SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2503.14553 · 2025

Redefining non-IID Data in Federated Learning for Computer Vision Tasks: Migrating from Labels to Embeddings for Task-Specific Data Distributions

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 8 functions out of this paper's own repositories and ran 7 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
kasraborazjani/task-perspective-het canonical 7 of 8
FunctionStatusWhere it lives
compute_pck Ran kasraborazjani/task-perspective-het/calc_pck_keypoints3d.py
code served (permissive licence) · get_code("38461541e0c3c23e")
get_chkpt_name Ran kasraborazjani/task-perspective-het/train_centralized_task_model_multidataset.py
code served (permissive licence) · get_code("dcf3a3d8b3b21c39")
get_default_chkpt_name Ran kasraborazjani/task-perspective-het/extract_cluster_and_dirichlet_viz.py
code served (permissive licence) · get_code("9e518d54b8b148f7")
get_run_dir Ran kasraborazjani/task-perspective-het/extract_cluster_and_dirichlet_viz.py
code served (permissive licence) · get_code("11f8754fc2a611c1")
get_semseg_fname Ran kasraborazjani/task-perspective-het/make_pseudo_scene_labels_pascal.py
code served (permissive licence) · get_code("07c2149680861df5")
heatmaps_to_coords Ran kasraborazjani/task-perspective-het/calc_pck_keypoints3d.py
code served (permissive licence) · get_code("1c804d0a3b938905")
return_task_loss Ran kasraborazjani/task-perspective-het/losses.py
code served (permissive licence) · get_code("26a10a0eea32335f")
return_task_loss Not yet run kasraborazjani/task-perspective-het/pascal_losses.py
code served (permissive licence) · get_code("204faf9fe3e3fefb")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Federated Learning (FL) has emerged as one of the prominent paradigms for distributed machine learning (ML). However, it is well-established that its performance can degrade significantly under non-IID (non-independent and identically distributed) data distributions across clients. To study this effect, the existing works predominantly emulate data heterogeneity by imposing label distribution skew across clients. In this paper, we show that label distribution skew fails to fully capture the data heterogeneity in computer vision tasks beyond classification, exposing an overlooked gap in the literature. Motivated by this, by utilizing pre-trained deep neural networks to extract task-specific data embeddings, we define task-specific data heterogeneity through the lens of each vision task and introduce a new level of data heterogeneity called embedding-based data heterogeneity. Our methodology involves clustering data points based on embeddings and distributing them among clients using the Dirichlet distribution. Through extensive experiments, we evaluate the performance of different FL methods under our revamped notion of data heterogeneity, introducing new benchmark performance measures to the literature. For instance, across seven representative computer vision tasks, our embedding-based heterogeneity formulation leads to up to around 60% increase in the observed loss under FedAvg, indicating that it more accurately exposes the performance degradation caused by data heterogeneity. We further unveil a series of open research directions that can be pursued. (Code: https://github.com/KasraBorazjani/task-perspective-het.git)

For agents

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

get_harvested_code_for_paper("2503.14553")
get_code_for_paper("2503.14553")
have("2503.14553")

Connect an agent — have() is free.