Existing neural models have difficulty generalizing to unseen combinations of seen
components. To achieve compositional generalization, models are required to consistently …

Abstract Machine learning (ML) systems in natural language processing (NLP) face
significant challenges in generalizing to out-of-distribution (OOD) data, where the test …

Existing neural models are demonstrated to struggle with compositional generalization (CG),
ie, the ability to systematically generalize to unseen compositions of seen components. A …

Transformers generalize to novel compositions of structures and entities after being trained
on a complex dataset, but easily overfit on datasets of insufficient complexity. We observe …

Conventional neural machine translation (NMT) models typically use subwords and words
as the basic units for model input and comprehension. However, complete words and …

In Industry 4.0, achieving semantic interoperability is a significant problem due to the
complexities of current automation systems and the numerous standards involved. The study …

In a multilingual neural machine translation model that fully shares parameters across all
languages, a popular approach is to use an artificial language token to guide translation into …

Human intelligence demonstrates compositionality, the algebraic capacity that enables us to
understand and produce a potentially infinite number of novel combinations of known …