DeepRewind: Predicting and Repairing Premature Commitments in Deep Research Agents
Amirhossein Abaskohi, Peter West, Giuseppe Carenini and colleagues at the University of British Columbia introduce DeepRewind, a control layer for deep-research agents that blocks risky early commitments and rolls them back when later evidence contradicts them.
Ask this paper
Typed epistemic graph. The agent's state is a graph of sources, evidence, claims, hypotheses, assumptions, commitments, plans and drafts.
Predict before committing. A prompt-based world model estimates each intermediate conclusion's impact and reversibility from hypothesis narrowing, information loss, recovery cost and contradiction-trigger coverage; a binary controller blocks risky commitments.
Dependency-aware rollback. A consistency monitor undoes a commitment and everything built on it when new evidence invalidates it.
Results. On DRBench and LiveDRBench, against Open Deep Research, insight recall rises 3.6 points and premature commitments fall by 59.1%.
Abstract
Deep-research agents conduct long-horizon investigations through iterative search, evidence evaluation, belief revision, and synthesis. However, they may commit to claims before sufficient evidence is available, causing later reasoning to reinforce an incorrect interpretation. We introduce DeepRewind, an additive control layer for reversible deep research that represents the agent's evolving epistemic state as a typed graph of sources, evidence, claims, hypotheses, assumptions, commitments, plans, and drafts. Before accepting an intermediate conclusion, a prompt-based world model predicts its impact and estimates reversibility based on hypothesis narrowing, information loss, recovery cost, and contradiction-trigger coverage. A binary controller blocks risky commitments, while a consistency monitor performs dependency-aware rollback when later evidence invalidates them. Across DRBench and LiveDRBench, DeepRewind improves insight recall by 3.6 percentage points and reduces premature commitments by 59.1% relative to Open Deep Research.