paper-with-me

Papers

Minimizing Memorization in Meta-learning: A Causal Perspective

2021-09-29 · Yinjie Jiang, Zhengyu Chen, Luotian Yuan, Ying WEI, Kun Kuang, Xinhai Ye, Zhihua Wang, Fei Wu

Meta-learning has emerged as a potent paradigm for quick learning of few-shot tasks, by leveraging the meta-knowledge learned from meta-training tasks. Well-generalized meta-knowledge that facilitates fast adaptation in each task is preferred; however, recent evidence suggests the undesirable memorization effect where the meta-knowledge simply memorizing all meta-training tasks discourages task-specific adaptation and poorly generalizes. There have been several solutions to mitigating the effect, including both regularizer-based and augmentation-based methods, while a systematic understanding of these methods in a single framework is still lacking. In this paper, we offer a novel causal perspective of meta-learning. Through the lens of causality, we conclude the universal label space as a confounder to be the causing factor of memorization and frame the two lines of prevailing methods as different deconfounder approaches. Remarkably, derived from the causal inference principle of front-door adjustment, we propose two frustratingly easy but effective deconfounder algorithms, i.e., sampling multiple versions of the meta-knowledge via Dropout and grouping the meta-knowledge into multiple bins. The proposed causal perspective not only brings in the two deconfounder algorithms that surpass previous works in four benchmark datasets towards combating memorization, but also opens a promising direction for meta-learning.

📄 PDF Abstract BibTeX

Code (0)

등록된 구현이 없습니다.

Tasks

Causal InferenceMemorizationMeta-Learning

Methods 이 논문이 사용한 방법론

Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…

Similar Papers 제목 키워드 기반

DMCD: Semantic-Statistical Framework for Causal Discovery

2026-02-23 · Samarth KaPatel, Sofia Nikiforova, Giacinto Paolo Saggese, Paul Smith arxiv

We present DMCD (DataMap Causal Discovery), a two-phase causal discovery framework that integrates LLM-based semantic drafting from variable metadata with statistical validation on observational data. In Phase I, a large…

Systems with Switching Causal Relations: A Meta-Causal Perspective

2024-10-16 · Moritz Willig, Tim Nelson Tobiasch, Florian Peter Busch, Jonas Seng 외

Most work on causality in machine learning assumes that causal relationships are driven by a constant underlying process. However, the flexibility of agents' actions or tipping points in the environmental process can cha…

Interrogating LLM design under a fair learning doctrine

2025-02-22 · Johnny Tian-Zheng Wei, Maggie Wang, Ameya Godbole, Jonathan H. Choi 외

The current discourse on large language models (LLMs) and copyright largely takes a "behavioral" perspective, focusing on model outputs and evaluating whether they are substantially similar to training data. However, sub…

Memorization

Improving Generalization in Meta-Learning via Meta-Gradient Augmentation

2023-06-14 · Ren Wang, Haoliang Sun, Qi Wei, Xiushan Nie 외

Meta-learning methods typically follow a two-loop framework, where each loop potentially suffers from notorious overfitting, hindering rapid adaptation and generalization to new tasks. Existing schemes solve it by enhanc…

Few-Shot LearningMemorizationMeta-LearningNetwork Pruning

A Meta-Learning Perspective on Transformers for Causal Language Modeling

2023-10-09 · Xinbo Wu, Lav R. Varshney

The Transformer architecture has become prominent in developing large causal language models. However, mechanisms to explain its capabilities are not well understood. Focused on the training process, here we establish a …

Causal Language ModelingLanguage ModelingLanguage ModellingMeta-Learning