Newest papers

Tandem: Serving Hybrid Linear-Attention GUI Agents on Mobile NPUs

Zhuomin Chen, Jingchao Ni, Xu Zheng, Yong Chen, Mo Sha, Wei Cheng, Dongsheng Luo · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26660 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

Las unidades de procesamiento neural móviles (NPUs) permiten la ejecución local, manteniendo el contenido de la pantalla en el teléfono, pero la latencia del modelo se acumula con cada acción. Los modelos de atención lineal híbrido reducen los requisitos de memoria, pero la reutilización de su estado recurrente requiere puntos de control en las posiciones de entrada apropiadas. En las solicitudes del agente se retienen las instrucciones pero se reescribe el historial reciente, por lo que las políticas del punto de control diseñadas para la configuración de chat pueden perder prefixos reutilizables, pero la latencia del modelo se acumula con cada acción. Los modelos de atención lineal híbridos también pueden mover el procesamiento prompt de la unidad de matriz de la NPU, añadiendo sobrecargas que superan los ahorros de decodif

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

Abstract

Smartphone GUI agents complete tasks in mobile apps by repeatedly reading the screen and asking a model for the next action. Mobile neural processing units (NPUs) enable local execution, keeping screen contents on the phone, but model latency accumulates with every action. Hybrid linear attention models reduce memory requirements, yet reusing their recurrent state requires checkpoints at the appropriate input positions. Agent requests retain instructions but rewrite recent history, so checkpoint policies designed for chat continuation can miss reusable prefixes. Retaining rollback states for speculative decoding can also move prompt processing off the NPU's matrix unit, adding overhead that outweighs decoding savings on short action outputs. A loaded model can further hold the NPU awake between tasks. We present Tandem, a runtime for hybrid GUI agents on mobile NPUs. Tandem saves recurrent state where consecutive prompts diverge, allowing later requests to reuse previously processed context. It keeps most prefill on the matrix unit and produces rollback states through a short per-token tail. Its verification kernel reads each weight block once for all proposed tokens. Between request bursts, Tandem releases NPU power votes without unloading the model. We implement Tandem in llama.cpp and evaluate it on recorded agent interactions across two Snapdragon generations. On a Snapdragon 8 Elite, Tandem reduces per-step latency by 30% for text-based screen descriptions and 16% for screenshots relative to llama.cpp's fastest NPU configuration, reduces energy per text step by 33%, and lowers standby power with the model loaded from 0.76 to 0.22 W.

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

Topics: Computer science, Latency (audio), Operating system, Decoding methods, Reuse, Rollback

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