Expressiveness Limits of Autoregressive Semantic ID Generation in Generative Recommendation

• URL: http://arxiv.org/abs/2605.06331v1
• Date: Thu, 07 May 2026 14:25:10 GMT
• Title: Expressiveness Limits of Autoregressive Semantic ID Generation in Generative Recommendation
• Abstract summary: Generative recommendation (GR) models generate items by autoregressively producing a sequence of discrete tokens that jointly index the target item.<n>We propose Latte, a simple modification that injects a latent token before each semantic ID, reshaping the decoding space from a single tree into multiple latent-token-conditioned trees.
• Score: 37.7230289841301
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Generative recommendation (GR) models generate items by autoregressively producing a sequence of discrete tokens that jointly index the target item. However, this autoregressive generation process also induces a structured decoding space whose impact on model expressiveness remains underexplored. Specifically, token-by-token generation can be viewed as traversing a decoding tree induced by semantic ID tokens, where leaf nodes correspond to candidate items. We observe that the item probabilities produced by GR models are strongly correlated with this tree structure: items that are close in the tree tend to receive similar probabilities for any given user, making it difficult to distinguish among them based on user-specific preferences. We further show theoretically that such structural correlations prevent GR models from representing even simple patterns that can be well captured by conventional collaborative filtering models. To mitigate this issue, we propose Latte, a simple modification that injects a latent token before each semantic ID, reshaping the decoding space from a single tree into multiple latent-token-conditioned trees. This design creates multiple paths with varying tree distances between items, relaxing tree-induced probability coupling and yielding an average of 3.45% relative improvement on NDCG@10. Our code is available at https://github.com/hyp1231/Latte.

Related papers

• Tree Matching Networks for Natural Language Inference: Parameter-Efficient Semantic Understanding via Dependency Parse Trees [0.0]
  Tree Matching Networks (TMN) are able to leverage prior encoded information about relationships without having to learn them from scratch.<n>TMN is able to achieve significantly better results with a significantly reduced memory footprint and much less training time than the BERT based model on the SNLI task.
  arXiv  Detail & Related papers  (2025-11-28T21:06:11Z)
• CoFiRec: Coarse-to-Fine Tokenization for Generative Recommendation [55.783414010717074]
  CoFiRec is a novel generative recommendation framework that decomposes item information into semantic levels.<n>We show that CoFiRec outperforms existing methods, offering a new perspective for generative recommendation.
  arXiv  Detail & Related papers  (2025-11-27T18:59:35Z)
• Generating Long Semantic IDs in Parallel for Recommendation [29.97624755406803]
  We propose RPG, a lightweight framework for semantic ID-based recommendation.<n>We train the model to predict each token independently using a multi-token prediction loss.<n> Experiments show that scaling up semantic ID length to 64 enables RPG to outperform generative baselines.
  arXiv  Detail & Related papers  (2025-06-06T06:20:37Z)
• RASD: Retrieval-Augmented Speculative Decoding [5.3926068062773895]
  Speculative decoding accelerates inference in large language models (LLMs)<n>This paper proposes RASD (Retrieval-Augmented Speculative Decoding), which adopts retrieval methods to enhance model-based speculative decoding.
  arXiv  Detail & Related papers  (2025-03-05T12:10:14Z)
• C2T: A Classifier-Based Tree Construction Method in Speculative Decoding [9.663330370149428]
  Speculative decoding methods often face inefficiencies in the construction of token trees and the verification of candidate tokens.<n>We propose a novel method named C2T that adopts a lightweight classifier to generate and prune token trees dynamically.
  arXiv  Detail & Related papers  (2025-02-19T11:57:02Z)
• Efficient Link Prediction via GNN Layers Induced by Negative Sampling [86.87385758192566]
  Graph neural networks (GNNs) for link prediction can loosely be divided into two broad categories.<n>We propose a novel GNN architecture whereby the emphforward pass explicitly depends on emphboth positive (as is typical) and negative (unique to our approach) edges.<n>This is achieved by recasting the embeddings themselves as minimizers of a forward-pass-specific energy function that favors separation of positive and negative samples.
  arXiv  Detail & Related papers  (2023-10-14T07:02:54Z)
• Tree Variational Autoencoders [5.992683455757179]
  We propose a new generative hierarchical clustering model that learns a flexible tree-based posterior distribution over latent variables. TreeVAE hierarchically divides samples according to their intrinsic characteristics, shedding light on hidden structures in the data.
  arXiv  Detail & Related papers  (2023-06-15T09:25:04Z)
• SequenceMatch: Imitation Learning for Autoregressive Sequence Modelling with Backtracking [60.109453252858806]
  A maximum-likelihood (MLE) objective does not match a downstream use-case of autoregressively generating high-quality sequences. We formulate sequence generation as an imitation learning (IL) problem. This allows us to minimize a variety of divergences between the distribution of sequences generated by an autoregressive model and sequences from a dataset. Our resulting method, SequenceMatch, can be implemented without adversarial training or architectural changes.
  arXiv  Detail & Related papers  (2023-06-08T17:59:58Z)
• Complex Event Forecasting with Prediction Suffix Trees: Extended Technical Report [70.7321040534471]
  Complex Event Recognition (CER) systems have become popular in the past two decades due to their ability to "instantly" detect patterns on real-time streams of events. There is a lack of methods for forecasting when a pattern might occur before such an occurrence is actually detected by a CER engine. We present a formal framework that attempts to address the issue of Complex Event Forecasting.
  arXiv  Detail & Related papers  (2021-09-01T09:52:31Z)
• Rethinking Learnable Tree Filter for Generic Feature Transform [71.77463476808585]
  Learnable Tree Filter presents a remarkable approach to model structure-preserving relations for semantic segmentation. To relax the geometric constraint, we give the analysis by reformulating it as a Markov Random Field and introduce a learnable unary term. For semantic segmentation, we achieve leading performance (82.1% mIoU) on the Cityscapes benchmark without bells-and-whistles.
  arXiv  Detail & Related papers  (2020-12-07T07:16:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.