AI-KU: Using Co-Occurrence Modeling for Semantic Similarity

Abstract

In this paper, we describe our unsupervised method submitted to the Cross-Level Semantic Similarity task in Semeval 2014 that computes semantic similarity between two different sized text fragments. Our method models each text fragment by using the cooccurrence statistics of either occurred words or their substitutes. The co-occurrence modeling step provides dense, low-dimensional embedding for each fragment which allows us to calculate semantic similarity using various similarity metrics. Although our current model avoids the syntactic information, we achieved promising results and outperformed all baselines.

Voltron: A Hybrid System For Answer Validation Based On Lexical And Distance Features

“…Although their applications have some differences from the system described in this paper, we consider them relevant because they deal with semantic similarity. (Başkaya, 2014) uses Vector Space Models which have some similarity to our usage of word2vec centroid metrics with the difference that we do not organize the whole text according to the structure of the result matrix, as the VSMs do. The cosine similarity is common for both systems.…”

Large Scale Substitution-based Word Sense Induction

“…Goldberg (2018, 2019) show how contextualized embeddings can be used for achieving state-of-the-art WSI results. The core idea of their WSI algorithm is based on the intuition, first proposed by Başkaya et al (2013), that occurrences of a word that share a sense, also share in-context substitutes. An MLM is then used to derive top-k word substitutes for each word, and these substitutevectors are clustered to derive word senses.…”