Deriving the DPO Objective Under the Bradley-Terry Model

Learn how to derive the DPO objective under the bradley-terry model.


This content originally appeared on HackerNoon and was authored by Writings, Papers and Blogs on Text Models

:::info Authors:

(1) Rafael Rafailo, Stanford University and Equal contribution; more junior authors listed earlier;

(2) Archit Sharma, Stanford University and Equal contribution; more junior authors listed earlier;

(3) Eric Mitchel, Stanford University and Equal contribution; more junior authors listed earlier;

(4) Stefano Ermon, CZ Biohub;

(5) Christopher D. Manning, Stanford University;

(6) Chelsea Finn, Stanford University.

:::

Abstract and 1. Introduction

2 Related Work

3 Preliminaries

4 Direct Preference Optimization

5 Theoretical Analysis of DPO

6 Experiments

7 Discussion, Acknowledgements, and References

Author Contributions

\ A Mathematical Derivations

A.1 Deriving the Optimum of the KL-Constrained Reward Maximization Objective

A.2 Deriving the DPO Objective Under the Bradley-Terry Model

A.3 Deriving the DPO Objective Under the Plackett-Luce Model

A.4 Deriving the Gradient of the DPO Objective and A.5 Proof of Lemma 1 and 2

A.6 Proof of Theorem 1

\ B DPO Implementation Details and Hyperparameters

\ C Further Details on the Experimental Set-Up and C.1 IMDb Sentiment Experiment and Baseline Details

C.2 GPT-4 prompts for computing summarization and dialogue win rates

C.3 Unlikelihood baseline

\ D Additional Empirical Results

D.1 Performance of Best of N baseline for Various N and D.2 Sample Responses and GPT-4 Judgments

D.3 Human study details

A.2 Deriving the DPO Objective Under the Bradley-Terry Model

It is straightforward to derive the DPO objective under the Bradley-Terry preference model as we have

\

\ In Section 4 we showed that we can express the (unavailable) ground-truth reward through its corresponding optimal policy:

\

\ Substituting Eq. 17 into Eq. 16 we obtain:

\

\

:::info This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license.

:::

\


This content originally appeared on HackerNoon and was authored by Writings, Papers and Blogs on Text Models


Print Share Comment Cite Upload Translate Updates
APA

Writings, Papers and Blogs on Text Models | Sciencx (2024-08-25T21:14:08+00:00) Deriving the DPO Objective Under the Bradley-Terry Model. Retrieved from https://www.scien.cx/2024/08/25/deriving-the-dpo-objective-under-the-bradley-terry-model/

MLA
" » Deriving the DPO Objective Under the Bradley-Terry Model." Writings, Papers and Blogs on Text Models | Sciencx - Sunday August 25, 2024, https://www.scien.cx/2024/08/25/deriving-the-dpo-objective-under-the-bradley-terry-model/
HARVARD
Writings, Papers and Blogs on Text Models | Sciencx Sunday August 25, 2024 » Deriving the DPO Objective Under the Bradley-Terry Model., viewed ,<https://www.scien.cx/2024/08/25/deriving-the-dpo-objective-under-the-bradley-terry-model/>
VANCOUVER
Writings, Papers and Blogs on Text Models | Sciencx - » Deriving the DPO Objective Under the Bradley-Terry Model. [Internet]. [Accessed ]. Available from: https://www.scien.cx/2024/08/25/deriving-the-dpo-objective-under-the-bradley-terry-model/
CHICAGO
" » Deriving the DPO Objective Under the Bradley-Terry Model." Writings, Papers and Blogs on Text Models | Sciencx - Accessed . https://www.scien.cx/2024/08/25/deriving-the-dpo-objective-under-the-bradley-terry-model/
IEEE
" » Deriving the DPO Objective Under the Bradley-Terry Model." Writings, Papers and Blogs on Text Models | Sciencx [Online]. Available: https://www.scien.cx/2024/08/25/deriving-the-dpo-objective-under-the-bradley-terry-model/. [Accessed: ]
rf:citation
» Deriving the DPO Objective Under the Bradley-Terry Model | Writings, Papers and Blogs on Text Models | Sciencx | https://www.scien.cx/2024/08/25/deriving-the-dpo-objective-under-the-bradley-terry-model/ |

Please log in to upload a file.




There are no updates yet.
Click the Upload button above to add an update.

You must be logged in to translate posts. Please log in or register.