AI Paper Atlas

Research map / Continual learning / Online and replay-based

$\phi$-DPO: Fairness Direct Preference Optimization Approach to Continual Learning in Large Multimodal Models

Thanh-Dat Truong, Huu-Thien Tran, Jackson Cothren, Bhiksha Raj, Khoa Luu

CVPR 2026 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use DPO preference loss instead of knowledge distillation (ablation)ϕ-DPO (LLaVA-7B) · MLLM-CL DomainMFN: 69.21 -> 71.74Improved
use Fair DPO (γ=2) instead of vanilla DPO (ablation)ϕ-DPO (LLaVA-7B) · MLLM-CL DomainMFN: 71.74 -> 74.00Improved
use Fair DPO instead of vanilla DPO (ablation)ϕ-DPO (LLaVA-7B) · MLLM-CL DomainBWT: -1.33 -> -0.37Improved
lower β from 0.10 to 0.01 (ablation)ϕ-DPO · MLLM-CL DomainMFT / BWT: MFT 74.29 -> 75.43; BWT -0.37 -> -2.66Mixed
raise β from 0.10 to 0.50 (ablation)ϕ-DPO · MLLM-CL DomainMFN: 74.00 -> 71.85Hurt
increase γ from 0 to 2.0 (ablation)ϕ-DPO · MLLM-CL DomainMFN: 71.74 -> 74.00Improved
raise γ from 2.0 to 5.0 (ablation)ϕ-DPO · MLLM-CL DomainMFN: 74.00 -> 72.18Hurt
use ϕ-DPO instead of DPO on LLaVA-13B (ablation)ϕ-DPO · MLLM-CL DomainMFN: 74.93 -> 75.81Improved
use ϕ-DPO instead of DPO on InternVL-7B (ablation)ϕ-DPO · MLLM-CL DomainMFN: 72.37 -> 74.23Improved
scale backbone from LLaVA-7B to LLaVA-13B (ablation)ϕ-DPO · MLLM-CL DomainBWT: MFN 74.00 -> 75.81; BWT -0.37 -> -0.59Mixed
ϕ-DPO vs best prior method MR-LoRAϕ-DPO · MLLM-CL DomainMFN: 69.63 -> 74.00Improved
ϕ-DPO vs HiDe on CoINϕ-DPO · CoINMFN: 62.82 -> 68.86Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy