Sequential recommendation (SR) models capture continuously observed behavior, but a returning user may have no interactions for months or years. We define this setting as Zero-Observation Reactivation: the user has a pre-gap history, while the platform observes no behavioral signals during a macro-gap Delta t.
arXiv:2607. 19802v1 Announce Type: cross Abstract: Sequential recommendation (SR) models capture continuously observed behavior, but a returning user may have no interactions for months or years.
arXiv:2603. 22016v3 Announce Type: replace-cross Abstract: Large Reasoning Models (LRMs) often reach a correct solution before their long Chain-of-Thought trace ends, yet continue with redundant verification, repeated attempts, or unnecessary exploration that wastes computation and can even overturn the correct answer.
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.
The paper presents a strictly causal streaming video anomaly detector that updates a fixed‑size state in constant time per frame, eliminating the need for clip buffering or lookahead. Its core is a diagonal linear state‑space recurrence with a decay gate, trained via self‑supervised next‑embedding prediction on a frozen visual backbone. The authors derive a closed‑form link between the recurrence’s decay spectrum and detection delay, validate on UCSD Ped2 and CUHK Avenue, and report real‑time latency on Apple M3 Pro hardware (≈0.75 ms per frame).
SparseGDN is a hardware‑aware framework that makes Gated Delta Networks (GDNs) more efficient for long‑context sequence modeling. It combines structured intra‑chunk approximation with an input‑dependent recurrent‑event selector, producing outputs for every token while reducing dense computation. The GPU implementation updates only selected events and reconstructs token outputs from compact state snapshots, cutting computation and memory traffic.
The paper introduces CHASE, a cache‑hole‑adapted skip‑exit mechanism for looped state‑space language models, specifically Looped Mamba and Looped Hybrid Mamba‑Transformer. It shows that looping these architectures improves performance on controlled reasoning tasks and remains competitive in pre‑training benchmarks while using fewer distinct parameters. The cache‑hole adaptation allows selective skipping of recurrent steps during inference, maintaining perplexity close to full computation and achieving significant speedups.
arXiv:2609.08189v1 Announce Type: new Abstract: Dynamic layer routing reduces the inference cost of Large Language Models (LLMs) by learning to skip layers for individual tokens. Existing methods, ho...
arXiv:2609.40063v1 Announce Type: cross Abstract: Low-Rank Adaptive Residual Connections (LARC) give a frozen model a compact numerical state that can learn from feedback. The map $h+BAh$ adds a low-...
BehaveRec is a three-signal framework that combines a causal session transformer, a bidirectional User History Encoder over a 90‑day window, and a GraphSAGE module built from behavioral embedding similarity. On an industrial dataset of 137M sequences, it improves R@5 by +3.2% over SASRec, and its advantage grows with user history length, becoming significant for users with 101+ interactions. The study also establishes a practical deployment rule of roughly 20 interactions per user for when long‑term history and behavioral‑graph signals are worthwhile.
BF1 is a deterministic block‑aligned dyadic sparse‑attention retrofit designed to reduce the cost of causal attention in long‑context transformers. It combines a small exact local neighborhood, a global first block, and logarithmically spaced historical blocks, achieving O(n log n) token interactions per layer with O(log n) communication depth. On an NVIDIA RTX PRO 6000 Blackwell GPU, BF1 outperforms dense attention for 2K–4K tokens and delivers up to a 10.91× prefill speedup at 32K tokens, while retrofitting eight of 28 Qwen3‑0.6B layers reduces first‑token latency by up to 15.3% at 32K tokens and yields the lowest perplexity among compared sparse and dense training protocols.
The paper introduces recirculation, an inference‑time architectural enhancement for foundation models that reduces perplexity and improves accuracy on generation and reasoning tasks without adding significant latency. Recirculation adds a specific form of recurrence, enabling the model to function as a dynamical system that tracks belief states, and is distinct from chain‑of‑thought or depth‑recurrence methods. An adaptive variant requires minimal hyperparameter tuning and achieves notable gains on the Gemma3 family, including a 23% perplexity drop and a 21% accuracy increase on GSM8k.
arXiv:2604. 24806v2 Announce Type: replace-cross Abstract: Modern Deep Learning Recommendation Models (DLRMs) follow scaling laws with sequence length, driving the frontier toward ultra-long User Interaction History (UIH).