NeFut Logo NeFut
Admin Login

[CS.AI] EchoPath: Execution-Level Replayable Memory for GUI Agents

Published at: 2026-09-16 22:00 Last updated: 2026-09-18 00:46
#algorithm #AI #Machine Learning

EchoPath is a model‑agnostic harness that converts artifact‑validated GUI trajectories into standardized, parameter‑controlled callable memories. Each memory stores task‑intent keys, application and state preconditions, flexible input parameters, visual GUI evidence, validation provenance, and lifecycle state, allowing the host agent to invoke a procedure only when it can be deterministically replayed in the current runtime. The core mechanism is an image‑based target‑reaiming algorithm that treats stored coordinates as visual evidence, matches the remembered GUI target against the current screen, and emits corrected operation coordinates before execution. During replay, EchoPath rebinds only declared modifiable inputs and rejects ambiguous or incompatible steps, falling back to bounded grounding repair or fresh planning. Experiments on real computer‑use tasks show that EchoPath reduces median token cost by more than 90% and median execution time by about 60%. These results support a bounded form of enterprise GUI memory: validated execution experience can become a controllable callable asset for recurrent work.

Review: EchoPath dramatically improves the efficiency of GUI‑based tasks, illustrating the promise of controllable memory for enterprise automation.

Original Source: https://arxiv.org/abs/2609.16635

[h] Back to Home