-Domain Offline Reinforcement Learning Against Dynamics Shifts

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

Human-Object Interaction (HOI) detection plays a core role in activity understanding. As a compositional learning problem (human-verb-object), studying its generalization matters. However, widely-used metric mean average precision (mAP) maybe not enough to model the compositional generalization well. Here, we propose a novel metric, mPD (mean Performance Degradation), as a complementary of mAP to evaluate the performance gap among compositions of different objects and the same verb. Surprisingly, mPD reveals that previous state-of-the-arts usually do not generalize well. With mPD as a cue, we propose Object Category (OC) Immunity to advance HOI generalization. Concretely, our core idea is to prevent model from learning spurious object-verb correlations as a short-cut to over-fit the train set. To achieve OCimmunity, we propose an OC-immune network that decouples the inputs from OC, extracts OC-immune representations and leverages uncertainty quantification to generalize to unseen objects. In both conventional and zero-shot experiments, our method achieves decent improvements. To fully evaluate the generalization, we design a new and more difficult benchmark, on which we present significant advantage. The code is available at https://github.com/Foruck/OC-Immunity.

For agents

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

get_harvested_code_for_paper("2202.09492")
get_code_for_paper("2202.09492")
have("2202.09492")

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

curl https://syntology.ai/api/ran/2202.09492.json

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

[![Syntology run record](https://syntology.ai/api/ran/2202.09492.svg)](https://syntology.ai/paper/2202.09492)

Connect an agent — have() is free.