AffectFlow-DINO: Uncertainty-Aware Multi-Task Affect Estimation via Conditional Rectified Flow

Hugging Face Daily Papers Papers
affect-estimation multi-task-learning uncertainty-aware rectified-flow facial-expression-recognition action-units abaw-challenge

Summary

This paper presents AffectFlow-DINO, a multi-task learning system for the 11th ABAW challenge that uses a conditional rectified-flow head to model uncertainty in in-the-wild facial behavior estimation, achieving substantial improvements over the baseline.

We present AffectFlow-DINO, a multi-task learning system for the 11th ABAW challenge that extends a standard deterministic architecture with a conditional rectified-flow head to model the inherent ambiguity of in-the-wild facial behavior. Instead of predicting a single affect estimate, the model learns a conditional generative distribution, enabling uncertainty-aware one-to-many predictions through Monte Carlo sampling. The system jointly estimates continuous valence-arousal, classifies eight facial expressions, and detects twelve Action Units from static face images. Built on a frozen DINOv3 ViT-S/16 backbone, extensive ablation studies show that rectified-flow decoding consistently improves deterministic prediction, particularly for valence-arousal estimation (CCC-V +0.058). We further show that post-hoc threshold calibration effectively recovers performance on severely imbalanced rare classes (e.g., Fear: 3.8% rightarrow 33.1%) without retraining. Combined with backbone fine-tuning and flow retuning, the final model achieves P_{MTL=1.177}, substantially outperforming the official challenge baseline of P_{MTL}=0.45.
Original Article
View Cached Full Text

Cached at: 07/16/26, 01:43 PM

Paper page - AffectFlow-DINO: Uncertainty-Aware Multi-Task Affect Estimation via Conditional Rectified Flow

Source: https://huggingface.co/papers/2607.13250

Abstract

WepresentAffectFlow-DINO,amulti-tasklearningsystemforthe11thABAWchallengethatextendsastandarddeterministicarchitecturewithaconditionalrectified-flowheadtomodeltheinherentambiguityofin-the-wildfacialbehavior.Insteadofpredictingasingleaffectestimate,themodellearnsaconditionalgenerativedistribution,enablinguncertainty-awareone-to-manypredictionsthroughMonteCarlosampling.Thesystemjointlyestimatescontinuousvalence-arousal,classifieseightfacialexpressions,anddetectstwelveActionUnitsfromstaticfaceimages.BuiltonafrozenDINOv3ViT-S/16backbone,extensiveablationstudiesshowthatrectified-flowdecodingconsistentlyimprovesdeterministicprediction,particularlyforvalence-arousalestimation(CCC-V+0.058).Wefurthershowthatpost-hocthresholdcalibrationeffectivelyrecoversperformanceonseverelyimbalancedrareclasses(e.g.,Fear:3.8%rightarrow33.1%)withoutretraining.Combinedwithbackbonefine-tuningandflowretuning,thefinalmodelachievesP_{MTL=1.177},substantiallyoutperformingtheofficialchallengebaselineofP_{MTL}=0.45.

View arXiv pageView PDFGitHubAdd to collection

Get this paper in your agent:

hf papers read 2607\.13250

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper1

#### Bekhouche/AffectFlow-DINO Image Classification• Updatedabout 6 hours ago

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.13250 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.13250 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Uncertainty-aware Multi-fidelity Closure via Conditional Normalizing Flows

arXiv cs.LG

This paper proposes an uncertainty-aware multi-fidelity framework based on conditional normalizing flows to improve the predictive accuracy of reduced-order models (ROMs) for complex multiscale systems. The method learns a probabilistic mapping from low-fidelity to high-fidelity coefficients and is demonstrated on a vortex merging problem, showing improved accuracy with uncertainty quantification.