Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.22240 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917519683485696 |
|---|---|
| author | Zhang, Hongbin Gao, Ning Dai, Yuqin Wu, Ruiyuan Wang, Jinpeng Gao, Rena Wei Tan, Bingdong Gao, Shuzheng Li, Zongjie Wang, Chaozheng |
| author_facet | Zhang, Hongbin Gao, Ning Dai, Yuqin Wu, Ruiyuan Wang, Jinpeng Gao, Rena Wei Tan, Bingdong Gao, Shuzheng Li, Zongjie Wang, Chaozheng |
| contents | Proactive task-oriented dialogue (TOD), such as outbound sales, demands a persuasive agent that actively probes the user's concerns and steers the conversation toward acceptance within a bounded number of turns. Yet post-trained LLMs are inherently conservative, and reward-shaping RL (e.g., GRPO) struggles since it only re-weights what an already passive policy samples. We show that conditioning on the user's latent concerns unlocks proactive capability that no amount of sampling can undermine, establishing these concerns as a pivotal training-time signal. To operationalize this finding, we build the \textbf{Cognitive User Simulator}, which models each user as a stratified persona comprising observable external traits and hidden internal concerns. The simulator produces faithful and diverse interactions, while emitting per-turn state dynamics that track persuasion progress. We then introduce \textbf{Simulator-Induced Asymmetric-View Policy Optimization}, which converts the modeled concerns and the simulation state transition into complementary training objectives: (1) \emph{Asymmetric On-Policy Self-Distillation} that transfers concern-aware behavior from a privileged view of the same policy into its deployable, conversation-only view; and (2) \emph{State-Transition Policy Refinement} ... |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_22240 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Unlocking Proactivity in Task-Oriented Dialogue Zhang, Hongbin Gao, Ning Dai, Yuqin Wu, Ruiyuan Wang, Jinpeng Gao, Rena Wei Tan, Bingdong Gao, Shuzheng Li, Zongjie Wang, Chaozheng Artificial Intelligence Proactive task-oriented dialogue (TOD), such as outbound sales, demands a persuasive agent that actively probes the user's concerns and steers the conversation toward acceptance within a bounded number of turns. Yet post-trained LLMs are inherently conservative, and reward-shaping RL (e.g., GRPO) struggles since it only re-weights what an already passive policy samples. We show that conditioning on the user's latent concerns unlocks proactive capability that no amount of sampling can undermine, establishing these concerns as a pivotal training-time signal. To operationalize this finding, we build the \textbf{Cognitive User Simulator}, which models each user as a stratified persona comprising observable external traits and hidden internal concerns. The simulator produces faithful and diverse interactions, while emitting per-turn state dynamics that track persuasion progress. We then introduce \textbf{Simulator-Induced Asymmetric-View Policy Optimization}, which converts the modeled concerns and the simulation state transition into complementary training objectives: (1) \emph{Asymmetric On-Policy Self-Distillation} that transfers concern-aware behavior from a privileged view of the same policy into its deployable, conversation-only view; and (2) \emph{State-Transition Policy Refinement} ... |
| title | Unlocking Proactivity in Task-Oriented Dialogue |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2605.22240 |