Word Sense Linking: Disambiguating Outside the Sandbox
Andrei Stefan Bejgu, Edoardo Barba, Luigi Procopio, Alberte Fernández-Castro, Roberto Navigli
2024-12-13
Summary
This paper talks about Word Sense Linking (WSL), a new task that helps machines understand the correct meaning of words in context by identifying which words need clarification and linking them to their appropriate meanings.
What's the problem?
Word Sense Disambiguation (WSD) is a process where computers try to figure out which meaning of a word is being used in a sentence. However, current methods struggle because they assume that all the words needing clarification are already identified and that all possible meanings are provided. This makes it hard for these systems to work effectively with regular text, where meanings can vary widely based on context.
What's the solution?
The authors propose WSL, which not only identifies the words that need disambiguation but also connects them to their correct meanings using a reference list. They developed a transformer-based model that improves on traditional WSD methods by relaxing the strict assumptions previously required. This allows for better performance in understanding and linking word meanings in real-world text.
Why it matters?
This research is important because it enhances how machines process language, making it easier for them to understand and use human language accurately. By improving WSD techniques, WSL can lead to better applications in areas like chatbots, translation services, and any technology that relies on natural language understanding.
Abstract
Word Sense Disambiguation (WSD) is the task of associating a word in a given context with its most suitable meaning among a set of possible candidates. While the task has recently witnessed renewed interest, with systems achieving performances above the estimated inter-annotator agreement, at the time of writing it still struggles to find downstream applications. We argue that one of the reasons behind this is the difficulty of applying WSD to plain text. Indeed, in the standard formulation, models work under the assumptions that a) all the spans to disambiguate have already been identified, and b) all the possible candidate senses of each span are provided, both of which are requirements that are far from trivial. In this work, we present a new task called Word Sense Linking (WSL) where, given an input text and a reference sense inventory, systems have to both identify which spans to disambiguate and then link them to their most suitable meaning.We put forward a transformer-based architecture for the task and thoroughly evaluate both its performance and those of state-of-the-art WSD systems scaled to WSL, iteratively relaxing the assumptions of WSD. We hope that our work will foster easier integration of lexical semantics into downstream applications.