AsymRec decouples input and output representations in generative recommendation via multi-expert semantic projection and multi-faceted hierarchical quantization, outperforming prior models by 15.8% on average.
Canonical reference
Title resolution pending
Canonical reference. 80% of citing Pith papers cite this work as background.
23
Pith papers citing it
Background
80% of classified citations
citation-role summary
background 4
method 1
citation-polarity summary
representative citing papers
QCMP-CL introduces learnable collaborative sequence augmentation from same-target and similar sequences plus a quality-aware weighting mechanism based on augmentation confidence, outperforming prior CL-based sequential recommendation methods on three real-world datasets.
FAERec fuses collaborative ID embeddings with LLM semantic embeddings using adaptive gating and dual-level alignment to enhance tail-item sequential recommendations.
LLMAR applies LLM reasoning with a self-correction reflection loop to generate semantic user motives for tuning-free recommendations, showing up to 54.6% nDCG@10 gains on a sparse industrial dataset over trained baselines.
The paper proposes an item-aware attention mechanism with intra-item and inter-item layers to let LLMs capture item-level collaborative relations instead of only token-level ones.
GenRecEdit injects cold-start items into generative recommendation models via context-aware token editing and interference-reducing triggers, boosting cold-start accuracy while using only 9.5% of retraining time.
BEAR adds a beam-search-aware regularization to LLM fine-tuning for recommendations that forces positive-item tokens to rank in the top-B candidates at each decoding step to avoid premature pruning.
DIGER makes semantic IDs in generative recommendation differentiable via Gumbel noise and decay schedules, yielding consistent gains on public datasets by aligning indexing and recommendation losses.
ItemRAG augments LLM recommendation prompts with item-level retrievals that blend semantic and co-purchase signals, outperforming user-history RAG in both standard and cold-start settings.
Long-Term Embeddings anchor sequential recommendation models to fixed content-based item representations to capture stable preferences and ensure version compatibility, resulting in uplifts in user engagement and financial metrics.
CoDiS applies variational context adjustment, expert isolation, and adversarial disentanglement to separate domain-shared and domain-specific preferences in cross-domain sequential recommendation, outperforming baselines on three datasets.
ReAd retrieves collaboratively similar items, builds an augmentation embedding via a lightweight module, and fuses it to refine sequential recommendation predictions, outperforming baselines on five datasets.
FAVE replaces multi-step flow generation with a learned global average velocity from a semantic anchor prior, delivering SOTA accuracy and roughly 10x faster inference on recommendation benchmarks.
FLAME condenses ensemble diversity into a single network via modular ensemble simulation and guided mutual learning during training, delivering ensemble-level performance with single-network inference speed on sequential recommendation tasks.
BIPCL improves sequential recommendation accuracy by bilaterally injecting collective intent prototypes into representations and enforcing contrastive alignment via bounded embedding perturbations.
ADS-POI decomposes user mobility sequences into multiple parallel evolving latent sub-states with context-conditioned aggregation to improve next POI recommendation accuracy.
CaST-POI improves next POI recommendation by conditioning user history attention on each candidate and adding candidate-relative temporal and spatial biases.
SpecTran applies a spectral-aware transformer adapter with learnable position encoding to aggregate informative components across the full spectrum of LLM embeddings, yielding 9.17% average gains on sequential recommendation tasks.
AsarRec learns adaptive sequence augmentations via transformation matrices and Semi-Sinkhorn projection to improve robustness of self-supervised sequential recommenders under noise.
Douyin deploys stacked target-to-history cross attention and request-level batching to scale end-to-end recommendation modeling to 10k-length histories, observing scaling-law gains and live engagement improvements.
SDA uses structural alignment as a soft teacher and gated low-rank expert paths to adapt LVLMs for multimodal recommendation, reporting 6.15% Hit@10 and 8.64% NDCG@10 average gains plus larger long-tail improvements on Amazon datasets.
HaNoRec dynamically weights harder preference samples and applies Gaussian perturbations to output distributions to improve multimodal LLM performance on sequential recommendation tasks.
OneRec unifies retrieval and ranking in a generative recommender using session-wise decoding and iterative DPO-based preference alignment, achieving real-world gains on Kuaishou.
citing papers explorer
-
Asymmetric Generative Recommendation via Multi-Expert Projection and Multi-Faceted Hierarchical Quantization
AsymRec decouples input and output representations in generative recommendation via multi-expert semantic projection and multi-faceted hierarchical quantization, outperforming prior models by 15.8% on average.
-
Quality-Aware Collaborative Multi-Positive Contrastive Learning for Sequential Recommendation
QCMP-CL introduces learnable collaborative sequence augmentation from same-target and similar sequences plus a quality-aware weighting mechanism based on augmentation confidence, outperforming prior CL-based sequential recommendation methods on three real-world datasets.
-
Fusion and Alignment Enhancement with Large Language Models for Tail-item Sequential Recommendation
FAERec fuses collaborative ID embeddings with LLM semantic embeddings using adaptive gating and dual-level alignment to enhance tail-item sequential recommendations.
-
LLMAR: A Tuning-Free Recommendation Framework for Sparse and Text-Rich Industrial Domains
LLMAR applies LLM reasoning with a self-correction reflection loop to generate semantic user motives for tuning-free recommendations, showing up to 54.6% nDCG@10 gains on a sparse industrial dataset over trained baselines.
