Meta-learning enables rapid generalization to new tasks by learning meta-knowledge from a variety of tasks. It is intuitively assumed that the more tasks a model learns in one training batch, the richer knowledge it acquires, leading to better generalization performance. However, contrary to this intuition, our experiments reveal an unexpected result: adding more tasks within a single batch actually degrades the generalization performance. To explain this unexpected phenomenon, we conduct a Structural Causal Model (SCM) for causal analysis. Our investigation uncovers the presence of spurious correlations between task-specific causal factors and labels in meta-learning. Furthermore, the confounding factors differ across different batches. We refer to these confounding factors as ``Task Confounders". Based on this insight, we propose a plug-and-play Meta-learning Causal Representation Learner (MetaCRL) to eliminate task confounders. It encodes decoupled causal factors from multiple tasks and utilizes an invariant-based bi-level optimization mechanism to ensure their causality for meta-learning. Extensive experiments on various benchmark datasets demonstrate that our work achieves state-of-the-art (SOTA) performance.
The long-term goal of machine learning is to learn general visual representations from a small amount of data without supervision, mimicking three advantages of human cognition: i) no need for labels, ii) robustness to data scarcity, and iii) learning from experience. Self-supervised learning and meta-learning are two promising techniques to achieve this goal, but they both only partially capture the advantages and fail to address all the problems. Self-supervised learning struggles to overcome the drawbacks of data scarcity, while ignoring prior knowledge that can facilitate learning and generalization. Meta-learning relies on supervised information and suffers from a bottleneck of insufficient learning. To address these issues, we propose a novel Bootstrapped Meta Self-Supervised Learning (BMSSL) framework that aims to simulate the human learning process. We first analyze the close relationship between meta-learning and self-supervised learning. Based on this insight, we reconstruct tasks to leverage the strengths of both paradigms, achieving advantages i and ii. Moreover, we employ a bi-level optimization framework that alternates between solving specific tasks with a learned ability (first level) and improving this ability (second level), attaining advantage iii. To fully harness its power, we introduce a bootstrapped target based on meta-gradient to make the model its own teacher. We validate the effectiveness of our approach with comprehensive theoretical and empirical study.
In recent years, self-supervised learning (SSL) has emerged as a promising approach for extracting valuable representations from unlabeled data. One successful SSL method is contrastive learning, which aims to bring positive examples closer while pushing negative examples apart. Many current contrastive learning approaches utilize a parameterized projection head. Through a combination of empirical analysis and theoretical investigation, we provide insights into the internal mechanisms of the projection head and its relationship with the phenomenon of dimensional collapse. Our findings demonstrate that the projection head enhances the quality of representations by performing contrastive loss in a projected subspace. Therefore, we propose an assumption that only a subset of features is necessary when minimizing the contrastive loss of a mini-batch of data. Theoretical analysis further suggests that a sparse projection head can enhance generalization, leading us to introduce SparseHead - a regularization term that effectively constrains the sparsity of the projection head, and can be seamlessly integrated with any self-supervised learning (SSL) approaches. Our experimental results validate the effectiveness of SparseHead, demonstrating its ability to improve the performance of existing contrastive methods.
Through experiments on various meta-learning methods, task samplers, and few-shot learning tasks, this paper arrives at three conclusions. Firstly, there are no universal task sampling strategies to guarantee the performance of meta-learning models. Secondly, task diversity can cause the models to either underfit or overfit during training. Lastly, the generalization performance of the models are influenced by task divergence, task entropy, and task difficulty. In response to these findings, we propose a novel task sampler called Adaptive Sampler (ASr). ASr is a plug-and-play task sampler that takes task divergence, task entropy, and task difficulty to sample tasks. To optimize ASr, we rethink and propose a simple and general meta-learning algorithm. Finally, a large number of empirical experiments demonstrate the effectiveness of the proposed ASr.
Diffusion models are a powerful class of generative models that can produce high-quality images, but they may suffer from data bias. Data bias occurs when the training data does not reflect the true distribution of the data domain, but rather exhibits some skewed or imbalanced patterns. For example, the CelebA dataset contains more female images than male images, which can lead to biased generation results and affect downstream applications. In this paper, we propose a novel method to mitigate data bias in diffusion models by applying manifold guidance. Our key idea is to estimate the manifold of the training data using a learnable information-theoretic approach, and then use it to guide the sampling process of diffusion models. In this way, we can encourage the generated images to be uniformly distributed on the data manifold, without changing the model architecture or requiring labels or retraining. We provide theoretical analysis and empirical evidence to show that our method can improve the quality and unbiasedness of image generation compared to standard diffusion models.
Generative Adversarial Networks (GANs) and their variants have achieved remarkable success on natural images. It aims to approximate the distribution of the training datasets. However, their performance degrades when applied to remote sensing (RS) images, and the discriminator often suffers from the overfitting problem. In this paper, we examine the differences between natural and RS images and find that the intrinsic dimensions of RS images are much lower than those of natural images. Besides, the low-dimensional data manifold of RS images may exacerbate the uneven sampling of training datasets and introduce biased information. The discriminator can easily overfit to the biased training distribution, leading to a faulty generation model, even the mode collapse problem. While existing GANs focus on the general distribution of RS datasets, they often neglect the underlying data manifold. In respond, we propose a learnable information-theoretic measure that preserves the intrinsic structures of the original data, and establish a unified GAN framework for manifold alignment in supervised and unsupervised RS image generation.