English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Sequential recommendation is an important recommendation task that aims topredict the next item in a sequence. Recently, adaptations of language models,particularly Transformer-based models such as SASRec and BERT4Rec, haveachieved state-of-the-art results in sequential recommendation. In thesemodels, item ids replace tokens in the original language models. However, thisapproach has limitations. First, the vocabulary of item ids may be many timeslarger than in language models. Second, the classical Top-K recommendationapproach used by these models may not be optimal for complex recommendationobjectives, including auxiliary objectives such as diversity, coverage orcoherence. Recent progress in generative language models inspires us to revisitgenerative approaches to address these challenges. This paper presents theGPTRec sequential recommendation model, which is based on the GPT-2architecture. GPTRec can address large vocabulary issues by splitting item idsinto sub-id tokens using a novel SVD Tokenisation algorithm based on quantiseditem embeddings from an SVD decomposition of the user-item interaction matrix.The paper also presents a novel Next-K recommendation strategy, which generatesrecommendations item-by-item, considering already recommended items. The Next-Kstrategy can be used for producing complex interdependent recommendation lists.We experiment with GPTRec on the MovieLens-1M dataset and show that usingsub-item tokenisation GPTRec can match the quality of SASRec while reducing theembedding table by 40%. We also show that the recommendations generated byGPTRec on MovieLens-1M using the Next-K recommendation strategy match thequality of SASRec in terms of NDCG@10, meaning that the model can serve as astrong starting point for future research.

Generative Sequential Recommendation with GPTRec