Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

generalisation error

Generalisation error is a measure of how accurately a machine learning model predicts outcomes on previously unseen data rather than on the specific dataset used during training. Also referred to as out-of-sample error, it reflects the discrepancy between a model expected performance over the broader underlying data distribution and its empirical error measured on the training set. A high generalisation error indicates that a model has overfitted to noise or idiosyncratic details in the training examples, whereas a low generalisation error demonstrates that it has learned underlying patterns capable of transferring to novel inputs. In statistical learning theory, calculating or bounding the generalisation error is essential for assessing algorithm reliability, tuning model complexity, and ensuring effective performance across new tasks and domains.

1 item

An embarrassingly simple approach to zero-shot learning

An embarrassingly simple approach to zero-shot learning

Bernardino Romera-Paredes, Philip H. S. Torr

OrganizationsUniversity of Oxford

Why you should read this

Proposes a two-layer linear framework for zero-shot learning that is implementable in a single line of code, establishes theoretical generalization bounds through domain adaptation, and surpasses complex state-of-the-art methods across standard benchmarks by up to 17%.

Zero-shot learning consists in learning how to recognise new concepts by just having a description of them. Many sophisticated approaches have been proposed to address the challenges this problem comprises. In this paper we describe a zero-shot learning approach that can be implemented in just one line of code, yet it is able to outperform state of the art approaches on standard datasets. The approach is based on a more general framework which models the relationships between features, attributes, and classes as a two linear layers network, where the weights of the top layer are not learned but are given by the environment. We further provide a learning bound on the generalisation error of this kind of approaches, by casting them as domain adaptation methods. In experiments carried out on three standard real datasets, we found that our approach is able to perform significantly better than the state of art on all of them, obtaining a ratio of improvement up to 17%.

Added

2026-09-25