Tech Meridian ← LIVE FEED
PROMY MERIDIAN RU

RESEARCH · RESEARCH · #1858

CLEAT: Co-training a clarification-value controller and task policy for interactive agents

The paper (arXiv:2610.12655v1) introduces CLEAT, a framework that co-trains a clarification-value controller and a task policy using paired rollouts to estimate a residual query value — the task utility gained by asking versus acting. Evaluated with Qwen3-4B across four benchmarks, CLEAT achieves the highest mean score in ten evaluation settings, with gains up to 13.2 and 6.5 percentage points on UserGym Travel and IN3 underspecification judgment accuracy respectively.

KEY POINTS

  1. The paper (arXiv:2610.12655v1) introduces CLEAT, a framework that co-trains a clarification-value controller and a task policy using paired rollouts to estimate a residual query value — the task utility gained by asking versus acting.
  2. Evaluated with Qwen3-4B across four benchmarks, CLEAT achieves the highest mean score in ten evaluation settings, with gains up to 13.2 and 6.5 percentage points on UserGym Travel and IN3 underspecification judgment accuracy respectively.
  3. By learning which clarifying questions actually improve task outcomes and adapting guidance as the agent's policy changes, CLEAT helps agents align actions with users' unstated goals and constraints.

WHY IT MATTERS

By learning which clarifying questions actually improve task outcomes and adapting guidance as the agent's policy changes, CLEAT helps agents align actions with users' unstated goals and constraints.

SOURCES & TIMELINE

1