REVIEW 8 cited by
Recommender Systems with Generative Retrieval
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Modern recommender systems perform large-scale retrieval by first embedding queries and item candidates in the same unified space, followed by approximate nearest neighbor search to select top candidates given a query embedding. In this paper, we propose a novel generative retrieval approach, where the retrieval model autoregressively decodes the identifiers of the target candidates. To that end, we create semantically meaningful tuple of codewords to serve as a Semantic ID for each item. Given Semantic IDs for items in a user session, a Transformer-based sequence-to-sequence model is trained to predict the Semantic ID of the next item that the user will interact with. To the best of our knowledge, this is the first Semantic ID-based generative model for recommendation tasks. We show that recommender systems trained with the proposed paradigm significantly outperform the current SOTA models on various datasets. In addition, we show that incorporating Semantic IDs into the sequence-to-sequence model enhances its ability to generalize, as evidenced by the improved retrieval performance observed for items with no prior interaction history.
Forward citations
Cited by 8 Pith papers
-
Hierarchical Residual Policy Optimization for Generative Recommendations
HRPO decomposes item-level rewards into token-level 'residual credits' along semantic identifier hierarchies and optimizes the generator with a PPO-style objective, improving session utility in KuaiSim and production ...
-
Mosaic: A Fleet of User Embedding Specialists for Recommendation at Meta
Mosaic shows that a fleet of four heterogeneous user-embedding specialists, trained with redundancy-reduction and composite-label losses, improves downstream recommendation quality at Meta.
-
EGR: Embedding-Native Generative Retrieval with a Shared LLM
EGR jointly trains a single LLM to embed both item metadata and user histories in one dense-vector space, and reports superior retrieval and a +2.91% production CVR lift.
-
An LLM-powered Agentic Recommendation System for Connected TV Content Discovery
A hybrid system where an LLM agent generates CTV topic channels and traditional ML ranks the media inside them is engineered to run at production latency via async caching, but only its retrieval component is quality-...
-
Vectorizing the Trie: Efficient Constrained Decoding for LLM-based Generative Retrieval on Accelerators
Constrained decoding for generative retrieval can be made accelerator-friendly by flattening the trie of valid items into a CSR sparse matrix and doing branch-free vectorized lookups.
-
JointRank: Rank Large Set with Single Pass
JointRank partitions candidates into overlapping blocks, ranks each block in parallel with an LLM, and reconstructs a global ranking by aggregating the resulting pairwise comparisons.
-
Session-Level Optimization for Large-Scale Retrieval using REINFORCE with Multi-Step Off-Policy Correction
Off-policy REINFORCE with up to 10 importance-weight factors raises estimated discounted session reward over next-item and positive-only baselines in offline evaluation on the Yambda-5B dataset.
-
Serving the Long Tail: Training-Free LLM Candidate Generation for Vacation Rental Marketplaces
Union fusion of LLM metadata queries with IBKNN extends candidate coverage to cold-start and long-tail Vrbo listings while matching or beating IBKNN recall at every K and collapsing small-vs-frontier LLM gaps under 1%.
Discussion (0). Continue with ORCID to comment.