Andreu Vall

Music Playlist Continuation by Learning from Hand-Curated Examples and Song Features: Alleviating the Cold-Start Problem for Rare and Out-of-Set Songs

Automated music playlist generation is a specific form of music recommendation. Generally stated, the user receives a set of song suggestions defining a coherent listening session. We hypothesize that the best way to convey such playlist coherence to new recommendations is by learning it from actual curated examples, in contrast to imposing ad hoc constraints. Collaborative filtering methods can be used to capture underlying patterns in hand-curated playlists. However, the scarcity of thoroughly curated playlists and the bias towards popular songs result in the vast majority of songs occurring in very few playlists and thus being poorly recommended. To overcome this issue, we propose an alternative model based on a song-to-playlist classifier, which learns the underlying structure from actual playlists while leveraging song features derived from audio, social tags and independent listening logs. Experiments on two datasets of hand-curated playlists show competitive performance compared to collaborative filtering when sufficient training data is available and more robust performance when recommending rare and out-of-set songs. For example, both approaches achieve a recall@100 of roughly 35% for songs occurring in 5 or more training playists, whereas the proposed model achieves a recall@100 of roughly 15% for songs occurring in 4 or less training playlists, compared to the 3% achieved by collaborative filtering.

Exploring Music Diversity Needs Across Countries

Providing diversity in recommendations has shown to positively influence the users subjective evaluations such as satisfaction. However, it is often unknown how much diversity a recommendation set needs to consist of. In this work, we explored how music users of Last.fm apply diversity in their listening behavior. We analyzed a dataset with the music listening history of 53,309 Last.fm users capturing their total listening events until August 2014. We complemented this dataset with The Echo Nest features and Hofstedes cultural dimensions to explore how music diversity is applied across countries. Between 47 countries, we found distinct relationships between the cultural dimensions and music diversity variables. These results suggest that different country-based diversity measurements should be considered when applied to a recommendation set in order to maximize the users subjective evaluations. The country-based relationships also provide opportunities for recommender systems to personalize experiences when user data is limited by being able to rely on the users demographics.

The Role of Quick Response in Accelerating Sales of Fashion Goods

Quick response has been proposed as an appropriate operational strategy to serve volatile markets. In fashion, postponing design, production, and distribution as much as possible may indeed reduce the uncertainty related to product success. In this paper, we provide an empirical study of the influence of lead time and sourcing origin on product success, based on data provided by a European fast fashion retailer. We provide a model of sales diffusion over time where product success is characterized by the speed of sales. We then evaluate how the speed of sales is influenced by the design time and the time-to-market of each particular product. We find that delaying the time of design is very beneficial, because it allows the firm to learn about fashion trends. The effect of time-to-market is more subtle. For a shorter time-to-market, speed of sales is considerably higher, but there is limited learning obtained by postponing design. In contrast, for longer time-to-market, speed of sales is lower, but the learning is higher, so for products designed late in the season, the speed of sales is similar to that of items with short time-to-market.

End-to-end cross-modality retrieval with CCA projections and pairwise ranking loss

Cross-modality retrieval encompasses retrieval tasks where the fetched items are of a different type than the search query, e.g., retrieving pictures relevant to a given text query. The state-of-the-art approach to cross-modality retrieval relies on learning a joint embedding space of the two modalities, where items from either modality are retrieved using nearest-neighbor search. In this work, we introduce a neural network layer based on canonical correlation analysis (CCA) that learns better embedding spaces by analytically computing projections that maximize correlation. In contrast to previous approaches, the CCA layer allows us to combine existing objectives for embedding space learning, such as pairwise ranking losses, with the optimal projections of CCA. We show the effectiveness of our approach for cross-modality retrieval on three different scenarios (text-to-image, audio-sheet-music and zero-shot retrieval), surpassing both Deep CCA and a multi-view network using freely learned projections optimized by a pairwise ranking loss, especially when little training data is available (the code for all three methods is released at: https://github.com/CPJKU/cca_layer).

How item discovery enabled by diversity leads to increased recommendation list attractiveness

Applying diversity to a recommendation list has been shown to positively influence the user experience. A higher perceived diversity is argued to have a positive effect on the attractiveness of the recommendation list and a negative effect on the difficulty to make a choice. In a user study we presented 100 participants with several personalized lists of recommended music artists varying in levels of diversity. Participants were asked to assess these lists on perceived diversity and attractiveness, the experienced choice difficulty and discovery (i.e., the extent the list enriches their taste). We found that recommendation list attractiveness is influenced by two effects: 1) by diversity mediated through discovery; diverse recommendation lists are perceived to be more attractive if they enrich the users taste or 2) by the list familiarity; a higher list familiarity contributes to a higher list attractiveness. We additionally revealed how individual differences (i.e., familiarity) moderate the effects found. Our results have implications on the composition of diversified recommendation lists. Specifically recommended items should contribute in extending and/or deepening the users taste for the diversification to be effective.