AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

RESFL: An Uncertainty-Aware Framework for Responsible Federated Learning by Balancing Privacy, Fairness and Utility

Dawood Wasif, Terrence J Moore, Jin-Hee Cho

ICLR 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated learning with pretrained models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use RESFL (adversarial disentanglement + UFM-weighted aggregation) instead of FedAvgfederated YOLOv8 · FACETDelta EOP: 0.2362 -> 0.1959Improved
use RESFL instead of FedAvgfederated YOLOv8 · FACETMIA / AIA success rate: MIA 0.3341 -> 0.2093; AIA 0.4431 -> 0.1832Improved
use RESFL instead of FedAvgfederated YOLOv8 · FACETmAP: 0.6378 -> 0.6654Improved
use RESFL instead of FedAvgfederated YOLOv8 · FACET|1-DI|: 0.2159 -> 0.2287Hurt
use RESFL instead of FairFedfederated YOLOv8 · FACETmAP: 0.7013 -> 0.6654Hurt
add differential privacy (eps=1) to FedAvgFedAvg · FACETmAP: 0.6378 -> 0.4612Hurt
raise lambda_priv from 0.01 to 0.1 (lambda_fair=0.1) (ablation)RESFL · FACETmAP / Delta EOP: mAP 0.6654 -> 0.6430; Delta EOP 0.1959 -> 0.3143Hurt
raise lambda_priv from 0.01 to 1 (lambda_fair=0.1) (ablation)RESFL · FACETmAP / Delta EOP: mAP 0.6654 -> 0.5839; Delta EOP 0.1959 -> 0.4146Hurt
raise lambda_priv from 0.01 to 0.1 (lambda_fair=0.1) (ablation)RESFL · FACETMIA success rate: 0.2093 -> 0.1363Improved
raise lambda_fair from 0.01 to 1 (lambda_priv=1) (ablation)RESFL · FACETDelta EOP: 0.3459 -> 0.2513Improved
raise lambda_fair from 0.01 to 1 (lambda_priv=1) (ablation)RESFL · FACETmAP: 0.6056 -> 0.5953 (peak 0.6254 at 0.1)Mixed
use RESFL on CARLA under 100% fog vs other FL methodsRESFL (fine-tuned) · CARLA fog 100%mAP: 0.17 vs others <0.10Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy