UniGRec: Unified Generative Recommendation with Soft Identifiers for End-to-End Optimization

TL;DR

UniGRec introduces a unified, end-to-end trainable generative recommendation framework that addresses key challenges like soft-to-hard mismatch, identifier collapse, and collaborative signal deficiency, leading to superior performance.

Contribution

It proposes a novel unified model with differentiable soft identifiers, annealed inference alignment, regularization for diversity, and distillation for collaborative signals, enabling fully end-to-end optimization.

Findings

Outperforms state-of-the-art baselines on real datasets.

Effectively addresses soft-hard inference gap.

Maintains diverse and balanced item identifiers.

Abstract

Generative recommendation has recently emerged as a transformative paradigm that directly generates target items, surpassing traditional cascaded approaches. It typically involves two components: a tokenizer that learns item identifiers and a recommender trained on them. Existing methods often decouple tokenization from recommendation or rely on asynchronous alternating optimization, limiting full end-to-end alignment. To address this, we unify the tokenizer and recommender under the ultimate recommendation objective via differentiable soft item identifiers, enabling joint end-to-end training. However, this introduces three challenges: training-inference discrepancy due to soft-to-hard mismatch, item identifier collapse from codeword usage imbalance, and collaborative signal deficiency due to an overemphasis on fine-grained token-level semantics. To tackle these challenges, we propose UniGRec, a unified generative recommendation framework that addresses them from three perspectives. UniGRec employs Annealed Inference Alignment during tokenization to smoothly bridge soft training and hard inference, a Codeword Uniformity Regularization to prevent identifier collapse and encourage codebook diversity, and a Dual Collaborative Distillation mechanism that distills collaborative priors from a lightweight teacher model to jointly guide both the tokenizer and the recommender. Extensive experiments on real-world datasets demonstrate that UniGRec consistently outperforms state-of-the-art baseline methods. Our codes are available at https://github.com/Jialei-03/UniGRec.