-
Beyong Tokens: Item-aware Attention for LLM-based Recommendation
The paper proposes an item-aware attention mechanism with intra-item and inter-item layers to let LLMs capture item-level collaborative relations instead of only token-level ones.
-
GenRecEdit: Adapting Model Editing for Generative Recommendation with Cold-Start Items
GenRecEdit injects cold-start items into generative recommendation models via context-aware token editing and interference-reducing triggers, boosting cold-start accuracy while using only 9.5% of retraining time.
-
BEAR: Towards Beam-Search-Aware Optimization for Recommendation with Large Language Models
BEAR adds a beam-search-aware regularization to LLM fine-tuning for recommendations that forces positive-item tokens to rank in the top-B candidates at each decoding step to avoid premature pruning.
-
Differentiable Semantic ID for Generative Recommendation
DIGER makes semantic IDs in generative recommendation differentiable via Gumbel noise and decay schedules, yielding consistent gains on public datasets by aligning indexing and recommendation losses.
-
ItemRAG: Item-Based Retrieval-Augmented Generation for LLM-Based Recommendation
ItemRAG augments LLM recommendation prompts with item-level retrievals that blend semantic and co-purchase signals, outperforming user-history RAG in both standard and cold-start settings.
-
Long-Term Embeddings for Balanced Personalization
Long-Term Embeddings anchor sequential recommendation models to fixed content-based item representations to capture stable preferences and ensure version compatibility, resulting in uplifts in user engagement and financial metrics.
-
Context-Aware Disentanglement for Cross-Domain Sequential Recommendation: A Causal View
CoDiS applies variational context adjustment, expert isolation, and adversarial disentanglement to separate domain-shared and domain-specific preferences in cross-domain sequential recommendation, outperforming baselines on three datasets.
-
Retrieve-then-Adapt: Retrieval-Augmented Test-Time Adaptation for Sequential Recommendation
ReAd retrieves collaboratively similar items, builds an augmentation embedding via a lightweight module, and fuses it to refine sequential recommendation predictions, outperforming baselines on five datasets.
-
FAVE: Flow-based Average Velocity Establishment for Sequential Recommendation
FAVE replaces multi-step flow generation with a learned global average velocity from a semantic anchor prior, delivering SOTA accuracy and roughly 10x faster inference on recommendation benchmarks.
-
FLAME: Condensing Ensemble Diversity into a Single Network for Efficient Sequential Recommendation
FLAME condenses ensemble diversity into a single network via modular ensemble simulation and guided mutual learning during training, delivering ensemble-level performance with single-network inference speed on sequential recommendation tasks.
-
BIPCL: Bilateral Intent-Enhanced Sequential Recommendation via Embedding Perturbation Contrastive Learning
BIPCL improves sequential recommendation accuracy by bilaterally injecting collective intent prototypes into representations and enforcing contrastive alignment via bounded embedding perturbations.
-
ADS-POI: Agentic Spatiotemporal State Decomposition for Next Point-of-Interest Recommendation
ADS-POI decomposes user mobility sequences into multiple parallel evolving latent sub-states with context-conditioned aggregation to improve next POI recommendation accuracy.
-
CaST-POI: Candidate-Conditioned Spatiotemporal Modeling for Next POI Recommendation
CaST-POI improves next POI recommendation by conditioning user history attention on each candidate and adding candidate-relative temporal and spatial biases.
-
SpecTran: Spectral-Aware Transformer-based Adapter for LLM-Enhanced Sequential Recommendation
SpecTran applies a spectral-aware transformer adapter with learnable position encoding to aggregate informative components across the full spectrum of LLM embeddings, yielding 9.17% average gains on sequential recommendation tasks.
-
AsarRec: Adaptive Sequential Augmentation for Robust Self-supervised Sequential Recommendation
AsarRec learns adaptive sequence augmentations via transformation matrices and Semi-Sinkhorn projection to improve robustness of self-supervised sequential recommenders under noise.
-
Make It Long, Keep It Fast: End-to-End 10k-Sequence Modeling at Billion Scale on Douyin Recommendation
Douyin deploys stacked target-to-history cross attention and request-level batching to scale end-to-end recommendation modeling to 10k-length histories, observing scaling-law gains and live engagement improvements.
-
Structural and Disentangled Adaptation of Large Vision Language Models for Multimodal Recommendation
SDA uses structural alignment as a soft teacher and gated low-rank expert paths to adapt LVLMs for multimodal recommendation, reporting 6.15% Hit@10 and 8.64% NDCG@10 average gains plus larger long-tail improvements on Amazon datasets.
-
Multimodal Large Language Models with Adaptive Preference Optimization for Sequential Recommendation
HaNoRec dynamically weights harder preference samples and applies Gaussian perturbations to output distributions to improve multimodal LLM performance on sequential recommendation tasks.
-
OneRec: Unifying Retrieve and Rank with Generative Recommender and Iterative Preference Alignment
OneRec unifies retrieval and ranking in a generative recommender using session-wise decoding and iterative DPO-based preference alignment, achieving real-world gains on Kuaishou.