Related papers: Weak-to-Strong Generalization Through the Data-Cen…

We study generalization properties of weakly supervised learning. That is, learning where only a few "strong" labels (the actual target of our prediction) are present but many more "weak" labels are available. In particular, we show that…

One of the major open problems in machine learning is to characterize generalization in the overparameterized regime, where most traditional generalization bounds become inconsistent even for overparameterized linear regression. In many…

Aligning large-scale commercial models with user intent is crucial to preventing harmful outputs. Current methods rely on human supervision but become impractical as model complexity increases. When models surpass human knowledge, providing…

Future superhuman models will surpass the ability of humans and humans will only be able to \textit{weakly} supervise superhuman models. To alleviate the issue of lacking high-quality data for model alignment, some works on weak-to-strong…

Real-world datasets are often biased with respect to key demographic factors such as race and gender. Due to the latent nature of the underlying factors, detecting and mitigating bias is especially challenging for unsupervised machine…

A machine learning (ML) system must learn not only to match the output of a target function on a training set, but also to generalize to novel situations in order to yield accurate predictions at deployment. In most practical applications,…

Modern large language model (LLM) alignment techniques rely on human feedback, but it is unclear whether these techniques fundamentally limit the capabilities of aligned LLMs. In particular, it is unknown if it is possible to align…

Many recent studies have demonstrated that scaling arguments, such as the so-called hierarchical {\em ansatz}, are extremely useful in understanding the statistical properties of weak gravitational lensing. This is especially true on small…

Common methods for aligning already-capable models with desired behavior rely on the ability of humans to provide supervision. However, future superhuman models will surpass the capability of humans. Therefore, humans will only be able to…

Deep learning achieves remarkable generalization capability with overwhelming number of model parameters. Theoretical understanding of deep learning generalization receives recent attention yet remains not fully explored. This paper…

This paper presents a novel information-theoretic perspective on generalization in machine learning by framing the learning problem within the context of lossy compression and applying finite blocklength analysis. In our approach, the…

Weak-to-Strong generalization (W2SG) is a new trend to elicit the full capabilities of a strong model with supervision from a weak model. While existing W2SG studies focus on simple tasks like binary classification, we extend this paradigm…

Steering the behavior of a strong model pre-trained on internet-scale data can be difficult due to the scarcity of competent supervisors. Recent studies reveal that, despite supervisory noises, a strong student model may surpass its weak…

Machine learning has demonstrated remarkable prediction accuracy over i.i.d data, but the accuracy often drops when tested with data from another distribution. In this paper, we aim to offer another view of this problem in a perspective…

Weak-to-strong generalization (W2SG) refers to the phenomenon where a strong student model, trained on a dataset labeled by a weak teacher, ultimately outperforms the teacher on the target task. Recent studies attribute this performance…

As an emerging and challenging problem in the computer vision community, weakly supervised object localization and detection plays an important role for developing new generation computer vision systems and has received significant…

Recent studies have noted an intriguing phenomenon termed Neural Collapse, that is, when the neural networks establish the right correlation between feature spaces and the training targets, their last-layer features, together with the…

Weak-to-strong alignment offers a promising route to scalable supervision, but it can fail when a strong model becomes confidently wrong on examples that lie in the weak teacher's blind spots. Understanding such failures requires going…

Motivated in part by understanding average case analysis of fundamental algorithms in computer science, and in part by the wide array of network data available over the last decade, a variety of random graph models, with corresponding…

Generalization to unseen degradations remains a fundamental challenge for low-level vision models. This paper aims to investigate the underlying mechanism of this failure, using image deraining as a primary case study due to its…