Fazle Faisal
Research Subjects
Fazle Faisal has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Adriana Szekeres1 shared publication
- Hongbin Zhong
Georgia Institute of Technology Duke University VMware Research VMware Research Georgia Institute of Technology
1 shared publication - Kexin Rong
Georgia Institute of Technology
1 shared publication - Luis França1 shared publication
- Suman Nath
Microsoft Research Redmond WA
1 shared publication Tanakorn Leesatapornwongsa
1 shared publication
Latest Publications
- Open Access
ActionEngine: From Reactive to Programmatic GUI Agents via State Machine Memory
Existing Graphical User Interface (GUI) agents operate through step-by-step calls to vision language models--taking a screenshot, reasoning about the next action, executing it, then repeating on the new page--resulting in high costs and latency that scale with the number of reasoning steps, and limited accuracy due to no persistent memory of previously visited pages. We propose ActionEngine, a training-free framework that transitions from reactive execution to programmatic planning through a novel two-agent architecture: a Crawling Agent that constructs an updatable state-machine memory of the GUIs through offline exploration, and an Execution Agent that leverages this memory to synthesize complete, executable Python programs for online task execution. To ensure robustness against evolving interfaces, execution failures trigger a vision-based re-grounding fallback that repairs the failed action and updates the memory. This design drastically improves both efficiency and accuracy: on Reddit tasks from the WebArena benchmark, our agent achieves 95% task success with on average a single LLM call, compared to 66% for the strongest vision-only baseline, while reducing cost by 11.8x and end-to-end latency by 2x. Together, these components yield scalable and reliable GUI interaction by combining global programmatic planning, crawler-validated action templates, and node-level execution with localized validation and repair.
Open AccessActionEngine: From Reactive to Programmatic GUI Agents via State Machine Memory
Existing Graphical User Interface (GUI) agents operate through step-by-step calls to vision language models--taking a screenshot, reasoning about the next action, executing it, then repeating on the new page--resulting in high costs and latency that scale with the number of reasoning steps, and limited accuracy due to no persistent memory of previously visited pages. We propose ActionEngine, a training-free framework that transitions from reactive execution to programmatic planning through a novel two-agent architecture: a Crawling Agent that constructs an updatable state-machine memory of the GUIs through offline exploration, and an Execution Agent that leverages this memory to synthesize complete, executable Python programs for online task execution. To ensure robustness against evolving interfaces, execution failures trigger a vision-based re-grounding fallback that repairs the failed action and updates the memory. This design drastically improves both efficiency and accuracy: on Reddit tasks from the WebArena benchmark, our agent achieves 95% task success with on average a single LLM call, compared to 66% for the strongest vision-only baseline, while reducing cost by 11.8x and end-to-end latency by 2x. Together, these components yield scalable and reliable GUI interaction by combining global programmatic planning, crawler-validated action templates, and node-level execution with localized validation and repair.