Beyond Item IDs: Scaling Short-Form-Video Recommendation via Semantic-Native Long Sequence Modeling

• URL: http://arxiv.org/abs/2606.07546v1
• Date: Mon, 04 May 2026 00:00:53 GMT
• Title: Beyond Item IDs: Scaling Short-Form-Video Recommendation via Semantic-Native Long Sequence Modeling
• Abstract summary: We present a production-deployed framework for modeling ultra-long user behavior sequences at a billion-user scale.<n>We first address the representation bottleneck by adopting content-native Semantic IDs.<n>Second, to overcome the sequence scaling barrier, we introduce a Global-Aware Compression Transformer.
• Score: 13.58338102892064
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Capturing user interests across extensive watch histories is critical for short-form video recommendation, yet scaling sequence length is limited by two bottlenecks: the semantic sparsity of atomic Video IDs and the quadratic computational complexity of Transformers. Traditional orthogonal Video IDs fail to capture content relationships and demand large embedding tables, while the quadratic complexity of self-attention restricts the maximum sequence length under strict industrial latency and resource constraints. In this work, we present a production-deployed framework for modeling ultra-long user behavior sequences at a billion-user scale. We first address the representation bottleneck by adopting content-native Semantic IDs. By utilizing depth-truncated, coarse-grained Semantic IDs, we shrink the embedding table size from corpus cardinality. This compact representation naturally generalizes to cold-start content through shared semantic prefixes. Second, to overcome the sequence scaling barrier, we introduce a Global-Aware Compression Transformer that leverages non-parametric temporal folding and unified global query integration to effectively condense the sequence, alleviating both the memory and computational bottlenecks of standard self-attention. Offline profiling on our computing infrastructure demonstrates an order-of-magnitude reduction in peak memory footprint and a drastic decrease in computational overhead. This efficiency gain enables supporting longer sequence lengths at an affordable cost in production, yielding substantial online gains in satisfied user engagement and satisfied content consumption in large-scale online A/B tests.

Related papers

• SequenceO1: End-to-End Ultra-Long (100K) Sequence Modeling in Recommendation with Low-Rank Caching [17.335538522124555]
  SequenceO1 is an end-to-end framework for ultra-long user behavior sequence modeling.<n>It is deployed at full traffic on Douyin with histories of up to 100K interactions.<n>Results show consistent offline and online gains, while sketch retains most of the benefit of directly scaling end-to-end sequence ranking to 100K.
  arXiv  Detail & Related papers  (2026-09-08T08:49:44Z)
• Adaptive Tokenisation Via Temporal Redundancy Masking And Latent Inpainting [18.71433885255431]
  Adaptive video tokenisation seeks to dynamically allocate token budgets based on the underlying visual complexity of a sequence.<n>We show that the latent space of a frozen continuous video tokeniser inherently encodes temporal redundancy that can be exploited directly.<n>We introduce a parameter-free adaptive token allocation mechanism that applies a fixed threshold to per-position temporal-L1 differences.
  arXiv  Detail & Related papers  (2026-06-04T13:31:12Z)
• Seeing Further and Wider: Joint Spatio-Temporal Enlargement for Micro-Video Popularity Prediction [24.227528430107114]
  Micro-video popularity prediction (MVPP) aims to forecast the future popularity of videos on online media.<n>It is critical for MVPP approaches to understand both the temporal dynamics of a given video (temporal) and its historical relevance to other videos (spatial)
  arXiv  Detail & Related papers  (2026-04-22T08:11:06Z)
• S2GS: Streaming Semantic Gaussian Splatting for Online Scene Understanding and Reconstruction [57.07346645250984]
  Streaming Semantic Gaussian Splatting (S2GS) is a strictly causal, incremental 3D Gaussian semantic field framework.<n>It does not leverage future frames and continuously updates scene geometry, appearance, and instance-level semantics without reprocessing historical frames.<n>S2GS matches or outperforms strong offline baselines on joint reconstruction-and-understanding benchmarks.
  arXiv  Detail & Related papers  (2026-03-15T05:48:55Z)
• GEMs: Breaking the Long-Sequence Barrier in Generative Recommendation with a Multi-Stream Decoder [54.64137490632567]
  We propose a novel and unified framework designed to capture users' sequences from long-term history.<n>Generative Multi-streamers ( GEMs) break user sequences into three streams.<n>Extensive experiments on large-scale industrial datasets demonstrate that GEMs significantly outperforms state-the-art methods in recommendation accuracy.
  arXiv  Detail & Related papers  (2026-02-14T06:42:56Z)
• Mixture of Contexts for Long Video Generation [72.96361488755986]
  We recast long-context video generation as an internal information retrieval task.<n>We propose a simple, learnable sparse attention routing module, Mixture of Contexts (MoC), as an effective long-term memory retrieval engine.<n>As we scale the data and gradually sparsify the routing, the model allocates compute to salient history, preserving identities, actions, and scenes over minutes of content.
  arXiv  Detail & Related papers  (2025-08-28T17:57:55Z)
• LoViC: Efficient Long Video Generation with Context Compression [68.22069741704158]
  We introduce LoViC, a DiT-based framework trained on million-scale open-domain videos.<n>At the core of our approach is FlexFormer, an expressive autoencoder that jointly compresses video and text into unified latent representations.
  arXiv  Detail & Related papers  (2025-07-17T09:46:43Z)
• CItruS: Chunked Instruction-aware State Eviction for Long Sequence Modeling [52.404072802235234]
  We introduce Chunked Instruction-aware State Eviction (CItruS), a novel modeling technique that integrates the attention preferences useful for a downstream task into the eviction process of hidden states. Our training-free method exhibits superior performance on long sequence comprehension and retrieval tasks over several strong baselines under the same memory budget.
  arXiv  Detail & Related papers  (2024-06-17T18:34:58Z)
• Lean Attention: Hardware-Aware Scalable Attention Mechanism for the Decode-Phase of Transformers [4.674454841332859]
  Transformer-based models have emerged as one of the most widely used architectures for natural language processing.<n>These huge models are memory hungry and incur significant inference latency even on cutting edge AI-accelerators.<n>We propose LeanAttention, a scalable technique of computing self-attention for the token-generation phase.
  arXiv  Detail & Related papers  (2024-05-17T00:52:39Z)
• Cluster-Former: Clustering-based Sparse Transformer for Long-Range Dependency Encoding [90.77031668988661]
  Cluster-Former is a novel clustering-based sparse Transformer to perform attention across chunked sequences. The proposed framework is pivoted on two unique types of Transformer layer: Sliding-Window Layer and Cluster-Former Layer. Experiments show that Cluster-Former achieves state-of-the-art performance on several major QA benchmarks.
  arXiv  Detail & Related papers  (2020-09-13T22:09:30Z)

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.