SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.11499 · ECCV · 2024

Bridge Past and Future: Overcoming Information Asymmetry in Incremental Object Detection

Yipeng Gao, Wei-Shi Zheng, Shenghao Fu, Junkai Yan, Ancong Wu, Qijie Mo

arXiv · PDF · Open in the Atlas

Code that ran

6 of 9 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 5 failures of our instrument; 3 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/2407.11499.json · badge: /api/ran/2407.11499.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
isee-laboratory/bpf canonical (classifier label)
classifier basis: census
6 ran · 2 no recorded run
iSEE-Laboratory/BPF community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
add_attention_scores Ran
ran; our contract draft was wrong, not the code
isee-laboratory/bpf/maskrcnn_benchmark/modeling/attention_map.py
code served (permissive licence) · get_code("ed0338a30ec1e616")
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 torch
import torch.nn as nn
from torchvision.ops import roi_align

torch.manual_seed(0)

# Stand-in class for Proposal objects with the required interface
class Proposal:
    def __init__(self, bbox, size):
        self.bbox = bbox
        self.size = size
        self.fields = {}
    
    def add_field(self, name, value):
        self.fields[name] = value

# Create small sample data
# Two proposals with different bounding boxes
proposal_1_bbox = torch.tensor([[10.0, 20.0, 50.0, 80.0], 
                                  [5.0, 15.0, 40.0, 70.0]], dtype=torch.float32)
proposal_1 = Proposal(bbox=proposal_1_bbox, size=(100, 120))

proposal_2_bbox = torch.tensor([[15.0, 25.0, 60.0, 90.0]], dtype=torch.float32)
proposal_2 = Proposal(bbox=proposal_2_bbox, size=(100, 120))

# Create attention maps - shape (batch_size, height, width)
# Using small spatial dimensions for feature maps
attention_map_1 = torch.randn(7, 8, dtype=torch.float32)  # (H_fea=7, W_fea=8)
attention_map_2 = torch.randn(7, 8, dtype=torch.float32)

# Pack into proposals list
proposals = [proposal_1, proposal_2]
attention_maps = [attention_map_1, attention_map_2]

ARGS = {
    "proposals": proposals,
    "attention_maps": attention_maps
}
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 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
calculate_attention_scores_per_img_roi_align Ran
ran; our contract draft was wrong, not the code
isee-laboratory/bpf/maskrcnn_benchmark/modeling/attention_map.py
code served (permissive licence) · get_code("374fba1d23321a7b")
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 Tensor, ndim 2, dtype family float, shape (4, 1) 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 numpy as np
from torchvision.ops import roi_align

torch.manual_seed(0)

# Create a stand-in class for proposals that mimics the expected interface
class ProposalsObject:
    def __init__(self, bbox, size):
        self.bbox = bbox
        self.size = size

# Image dimensions
W_img, H_img = 224, 224

# Feature map dimensions
H_fea, W_fea = 7, 7

# Create bounding boxes in image coordinates (4 boxes with format [x1, y1, x2, y2])
proposals_boxes = torch.tensor([
    [10.0, 20.0, 50.0, 80.0],
    [60.0, 30.0, 150.0, 180.0],
    [5.0, 5.0, 100.0, 100.0],
    [120.0, 120.0, 200.0, 200.0]
], dtype=torch.float32)

# Create proposals object
proposals = ProposalsObject(
    bbox=proposals_boxes,
    size=(W_img, H_img)
)

# Create attention map (H_fea x W_fea)
attention_map = torch.randn(H_fea, W_fea, dtype=torch.float32)

ARGS = {
    "proposals": proposals,
    "attention_map": attention_map
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, torchvision — 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 (4, 1) — shape (4, 1) 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
interpolate Ran
ran; our contract draft was wrong, not the code
isee-laboratory/bpf/maskrcnn_benchmark/layers/misc.py
pointer only (licence: MIT); the stored copy get_code serves, from KanghoonYoon/torch-rasgg, is pointer only (licence: NONE) · get_code("2902bf4410253ff7")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from KanghoonYoon/torch-rasgg, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

next_free_port Ran
ran, honoured the contract we drafted
isee-laboratory/bpf/get_free_port.py
pointer only (licence: MIT); the stored copy get_code serves, from prinshul/JASCL, is pointer only (licence: NONE) · get_code("5bb42fad3fc38eff")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from prinshul/JASCL, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

sigmoid_focal_loss_cpu Ran
ran; our contract draft was wrong, not the code
isee-laboratory/bpf/maskrcnn_benchmark/layers/sigmoid_focal_loss.py
pointer only (licence: MIT); the stored copy get_code serves, from KanghoonYoon/torch-rasgg, is pointer only (licence: NONE) · get_code("426fd02a2beaf826")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from KanghoonYoon/torch-rasgg, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

smooth_l1_loss Ran
ran; our contract draft was wrong, not the code
isee-laboratory/bpf/maskrcnn_benchmark/layers/smooth_l1_loss.py
code served (permissive licence) · get_code("e261fa29066b37e5")
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 Tensor, 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 torch

torch.manual_seed(0)

