Newest papers

Think Harder Where It Matters: Item-Level Adaptive Recursion for Sequential Recommendation

Zhanshuo Cao, Jiakai Tang, Yuqi Zhou, Sunhao Dai, Jun Xu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.30128 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

As famílias de padrões interativos mais recentes revelam que os padrões de padrões interativos mais baixos geralmente alocam um orçamento computacional fixo a cada item no histórico de interação de um usuário, independentemente da contribuição do item para a previsão final. Nós desafiamos este projeto de facto com duas descobertas de diagnóstico. (1) Uma necessidade de interpretação computacional não uniforme: os itens dentro da mesma sequência variam drasticamente em quanto computação eles precisam. A maioria converge após o processamento liso, enquanto uma minoria crítica só se estabiliza em profundidade total. (2) A suboptimalidade da alocação fixa: dado o acesso do oráculo ao item alvo, uma alocação de profundidade não uniforme supera consistentemente a fixa, indicando que a computação uniforme existente tanto desperdiça

Machine translation (Meta M2M100); the English original follows.

Abstract

Existing sequential recommenders usually allocate a fixed computational budget to every item in a user's interaction history, regardless of the item's contribution to the final prediction. We challenge this de facto design with two diagnostic findings. (1) Non-uniform computational need: items within the same sequence vary drastically in how much computation they require. Most converge after shallow processing while a critical minority only stabilizes at full depth. (2) Suboptimality of fixed allocation: given oracle access to the target item, a non-uniform depth assignment consistently outranks the fixed one, indicating that existing uniform computation both wastes capacity on already-converged items and sacrifices overall recommendation quality. These observations motivate a shift from fixed-depth computation to item-level adaptive reasoning. To this end, we propose LoopRec. LoopRec recursively applies a single weight-shared block and introduces a lightweight gate at each recursion boundary that makes a binary exit-or-continue decision for every item independently. Items whose representations have sufficiently matured exit early, while those requiring deeper reasoning continue to subsequent recursions. This design converts the fixed global depth into a learned, per-item property without auxiliary depth labels, and is readily adaptable to diverse backbone architectures. Extensive experiments on public benchmarks with four backbone families demonstrate that LoopRec consistently outperforms matched fixed-depth counterparts. Further analysis reveals that LoopRec's learned allocation exhibits interpretable patterns aligned with our post-hoc oracle analysis: lower-frequency items and more recent interactions naturally receive deeper computation. Our code is available at https://github.com/Elpmis117/LoopRec.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Recursion (computer science), Computer science, Block (permutation group theory), Computation, Oracle, Theoretical computer science

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange