TeleTune improves software agent skills by learning from offline user logs
TeleTune: Evolving Agent Skills From Offline Telemetry
Artificial IntelligenceComputation and Language
Summary
Teaching computer agents how people use software is hard because logs don’t clearly say what users want, can’t be replayed to check learning, and mix different tasks together. The paper’s authors created TeleTune, a system that learns common skills from these messy logs by testing which changes best predict user actions. This helps the agent get better at understanding and performing tasks based on past user behavior, without needing to try things out live. Experiments show TeleTune works better than previous methods even when the data is noisy or mixed up.
What this means in practice
- •For software developers: Improve software assistants by teaching them reusable skills from existing user activity logs without requiring real-time interaction.
- •For technical support teams: Automatically retrieve relevant user demonstrations covering subgoals to guide help agents in solving user-reported issues efficiently.
Authors
Justin Chih-Yao Chen, Elias Stengel-Eskin, Yan Chen, Pol Llado, Scott Counts, Mohit Bansal, Benjamin Van Durme, Harsh Jhamtani, Gaurav Verma
Abstract
Computer-use agents need to capture procedural knowledge of how people use software. User telemetry offers a scalable source of this knowledge. However, learning reusable skills from these logs requires addressing three challenges: (1) Goal Underspecification, since logs do not record the goal behind each action; (2) Non-Replayability, since past activity cannot be replayed to evaluate skill updates; and (3) Interleaved Trajectories, since logs may mix several tasks without marking their boundaries. To address these, we introduce TeleTune, a framework for learning a textual skill library from offline logs without recorded goals, cannot be replayed during optimization, and may interleave tasks. TeleTune uses action-prediction errors on logged trajectories to propose library edits and keep only those that improve held-out action-prediction accuracy, which we call skill-guided progress. The learned workflows also enable retrieval of demonstrations that cover the subgoals of a new task. At test time, the agent is provided with the learned library and the workflow-based retrieved demonstrations. Experiments on WorkArena and Online-Mind2Web show that TeleTune outperforms random retrieval, Agent Workflow Memory (AWM), and their combination. We find that the best baseline varies by setting, whereas TeleTune achieves average success rates of 77.1% and 80.6%, respectively, improving over the strongest baseline on each benchmark by 6.7% and 7.7%. Under the heaviest perturbation of the WorkArena training data,TeleTune keeps the highest average success rate at 68.5%, 6.3% above the strongest baseline. Our analyses show (1) skill optimization and workflow-based retrieval are complementary, (2) optimizing on fixed logs costs 5 to 75 times fewer tokens than validating the same edits with live episodes, (3) skill-guided progress tracks the live success rate.