ClockRoPE: Random Fourier Rotations for Temporal Routine Modeling
arXiv:2607. 26369v1 Announce Type: new Abstract: Rotary Position Embedding (RoPE) has been widely adopted in transformer-based large language models.
T‑RoPE introduces a time‑aware Rotary Position Embedding for sequential recommendation, replacing index‑only rotations with timestamp‑based angles, learnable temporal coefficients, multiscale frequency banks, shifted query alignment, and non‑stationary key rotation. The authors prove that standard RoPE is time‑translation invariant and cannot capture seasonal contexts, while T‑RoPE breaks this invariance while preserving the RoPE interface. Across five public benchmarks and an industrial‑scale e‑commerce dataset, T‑RoPE outperforms baselines by up to 130 % in HR@10 and delivers significant online lift in conversion and order count.
arXiv:2607. 26369v1 Announce Type: new Abstract: Rotary Position Embedding (RoPE) has been widely adopted in transformer-based large language models.
The paper introduces ReST, a recommendation‑native Transformer scaling framework designed to handle noisy, irregular, and sparsely supervised user behavior sequences in production ranking. ReST employs a dual‑gated attention encoder with rotary positional and temporal embeddings, and a lightweight cross decoder that decouples heavy encoding from fast decoding, enabling efficient compute‑once, decode‑many‑times ranking. Experiments on industrial and public benchmarks show that ReST outperforms traditional Transformer blocks, achieving higher accuracy and consistent scaling across sequence length, depth, and width, and a one‑week online A/B test on a production advertising platform yielded a 1.31% AUC lift and an 11.93% increase in a core revenue metric within a 50 ms P99 latency budget.
Positional encoding is a fundamental component of Transformer-based generative recommendation models, where user histories are modeled as autoregressive item sequences. Most positional encoding methods are inherited from natural language processing and mainly represent discrete item order.
arXiv:2608. 16274v1 Announce Type: cross Abstract: Positional encoding is a fundamental component of Transformer-based generative recommendation models, where user histories are modeled as autoregressive item sequences.
arXiv:2610.27691v1 Announce Type: new Abstract: Modern industrial ranking systems increasingly adopt generative recommendation (GR) paradigms, in which user behavior sequences and non-sequential (NS)...
DeGRe is a dense‑supervised generative reranking framework designed to improve multi‑stage recommender systems by addressing label bias and credit assignment issues. It uses an offline Lookahead Evaluator with beam search to generate dense supervision signals, which are distilled into a lightweight Online Generator that can perform efficient greedy decoding at inference time. Experiments show that DeGRe outperforms baselines on public benchmarks and industrial datasets, and it has been successfully deployed on Taobao Flash Shopping to enhance online recommendations.
The paper introduces DSRec, a dual‑interest sequential recommendation model that separates item representations into long‑term and short‑term semantic contexts. Long‑term embeddings capture stable preferences through historical aggregation, while short‑term embeddings focus on local session intent modulated by inter‑click time intervals. Each branch is processed by a distinct State Space Model— a full‑sequence Mamba for long‑term modeling and a time‑modulated SSM for short‑term dynamics— and a residual cross‑fusion mechanism aligns the two granularities while preserving their independence. Experiments on public benchmarks show that DSRec outperforms state‑of‑the‑art methods.
SequenceO1 is an end‑to‑end framework that enables ultra‑long (up to 100K interactions) sequence modeling for recommendation systems. It compresses raw user histories into a fixed‑size sketch using Sketch Attention and then models short‑term and long‑term interests with Target‑to‑History Cross Attention. The system incorporates low‑rank caching, batching, pipeline lift, and a FlashSA kernel to keep training and inference efficient, achieving consistent offline and online performance gains when deployed at full traffic on Douyin.
Large language models (LLMs) have recently emerged as powerful backbones for recommender systems by reformulating recommendation as a token-level generation task. Despite their promise, we identify a pervasive yet underexplored issue: $\textit{Length Bias}$.
arXiv:2603.02561v2 Announce Type: replace-cross Abstract: Attention mechanism remains the defining operator in Transformers since it provides expressive global credit assignment, yet its quadratic co...
The paper introduces LatentRec, a framework that replaces inference-time chain-of-thought reasoning in large language model (LLM) recommender systems with training-time per-step supervision. It employs a 2.5 M-parameter Continuous Thought Module, Soft Token Selection, and a Per-step Reasoning Loss to guide intermediate hidden states toward the target semantic ID. After training, the module can be removed, yielding a 4–5 % relative improvement in HR@10 on Amazon benchmarks while reducing inference latency by 18.3×.
arXiv:2610.15864v1 Announce Type: cross Abstract: Large-scale personalization is shifting from embedding-table retrieval to LLM-based recommenders that generate the Semantic ID (SID) of the item to r...