Search papers, labs, and topics across Lattice.
The paper introduces Self-Evolution Agent (SEA), a 7B parameter computer use agent, designed to improve autonomous computer operation. SEA incorporates an automatic pipeline for generating verifiable task trajectories, Efficient Step-wise Reinforcement Learning to reduce computational costs, and a grounding/planning integration method for model enhancement. Results show SEA outperforms existing models of similar size and achieves performance comparable to much larger models (32B/72B) on computer use tasks.
A 7B parameter agent can now rival the performance of models 4-10x larger at autonomous computer operation, thanks to innovations in data generation, reinforcement learning, and model enhancement.
Computer use agents represent an emerging area in artificial intelligence, aiming to operate computers autonomously to fulfill user tasks, attracting significant attention from both industry and academia. However, the performance of existing agents remains insufficient for practical deployment. In this paper, we propose the Self-Evolution Agent (SEA) for computer operation, alongside three core innovations in data generation, reinforcement learning, and model enhancement to develop this agent. Specifically, we first design an automatic pipeline to generate verifiable task trajectories for training. Second, we propose Efficient Step-wise Reinforcement Learning to reduce the substantial computational overhead of long-horizon training. Finally, we introduce a model enhancement method that integrates grounding and planning capabilities into a single model without additional training. Leveraging these innovations, our SEA (with only 7B parameters) outperforms existing models of the same parameter scale and achieves performance comparable to larger models (e.g., 32B/72B parameters) on computer use tasks. We plan to release the model weights and related code as open-source resources in the future.
