https://arxiv.org/abs/2404.03715

Skip to main content
Cornell University
We gratefully acknowledge support from the Simons Foundation, member
institutions, and all contributors. Donate
 
arxiv logo > cs > arXiv:2404.03715
[                    ]

Help | Advanced Search

[All fields        ]
Search
arXiv logo
Cornell University Logo
[                    ] GO
quick links

  * Login
  * Help Pages
  * About

Computer Science > Machine Learning

arXiv:2404.03715 (cs)
[Submitted on 4 Apr 2024]

Title:Direct Nash Optimization: Teaching Language Models to
Self-Improve with General Preferences

Authors:Corby Rosset, Ching-An Cheng, Arindam Mitra, Michael
Santacroce, Ahmed Awadallah, Tengyang Xie
View a PDF of the paper titled Direct Nash Optimization: Teaching
Language Models to Self-Improve with General Preferences, by Corby
Rosset and 5 other authors
View PDF HTML (experimental)

    Abstract:This paper studies post-training large language models
    (LLMs) using preference feedback from a powerful oracle to help a
    model iteratively improve over itself. The typical approach for
    post-training LLMs involves Reinforcement Learning from Human
    Feedback (RLHF), which traditionally separates reward learning
    and subsequent policy optimization. However, such a reward
    maximization approach is limited by the nature of "point-wise"
    rewards (such as Bradley-Terry model), which fails to express
    complex intransitive or cyclic preference relations. While
    advances on RLHF show reward learning and policy optimization can
    be merged into a single contrastive objective for stability, they
    yet still remain tethered to the reward maximization framework.
    Recently, a new wave of research sidesteps the reward
    maximization presumptions in favor of directly optimizing over
    "pair-wise" or general preferences. In this paper, we introduce
    Direct Nash Optimization (DNO), a provable and scalable algorithm
    that marries the simplicity and stability of contrastive learning
    with theoretical generality from optimizing general preferences.
    Because DNO is a batched on-policy algorithm using a
    regression-based objective, its implementation is straightforward
    and efficient. Moreover, DNO enjoys monotonic improvement across
    iterations that help it improve even over a strong teacher (such
    as GPT-4). In our experiments, a resulting 7B parameter Orca-2.5
    model aligned by DNO achieves the state-of-the-art win-rate
    against GPT-4-Turbo of 33% on AlpacaEval 2.0 (even after
    controlling for response length), an absolute gain of 26% (7% to
    33%) over the initializing model. It outperforms models with far
    more parameters, including Mistral Large, Self-Rewarding LM (70B
    parameters), and older versions of GPT-4.

Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI);
          Computation and Language (cs.CL)
Cite as:  arXiv:2404.03715 [cs.LG]
          (or arXiv:2404.03715v1 [cs.LG] for this version)

Submission history

From: Corbin Rosset [view email]
[v1] Thu, 4 Apr 2024 17:56:41 UTC (288 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled Direct Nash Optimization: Teaching
    Language Models to Self-Improve with General Preferences, by
    Corby Rosset and 5 other authors
  * View PDF
  * HTML (experimental)
  * TeX Source
  * Other Formats

license icon view license
Current browse context:
cs.LG
< prev   |   next >
new | recent | 2404
Change to browse by:
cs
cs.AI
cs.CL

References & Citations

  * NASA ADS
  * Google Scholar
  * Semantic Scholar

a export BibTeX citation Loading...

BibTeX formatted citation

x
[loading...          ]
Data provided by:

Bookmark

BibSonomy logo Reddit logo
(*) Bibliographic Tools

Bibliographic and Citation Tools

[ ] Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
[ ] Litmaps Toggle
Litmaps (What is Litmaps?)
[ ] scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
( ) Code, Data, Media

Code, Data and Media Associated with this Article

[ ] Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
[ ] DagsHub Toggle
DagsHub (What is DagsHub?)
[ ] GotitPub Toggle
Gotit.pub (What is GotitPub?)
[ ] Links to Code Toggle
Papers with Code (What is Papers with Code?)
[ ] ScienceCast Toggle
ScienceCast (What is ScienceCast?)
( ) Demos

Demos

[ ] Replicate Toggle
Replicate (What is Replicate?)
[ ] Spaces Toggle
Hugging Face Spaces (What is Spaces?)
[ ] Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
( ) Related Papers

Recommenders and Search Tools

[ ] Link to Influence Flower
Influence Flower (What are Influence Flowers?)
[ ] Connected Papers Toggle
Connected Papers (What is Connected Papers?)
[ ] Core recommender toggle
CORE Recommender (What is CORE?)
[ ] IArxiv recommender toggle
IArxiv Recommender (What is IArxiv?)

  * Author
  * Venue
  * Institution
  * Topic

( ) About arXivLabs

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.

Which authors of this paper are endorsers? | Disable MathJax (What is
MathJax?)

  * About
  * Help

  * Click here to contact arXiv Contact
  * Click here to subscribe Subscribe

  * Copyright
  * Privacy Policy

  * Web Accessibility Assistance
  * arXiv Operational Status
    Get status notifications via email or slack