The paper proposes Counterfactual Preference Optimization, a DPO-style method that uses concept-graph-generated counterfactual reasoning trajectories to improve multimodal LLM reinforcement fine-tuning on chest X-ray tasks.
Causal diagrams for empirical research
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Walking the Tightrope: Disentangling Beneficial and Detrimental Drifts in Non-Stationary Custom-Tuning
The paper proposes Counterfactual Preference Optimization, a DPO-style method that uses concept-graph-generated counterfactual reasoning trajectories to improve multimodal LLM reinforcement fine-tuning on chest X-ray tasks.