Mitigating Modality Collapse in Multimodal VAEs via Impartial
Optimization
- URL: http://arxiv.org/abs/2206.04496v1
- Date: Thu, 9 Jun 2022 13:29:25 GMT
- Title: Mitigating Modality Collapse in Multimodal VAEs via Impartial
Optimization
- Authors: Adri\'an Javaloy, Maryam Meghdadi and Isabel Valera
- Abstract summary: We argue that this effect is a consequence of conflicting gradients during multimodal VAE training.
We show how to detect the sub-graphs in the computational graphs where gradients conflict.
We empirically show that our framework significantly improves the reconstruction performance, conditional generation, and coherence of the latent space across modalities.
- Score: 7.4262579052708535
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: A number of variational autoencoders (VAEs) have recently emerged with the
aim of modeling multimodal data, e.g., to jointly model images and their
corresponding captions. Still, multimodal VAEs tend to focus solely on a subset
of the modalities, e.g., by fitting the image while neglecting the caption. We
refer to this limitation as modality collapse. In this work, we argue that this
effect is a consequence of conflicting gradients during multimodal VAE
training. We show how to detect the sub-graphs in the computational graphs
where gradients conflict (impartiality blocks), as well as how to leverage
existing gradient-conflict solutions from multitask learning to mitigate
modality collapse. That is, to ensure impartial optimization across modalities.
We apply our training framework to several multimodal VAE models, losses and
datasets from the literature, and empirically show that our framework
significantly improves the reconstruction performance, conditional generation,
and coherence of the latent space across modalities.
Related papers
- Multimodal Unlearnable Examples: Protecting Data against Multimodal Contrastive Learning [53.766434746801366]
Multimodal contrastive learning (MCL) has shown remarkable advances in zero-shot classification by learning from millions of image-caption pairs crawled from the Internet.
Hackers may unauthorizedly exploit image-text data for model training, potentially including personal and privacy-sensitive information.
Recent works propose generating unlearnable examples by adding imperceptible perturbations to training images to build shortcuts for protection.
We propose Multi-step Error Minimization (MEM), a novel optimization process for generating multimodal unlearnable examples.
arXiv Detail & Related papers (2024-07-23T09:00:52Z) - Robust Latent Representation Tuning for Image-text Classification [9.789498730131607]
We propose a robust latent representation tuning method for large models.
Our approach introduces a modality latent translation module to maximize the correlation between modalities, resulting in a robust representation.
Within this framework, common semantics are refined during training, and robust performance is achieved even in the absence of one modality.
arXiv Detail & Related papers (2024-06-10T06:29:00Z) - HiVG: Hierarchical Multimodal Fine-grained Modulation for Visual Grounding [80.85164509232261]
HiVG consists of a multi-layer adaptive cross-modal bridge and a hierarchical multimodal low-rank adaptation (Hi LoRA) paradigm.
Hi LoRA prevents the accumulation of perceptual errors by adapting the cross-modal features from shallow to deep layers in a hierarchical manner.
arXiv Detail & Related papers (2024-04-20T14:57:31Z) - Lightweight In-Context Tuning for Multimodal Unified Models [57.10831399642176]
MultiModal In-conteXt Tuning (M$2$IXT) is a lightweight module to enhance the ICL capabilities of multimodal unified models.
When tuned on as little as 50K multimodal data, M$2$IXT can boost the few-shot ICL performance significantly.
arXiv Detail & Related papers (2023-10-08T10:47:24Z) - Improving Cross-modal Alignment for Text-Guided Image Inpainting [36.1319565907582]
Text-guided image inpainting (TGII) aims to restore missing regions based on a given text in a damaged image.
We propose a novel model for TGII by improving cross-modal alignment.
Our model achieves state-of-the-art performance compared with other strong competitors.
arXiv Detail & Related papers (2023-01-26T19:18:27Z) - Image Generation with Multimodal Priors using Denoising Diffusion
Probabilistic Models [54.1843419649895]
A major challenge in using generative models to accomplish this task is the lack of paired data containing all modalities and corresponding outputs.
We propose a solution based on a denoising diffusion probabilistic synthesis models to generate images under multi-model priors.
arXiv Detail & Related papers (2022-06-10T12:23:05Z) - Balanced Multimodal Learning via On-the-fly Gradient Modulation [10.5602074277814]
Multimodal learning helps to comprehensively understand the world, by integrating different senses.
We propose on-the-fly gradient modulation to adaptively control the optimization of each modality, via monitoring the discrepancy of their contribution towards the learning objective.
arXiv Detail & Related papers (2022-03-29T08:26:38Z) - Weakly supervised segmentation with cross-modality equivariant
constraints [7.757293476741071]
Weakly supervised learning has emerged as an appealing alternative to alleviate the need for large labeled datasets in semantic segmentation.
We present a novel learning strategy that leverages self-supervision in a multi-modal image scenario to significantly enhance original CAMs.
Our approach outperforms relevant recent literature under the same learning conditions.
arXiv Detail & Related papers (2021-04-06T13:14:20Z) - Robust Latent Representations via Cross-Modal Translation and Alignment [36.67937514793215]
Most multi-modal machine learning methods require that all the modalities used for training are also available for testing.
To address this limitation, we aim to improve the testing performance of uni-modal systems using multiple modalities during training only.
The proposed multi-modal training framework uses cross-modal translation and correlation-based latent space alignment.
arXiv Detail & Related papers (2020-11-03T11:18:04Z) - MuCAN: Multi-Correspondence Aggregation Network for Video
Super-Resolution [63.02785017714131]
Video super-resolution (VSR) aims to utilize multiple low-resolution frames to generate a high-resolution prediction for each frame.
Inter- and intra-frames are the key sources for exploiting temporal and spatial information.
We build an effective multi-correspondence aggregation network (MuCAN) for VSR.
arXiv Detail & Related papers (2020-07-23T05:41:27Z) - Multiscale Deep Equilibrium Models [162.15362280927476]
We propose a new class of implicit networks, the multiscale deep equilibrium model (MDEQ)
An MDEQ directly solves for and backpropagates through the equilibrium points of multiple feature resolutions simultaneously.
We illustrate the effectiveness of this approach on two large-scale vision tasks: ImageNet classification and semantic segmentation on high-resolution images from the Cityscapes dataset.
arXiv Detail & Related papers (2020-06-15T18:07:44Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.