# Create small tensors for input and target
input_tensor = torch.randn(2, 4)
target_tensor = torch.randn(2, 4)

# Create ARGS dict with all parameters
ARGS = {
    "input": input_tensor,
    "target": target_tensor,
    "beta": 1.0 / 9,
    "size_average": True,
}
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 Tensor, 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
calculate_roi_scores_distillation_losses_new_raw Not yet run
unverified (harvested, no recorded run)
iSEE-Laboratory/BPF/maskrcnn_benchmark/distillation/finetune_distillation_all.py
code served (permissive licence) · get_code("0c50b7e70bf34d44")
How it ran

no invocation recorded.

calculate_roi_scores_distillation_losses_old_raw Not yet run
unverified (harvested, no recorded run)
isee-laboratory/bpf/maskrcnn_benchmark/distillation/finetune_distillation_all.py
code served (permissive licence) · get_code("06d865291a39b351")
How it ran

no invocation recorded.

generate_attention_map Not yet run
unverified (harvested, no recorded run)
isee-laboratory/bpf/maskrcnn_benchmark/modeling/attention_map.py
code served (permissive licence) · get_code("910bc10f3fec4d5d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 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, 3 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
interpolate 38 other papers
36 dated earlier, 2 later
1703.06870 (2017-03)
Mask R-CNN
1803.03243 (2018-03)
Domain Adaptive Faster R-CNN for Object Detection in the Wild
1807.08865 (2018-07)
StereoNet: Guided Hierarchical Refinement for Real-Time Edge-Aware Depth Prediction
and 35 more not named here
smooth_l1_loss 16 other papers
16 dated earlier
1703.06870 (2017-03)
Mask R-CNN
1903.00241 (2019-03)
Mask Scoring R-CNN
1904.01169 (2019-04)
Res2Net: A New Multi-scale Backbone Architecture
and 13 more not named here
sigmoid_focal_loss_cpu 10 other papers
9 dated earlier, 1 later
1803.03243 (2018-03)
Domain Adaptive Faster R-CNN for Object Detection in the Wild
1904.01169 (2019-04)
Res2Net: A New Multi-scale Backbone Architecture
1904.01355 (2019-04)
FCOS: Fully Convolutional One-Stage Object Detection
and 7 more not named here
next_free_port 3 other papers
2 dated earlier, 1 later
Cermelli_Incremental_Learning_in_Semantic_Segmentation_From_Image_Labels_CVPR_2022_paper (2022 (year only))
Incremental Learning in Semantic Segmentation from Image Labels
2210.02326 (2022-10)
Learning Across Domains and Devices: Style-Driven Source-Free Domain Adaptation in Clustered Federated Learning
2605.20538 (2026-05)
Continual Segmentation under Joint Nonstationarity

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 incremental object detection, knowledge distillation has been proven to be an effective way to alleviate catastrophic forgetting. However, previous works focused on preserving the knowledge of old models, ignoring that images could simultaneously contain categories from past, present, and future stages. The co-occurrence of objects makes the optimization objectives inconsistent across different stages since the definition for foreground objects differs across various stages, which limits the model's performance greatly. To overcome this problem, we propose a method called "Bridge Past and Future" (BPF), which aligns models across stages, ensuring consistent optimization directions. In addition, we propose a novel Distillation with Future (DwF) loss, fully leveraging the background probability to mitigate the forgetting of old classes while ensuring a high level of adaptability in learning new classes. Extensive experiments are conducted on both Pascal VOC and MS COCO benchmarks. Without memory, BPF outperforms current state-of-theart methods under various settings. The code is available at https: //github.com/iSEE-Laboratory/BPF.

For agents

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

get_harvested_code_for_paper("2407.11499")
get_code_for_paper("2407.11499")
have("2407.11499")

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

curl https://syntology.ai/api/ran/2407.11499.json

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

[![Syntology run record](https://syntology.ai/api/ran/2407.11499.svg)](https://syntology.ai/paper/2407.11499)

Connect an agent — have() is free.