Discrimination Is Generation: Unifying Ranking and Retrieval from a Tokenizer Perspective

TL;DR

This paper introduces DIG, a unified model that embeds tokenizers within a discriminative ranking framework, aligning retrieval and ranking tasks for improved recommendation quality.

Contribution

DIG is the first end-to-end training approach that unifies tokenization and ranking, enhancing both retrieval and ranking performance in generative recommendation systems.

Findings

DIG improves ranking and retrieval metrics across multiple datasets.

Unified training leads to better recommendation quality than separate models.

The approach demonstrates effectiveness on both public benchmarks and industrial datasets.

Abstract

Semantic IDs (SIDs) define the generation space of generative recommendation and directly determine its personalization ceiling. However, existing tokenizers are trained independently with retrieval objectives, leaving personalization signals fully decoupled from the SID construction process -- a fundamental gap that causes generative retrieval to persistently lag behind discriminative ranking. In this paper, we rethink the essence of SIDs: \emph{ranking seeks argmax in item space while retrieval seeks argmax in token space; both are the same problem solved at different granularities.} Based on this insight, we propose \DIG (\textbf{D}iscrimination \textbf{I}s \textbf{G}eneration), which embeds the tokenizer inside a discriminative ranking model for end-to-end training -- the ranker naturally becomes a retrieval model, yielding two models from a single training run. \DIG is organized around a \emph{feature assignment taxonomy}: item-intrinsic static features are encoded into SIDs, user-item cross features (u2i) implicitly drive codebook boundaries toward recommendation decision boundaries during training, and an MLP$_\mathrm{u2t}$ distillation module approximates u2i at the token level for inference. Experiments on three public benchmarks and two industrial datasets demonstrate that \DIG simultaneously improves ranking, retrieval, and unified retrieval-ranking quality.