MixMo: Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworks - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

MixMo: Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworks

Alexandre Ramé
  • Fonction : Auteur
  • PersonId : 1120615
  • IdRef : 276144511
Rémy Sun
  • Fonction : Auteur
  • PersonId : 1218484
  • IdHAL : remy-sun

Résumé

Recent strategies achieved ensembling ""for free"" by fitting concurrently diverse subnetworks inside a single base network. The main idea during training is that each subnetwork learns to classify only one of the multiple inputs simultaneously provided. However, the question of how to best mix these multiple inputs has not been studied so far. In this paper, we introduce MixMo, a new generalized framework for learning multi-input multi-output deep subnetworks. Our key motivation is to replace the suboptimal summing operation hidden in previous approaches by a more appropriate mixing mechanism. For that purpose, we draw inspiration from successful mixed sample data augmentations. We show that binary mixing in features - particularly with rectangular patches from CutMix - enhances results by making subnetworks stronger and more diverse. We improve state of the art for image classification on CIFAR-100 and Tiny ImageNet datasets. Our easy to implement models notably outperform data augmented deep ensembles, without the inference and memory overheads. As we operate in features and simply better leverage the expressiveness of large networks, we open a new line of research complementary to previous works.
Fichier principal
Vignette du fichier
Rame_MixMo_Mixing_Multiple_Inputs_for_Multiple_Outputs_via_Deep_Subnetworks_ICCV_2021_paper.pdf (2.11 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03483387 , version 1 (16-12-2021)

Identifiants

  • HAL Id : hal-03483387 , version 1

Citer

Alexandre Ramé, Rémy Sun, Matthieu Cord. MixMo: Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworks. ICCV 2021, Mar 2021, Paris, France. pp.823-833. ⟨hal-03483387⟩
52 Consultations
54 Téléchargements

Partager

Gmail Facebook X LinkedIn More