LLaDA-Rec: Discrete Diffusion for Parallel Semantic ID Generation in Generative Recommendation

TL;DR

LLaDA-Rec introduces a parallel, bidirectional diffusion-based framework for recommendation, overcoming limitations of autoregressive models by modeling dependencies more effectively and reducing error propagation.

Contribution

It proposes a novel discrete diffusion approach with bidirectional attention and adaptive order generation for improved recommendation performance.

Findings

Outperforms existing ID-based and generative recommenders on real datasets.

Effectively models inter-item and intra-item dependencies.

Reduces error accumulation in sequence generation.

Abstract

Generative recommendation represents each item as a semantic ID, i.e., a sequence of discrete tokens, and generates the next item through autoregressive decoding. While effective, existing autoregressive models face two intrinsic limitations: (1) unidirectional constraints, where causal attention restricts each token to attend only to its predecessors, hindering global semantic modeling; and (2) error accumulation, where the fixed left-to-right generation order causes prediction errors in early tokens to propagate to the predictions of subsequent token. To address these issues, we propose LLaDA-Rec, a discrete diffusion framework that reformulates recommendation as parallel semantic ID generation. By combining bidirectional attention with the adaptive generation order, the approach models inter-item and intra-item dependencies more effectively and alleviates error accumulation. Specifically, our approach comprises three key designs: (1) a parallel tokenization scheme that produces semantic IDs for bidirectional modeling, addressing the mismatch between residual quantization and bidirectional architectures; (2) two masking mechanisms at the user-history and next-item levels to capture both inter-item sequential dependencies and intra-item semantic relationships; and (3) an adapted beam search strategy for adaptive-order discrete diffusion decoding, resolving the incompatibility of standard beam search with diffusion-based generation. Experiments on three real-world datasets show that LLaDA-Rec consistently outperforms both ID-based and state-of-the-art generative recommenders, establishing discrete diffusion as a new paradigm for generative recommendation.