Update README.md

rutgerswiselab · Oct 15, 2023 · bc9a4b1 · bc9a4b1
1 parent bf403a1
commit bc9a4b1
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/README.md b/README.md
@@ -11,7 +11,7 @@ A relevant repo (OpenP5) of benchmarking foundation models for recommendation is
 
 # Abstract
 
-Recommendation foundation models utilize large language models as their backbone, converting recommendation tasks into natural language tasks. Items are represented as texts and tokenized by language model tokenizers, resulting in multiple tokens and corresponding embeddings for each item, in contrast to the single vector representations in traditional recommender systems. Therefore, creating item ID indexing compatible with language models is essential for recommendation foundation models. In this study, we systematically examine the item indexing problem for recommendation foundation models, using P5 as the representative backbone model and replicating its results with various indexing methods. To emphasize the importance of item indexing, we discuss the issues of several basic methods, such as independent indexing, title indexing, and random indexing, from empirical perspectives. We then propose four simple yet effective solutions, including sequential indexing, collaborative indexing, semantic (content-based) indexing, and hybrid indexing. Our reproducibility study of P5 highlights the significant influence of indexing methods on the model performance, and our results on real-world datasets validate the effectiveness of our proposed solutions.
+Recommendation foundation model utilizes large language models (LLM) for recommendation by converting recommendation tasks into natural language tasks. It enables generative recommendation which directly generates the item(s) to recommend rather than calculating a ranking score for each and every candidate item as in traditional recommendation models, simplifying the recommendation pipeline from multi-stage filtering to single-stage filtering. To avoid generating excessively long text and hallucinated recommendations when deciding which item(s) to recommend, creating LLM-compatible item IDs to uniquely identify each item is essential for recommendation foundation models. In this study, we systematically examine the item ID creation and indexing problem for recommendation foundation models, using P5 as an example of the backbone LLM. To emphasize the importance of item indexing, we first discuss the issues of several trivial item indexing methods, such as random indexing, title indexing, and independent indexing. We then propose four simple yet effective solutions, including sequential indexing, collaborative indexing, semantic (content-based) indexing, and hybrid indexing. Our study highlights the significant influence of item indexing methods on the performance of LLM-based recommendation, and our results on real-world datasets validate the effectiveness of our proposed solutions. The research also demonstrates how recent advances on language modeling and traditional IR principles such as indexing can help each other for better learning and inference.
 
 <img width="900" alt="LLM-RecSys-ID" src="image/LLM-RecSys-ID.png">