This paper presents Babelfy, a unified graph-based approach to Entity Linking (EL) and Word Sense Disambiguation (WSD). The approach leverages BabelNet, a multilingual semantic network, to address the lexical ambiguity in text. Babelfy consists of three main steps: (1) creating semantic signatures for each vertex in the semantic network, (2) identifying candidate meanings for textual fragments, and (3) linking fragments to the most coherent semantic interpretations using a densest subgraph heuristic. Experiments on six datasets, including multilingual settings, demonstrate state-of-the-art performance in both WSD and EL tasks. The approach effectively handles partial mentions and reduces ambiguity by enforcing high coherence in semantic interpretations.This paper presents Babelfy, a unified graph-based approach to Entity Linking (EL) and Word Sense Disambiguation (WSD). The approach leverages BabelNet, a multilingual semantic network, to address the lexical ambiguity in text. Babelfy consists of three main steps: (1) creating semantic signatures for each vertex in the semantic network, (2) identifying candidate meanings for textual fragments, and (3) linking fragments to the most coherent semantic interpretations using a densest subgraph heuristic. Experiments on six datasets, including multilingual settings, demonstrate state-of-the-art performance in both WSD and EL tasks. The approach effectively handles partial mentions and reduces ambiguity by enforcing high coherence in semantic interpretations.