吳明輝先生擁有20年軟件工程開(kāi)發(fā)和算法研究經(jīng)驗(yàn)、50余項(xiàng)國(guó)內(nèi)外發(fā)明專(zhuān)利、12年企業(yè)級(jí)服務(wù)領(lǐng)域創(chuàng)業(yè)經(jīng)驗(yàn)。他是首位獲邀在2014年IAB全球峰會(huì)上演講的中國(guó)企業(yè)家，并在2015年入選Campaign“TheA-List Digital China2015”名人堂，2017年4月入圍人工智能創(chuàng)業(yè)30人。吳明輝先生同時(shí)是一位出色的連續(xù)創(chuàng)業(yè)者，成功創(chuàng)立兩家獨(dú)角獸公司。秉持用AI技術(shù)賦能行業(yè)的使命感，他致力于帶領(lǐng)明略科技走出屬于中國(guó)自己的企業(yè)級(jí)服務(wù)道路。
Graphical user interfaces (GUIs) are the primary medium for human-computer interaction, yet automating GUI interactions remains challenging due to the complexity of visual elements, dynamic environments, and the need for multi-step reasoning. Existing methods based on vision-language models (VLMs) often suffer
from limited resolution, domain mismatch, and insufficient sequential decisionmaking capability. To address these issues, we propose Mano, a robust GUI agent built upon a multi-modal foundation model pre-trained on extensive web and computer system data. Our approach integrates a novel simulated environment for high-fidelity data generation, a three-stage training pipeline (supervised fine-tuning, offline reinforcement learning, and online reinforcement learning), and a verification module for error recovery. Mano demonstrates state-of-the-art
performance on multiple GUI benchmarks, including Mind2Web and OSWorld, achieving significant improvements in success rate and operational accuracy. Our work provides new insights into the effective integration of reinforcement learning with VLMs for practical GUI agent deployment, highlighting the importance of
domain-specific data, iterative training, and holistic reward design.
數(shù)字廣告測(cè)量及驗(yàn)證
一站式社媒營(yíng)銷(xiāo)監(jiān)測(cè)與洞察平臺(tái)
映會(huì)0918-Launch-Poster.jpg)
報(bào)告.jpg)
