# Cross-topic distributional semantic representations via unsupervised mappings

## TL;DR

This paper introduces a novel unsupervised method to generate multiple topic-specific distributional semantic representations for words, improving contextual similarity and downstream NLP task performance.

## Contribution

It proposes a new approach to learn and align multiple topic-based DSMs, capturing polysemy more effectively than traditional single-vector models.

## Key findings

- Achieves state-of-the-art results in contextual word similarity
- Outperforms single-prototype models in downstream NLP tasks
- Demonstrates robustness of semantic anchors across topics

## Abstract

In traditional Distributional Semantic Models (DSMs) the multiple senses of a polysemous word are conflated into a single vector space representation. In this work, we propose a DSM that learns multiple distributional representations of a word based on different topics. First, a separate DSM is trained for each topic and then each of the topic-based DSMs is aligned to a common vector space. Our unsupervised mapping approach is motivated by the hypothesis that words preserving their relative distances in different topic semantic sub-spaces constitute robust \textit{semantic anchors} that define the mappings between them. Aligned cross-topic representations achieve state-of-the-art results for the task of contextual word similarity. Furthermore, evaluation on NLP downstream tasks shows that multiple topic-based embeddings outperform single-prototype models.