Style2Vec: Representation Learning for Fashion Items from Style Sets

TL;DR

Style2Vec is a novel embedding method that learns fashion item representations from style sets, enabling better style-aware recommendations and semantic understanding of fashion items.

Contribution

It introduces a new vector representation model for fashion items based on style set co-occurrences, inspired by word embedding techniques.

Findings

Successfully captures fashion semantics like shapes, colors, and patterns.

Outperforms baseline methods in style classification tasks.

Enables meaningful analogies in fashion item representations.

Abstract

With the rapid growth of online fashion market, demand for effective fashion recommendation systems has never been greater. In fashion recommendation, the ability to find items that goes well with a few other items based on style is more important than picking a single item based on the user's entire purchase history. Since the same user may have purchased dress suits in one month and casual denims in another, it is impossible to learn the latent style features of those items using only the user ratings. If we were able to represent the style features of fashion items in a reasonable way, we will be able to recommend new items that conform to some small subset of pre-purchased items that make up a coherent style set. We propose Style2Vec, a vector representation model for fashion items. Based on the intuition of distributional semantics used in word embeddings, Style2Vec learns the representation of a fashion item using other items in matching outfits as context. Two different convolutional neural networks are trained to maximize the probability of item co-occurrences. For evaluation, a fashion analogy test is conducted to show that the resulting representation connotes diverse fashion related semantics like shapes, colors, patterns and even latent styles. We also perform style classification using Style2Vec features and show that our method outperforms other baselines.