Pith. sign in

REVIEW 7 cited by

On Softmax Direct Preference Optimization for Recommendation

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

arxiv 2406.09215 v3 pith:7TJJLVA6 submitted 2024-06-13 cs.IR cs.AI

classification cs.IRcs.AI
keywords preferencelm-basedrecommenderslosss-dposoftmaxdatalanguage
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Recommender systems aim to predict personalized rankings based on user preference data. With the rise of Language Models (LMs), LM-based recommenders have been widely explored due to their extensive world knowledge and powerful reasoning abilities. Most of the LM-based recommenders convert historical interactions into language prompts, pairing with a positive item as the target response and fine-tuning LM with a language modeling loss. However, the current objective fails to fully leverage preference data and is not optimized for personalized ranking tasks, which hinders the performance of LM-based recommenders. Inspired by the current advancement of Direct Preference Optimization (DPO) in human preference alignment and the success of softmax loss in recommendations, we propose Softmax-DPO (S-DPO) to instill ranking information into the LM to help LM-based recommenders distinguish preferred items from negatives, rather than solely focusing on positives. Specifically, we incorporate multiple negatives in user preference data and devise an alternative version of DPO loss tailored for LM-based recommenders, which is extended from the traditional full-ranking Plackett-Luce (PL) model to partial rankings and connected to softmax sampling strategies. Theoretically, we bridge S-DPO with the softmax loss over negative sampling and find that it has an inherent benefit of mining hard negatives, which assures its exceptional capabilities in recommendation tasks. Empirically, extensive experiments conducted on three real-world datasets demonstrate the superiority of S-DPO to effectively model user preference and further boost recommendation performance while providing better rewards for preferred items. Our codes are available at https://github.com/chenyuxin1999/S-DPO.

Discussion (0). Sign in to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. LBR: Towards Mitigating Length Bias in Large Language Models for Recommendation

    cs.IR 2026-07 accept novelty 6.5 of 10

    LBR removes length bias in LLM recommenders via length-aware attention offsets and Trie-branching information-length normalization, improving accuracy and fairness with negligible cost.

  2. Hierarchical Residual Policy Optimization for Generative Recommendations

    cs.IR 2026-08 conditional novelty 6.0 of 10

    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 ...

  3. Architecture is All You Need: Improving LLM Recommenders by Dropping the Text

    cs.IR 2025-06 conditional novelty 6.0 of 10

    A 2.2M-parameter Llama-style model using only item IDs beats a 7B-parameter text-based LLM recommender and SASRec on MovieLens-10M.

  4. CORONA: A Coarse-to-Fine Framework for Graph-based Recommendation with Large Language Models

    cs.IR 2025-06 conditional novelty 6.0 of 10

    CORONA uses LLM-generated preference and intent queries to prune the interaction graph in two stages, then applies a GNN to the remaining subgraph, achieving state-of-the-art recommendation accuracy.

  5. OneSug: The Unified End-to-End Generative Framework for E-commerce Query Suggestion

    cs.IR 2025-06 conditional novelty 6.0 of 10

    A unified encoder-decoder model with prefix representation enhancement and reward-weighted DPO outperforms Kuaishou's online multi-stage query suggestion system in offline and live A/B evaluations.

  6. Listwise Preference Alignment Optimization for Tail Item Recommendation

    cs.IR 2025-07 reject novelty 4.0 of 10

    The paper applies a listwise softmax preference loss with head-item negative sampling and tail reweighting to improve long-tail item recommendation.

  7. GR-LLMs: Recent Advances in Generative Recommendation Based on Large Language Models

    cs.IR 2025-07 unverdicted novelty 3.0 of 10

    A survey of LLM-based generative recommendation systems, covering application settings, training pipelines, industrial deployment challenges, and future directions.

Pith tools