- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
The paper introduces LEARNWEAK, a framework for specializing small computer-use agents by targeting their specific weaknesses. It identifies tasks where a stronger teacher succeeds but the student fails, generates practice tasks addressing those failures, and uses selective DPO to correct planning and execution errors. Across eight OSWorld domains, it improves EvoCUA-8B by 11.6 percentage points and OpenCUA-7B by 11.1 points using only a few dozen trajectories per domain. The key finding is that student-specific failure correction is more effective than scaling generic training data.
Critical strengths
- Good results on one dataset.
Critical weaknesses
- Only evaluated on OSWorld? Not sure how it generalize
- Not sure how this is new from previous papers on distillation. Maybe you should talk more on what previous papers did and how yours is different.
Questions for authors
- Can you run on more datasets?
- Please specify your diff from previous papers.