mDPO: Conditional Preference Optimization for Multimodal Large Language Models
- URL: http://arxiv.org/abs/2406.11839v1
- Date: Mon, 17 Jun 2024 17:59:58 GMT
- Title: mDPO: Conditional Preference Optimization for Multimodal Large Language Models
- Authors: Fei Wang, Wenxuan Zhou, James Y. Huang, Nan Xu, Sheng Zhang, Hoifung Poon, Muhao Chen,
- Abstract summary: Direct preference optimization (DPO) has shown to be an effective method for large language model (LLM) alignment.
Recent works have attempted to apply DPO to multimodal scenarios but have found it challenging to achieve consistent improvement.
We propose mDPO, a multimodal DPO objective that prevents the over-prioritization of language-only preferences by also optimizing image preference.
- Score: 52.607764280030196
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Direct preference optimization (DPO) has shown to be an effective method for large language model (LLM) alignment. Recent works have attempted to apply DPO to multimodal scenarios but have found it challenging to achieve consistent improvement. Through a comparative experiment, we identify the unconditional preference problem in multimodal preference optimization, where the model overlooks the image condition. To address this problem, we propose mDPO, a multimodal DPO objective that prevents the over-prioritization of language-only preferences by also optimizing image preference. Moreover, we introduce a reward anchor that forces the reward to be positive for chosen responses, thereby avoiding the decrease in their likelihood -- an intrinsic problem of relative preference optimization. Experiments on two multimodal LLMs of different sizes and three widely used benchmarks demonstrate that mDPO effectively addresses the unconditional preference problem in multimodal preference optimization and significantly improves model performance, particularly in reducing hallucination.
Related papers
- Hybrid Preference Optimization: Augmenting Direct Preference Optimization with Auxiliary Objectives [0.5120567378386615]
We propose a hybrid approach to aligning large language models (LLMs)
With a simple augmentation to the implicit reward decomposition of DPO, we allow for tuning LLMs to maximize a set of arbitrary auxiliary rewards.
The proposed method, Hybrid Preference Optimization (HPO), shows the ability to effectively generalize to both user preferences and auxiliary designer objectives.
arXiv Detail & Related papers (2024-05-28T08:35:48Z) - Provably Mitigating Overoptimization in RLHF: Your SFT Loss is Implicitly an Adversarial Regularizer [52.09480867526656]
We identify the source of misalignment as a form of distributional shift and uncertainty in learning human preferences.
To mitigate overoptimization, we first propose a theoretical algorithm that chooses the best policy for an adversarially chosen reward model.
Using the equivalence between reward models and the corresponding optimal policy, the algorithm features a simple objective that combines a preference optimization loss and a supervised learning loss.
arXiv Detail & Related papers (2024-05-26T05:38:50Z) - Multi-Reference Preference Optimization for Large Language Models [56.84730239046117]
We introduce a novel closed-form formulation for direct preference optimization using multiple reference models.
The resulting algorithm, Multi-Reference Preference Optimization (MRPO), leverages broader prior knowledge from diverse reference models.
Our experiments demonstrate that LLMs finetuned with MRPO generalize better in various preference data, regardless of data scarcity or abundance.
arXiv Detail & Related papers (2024-05-26T00:29:04Z) - Direct Preference Optimization with an Offset [58.7977683502207]
Direct preference optimization (DPO) is a successful strategy for aligning large language models with human preferences.
We propose a generalization of DPO, termed DPO with an offset (ODPO), that does not treat every preference pair equally during fine-tuning.
arXiv Detail & Related papers (2024-02-16T10:55:38Z) - Towards Efficient Exact Optimization of Language Model Alignment [93.39181634597877]
Direct preference optimization (DPO) was proposed to directly optimize the policy from preference data.
We show that DPO derived based on the optimal solution of problem leads to a compromised mean-seeking approximation of the optimal solution in practice.
We propose efficient exact optimization (EXO) of the alignment objective.
arXiv Detail & Related papers (2024-02-01T18:51:54Z) - Multi-Objective Bayesian Optimization with Active Preference Learning [18.066263838953223]
We propose a Bayesian optimization (BO) approach to identifying the most preferred solution in a multi-objective optimization (MOO) problem.
To minimize the interaction cost with the decision maker (DM), we also propose an active learning strategy for the preference estimation.
arXiv Detail & Related papers (2023-11-22T15:24:36Z) - Beyond One-Preference-Fits-All Alignment: Multi-Objective Direct
Preference Optimization [78.50294936259026]
We present Multi-Objective Direct Preference Optimization (MODPO) for multiple alignment objectives with minimal overheads.
MODPO folds language modeling directly into reward modeling, training LMs as implicit collective reward models (cRMs) that combine all objectives with specific weightings.
While theoretically guaranteed to produce the same optimal solutions as MORLHF, MODPO is practically more stable and computationally efficient.
arXiv Detail & Related papers (2023-10-05T17:35:26Z) - Statistical Rejection Sampling Improves Preference Optimization [42.57245965632205]
We introduce a novel approach to source preference data from the target optimal policy using rejection sampling.
We also propose a unified framework that enhances the loss functions used in both Sequence Likelihood (SLiC) and Direct Preference Optimization (DPO) from a preference modeling standpoint.
arXiv Detail & Related papers (2023-09-13T01:07:25Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.