Computer Science > Machine Learning
[Submitted on 5 Dec 2024 (v1), last revised 21 Feb 2025 (this version, v3)]
Title:SWEPO: Simultaneous Weighted Preference Optimization for Group Contrastive Alignment
View PDF HTML (experimental)Abstract:Direct Preference Optimization (DPO) has proven effective in aligning large language models with human preferences but is often constrained to pairwise comparisons -- overlooking additional positive and negative responses that are commonly available in real-world settings. We propose Simultaneous Weighted Preference Optimization (SWEPO), which incorporates multiple responses per query and prioritizes those that deviate most from the average reward. This deviation-based weighting focuses training on the most informative outliers, akin to a built-in curriculum. Theoretically, we prove that such multi-preference sampling lowers alignment bias, bounding the expected deviation from the true acceptable-response distribution at a rate of $\mathcal{O}(\tfrac{1}{\sqrt{k}})$. Empirically, SWEPO outperforms state-of-the-art baselines on the Ultra-Feedback dataset and demonstrates substantial improvements over DPO and InfoNCA, yielding boosts of up to $\sim 4$% on length-controlled win-rate on AlpacaEval.
Submission history
From: Taneesh Gupta [view email][v1] Thu, 5 Dec 2024 21:50:22 UTC (295 KB)
[v2] Wed, 8 Jan 2025 15:00:39 UTC (298 KB)
[v3] Fri, 21 Feb 2025 18:12:34 UTC (337 KB)
Current browse context:
cs.LG
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.