Language Model Decoding as Likelihood-Utility Alignment

10/13/2022
by   Martin Josifoski, et al.
11

A critical component of a successful language generation pipeline is the decoding algorithm. However, the general principles that should guide the choice of decoding algorithm remain unclear. Previous works only compare decoding algorithms in narrow scenarios and their findings do not generalize across tasks. To better structure the discussion, we introduce a taxonomy that groups decoding strategies based on their implicit assumptions about how well the model's likelihood is aligned with the task-specific notion of utility. We argue that this taxonomy allows a broader view of the decoding problem and can lead to generalizable statements because it is grounded on the interplay between the decoding algorithms and the likelihood-utility misalignment. Specifically, by analyzing the correlation between the likelihood and the utility of predictions across a diverse set of tasks, we provide the first empirical evidence supporting the proposed taxonomy, and a set of principles to structure reasoning when choosing a decoding algorithm. Crucially, our analysis is the first one to relate likelihood-based decoding strategies with strategies that rely on external information such as value-guided methods and prompting, and covers the most diverse set of tasks up-to-date.

READ FULL TEXT
research
03/29/2022

On Decoding Strategies for Neural Text Generators

When generating text from probabilistic models, the chosen decoding stra...
research
10/07/2022

An Analysis of the Effects of Decoding Algorithms on Fairness in Open-Ended Language Generation

Several prior works have shown that language models (LMs) can generate t...
research
04/22/2020

Trading Off Diversity and Quality in Natural Language Generation

For open-ended language generation tasks such as storytelling and dialog...
research
10/24/2022

Are Current Decoding Strategies Capable of Facing the Challenges of Visual Dialogue?

Decoding strategies play a crucial role in natural language generation s...
research
11/25/2016

A Simple, Fast Diverse Decoding Algorithm for Neural Generation

In this paper, we propose a simple, fast decoding algorithm that fosters...
research
02/06/2020

Consistency of a Recurrent Language Model With Respect to Incomplete Decoding

Despite strong performance on a variety of tasks, neural sequence models...

Please sign up or login with your details

Forgot password? Click here to reset