CESAR: Automatic Induction of Compositional Instructions for Multi-turn Dialogs

Taha Aksu, Devamanyu Hazarika, Shikib Mehri, Seokhwan Kim, Dilek Hakkani-Tur, Yang Liu, Mahdi Namazifar


Abstract
Instruction-based multitasking has played a critical role in the success of large language models (LLMs) in multi-turn dialog applications. While publicly available LLMs have shown promising performance, when exposed to complex instructions with multiple constraints, they lag against state-of-the-art models like ChatGPT. In this work, we hypothesize that the availability of large-scale complex demonstrations is crucial in bridging this gap. Focusing on dialog applications, we propose a novel framework, CESAR, that unifies a large number of dialog tasks in the same format and allows programmatic induction of complex instructions without any manual effort. We apply CESAR on InstructDial, a benchmark for instruction-based dialog tasks. We further enhance InstructDial with new datasets and tasks and utilize CESAR to induce complex tasks with compositional instructions. This results in a new benchmark called InstructDial++, which includes 63 datasets with 86 basic tasks and 68 composite tasks. Through rigorous experiments, we demonstrate the scalability of CESAR in providing rich instructions. Models trained on InstructDial++ can follow compositional prompts, such as prompts that ask for multiple stylistic constraints.
Anthology ID:
2023.emnlp-main.717
Volume:
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
11709–11737
Language:
URL:
https://aclanthology.org/2023.emnlp-main.717
DOI:
10.18653/v1/2023.emnlp-main.717
Bibkey:
Cite (ACL):
Taha Aksu, Devamanyu Hazarika, Shikib Mehri, Seokhwan Kim, Dilek Hakkani-Tur, Yang Liu, and Mahdi Namazifar. 2023. CESAR: Automatic Induction of Compositional Instructions for Multi-turn Dialogs. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 11709–11737, Singapore. Association for Computational Linguistics.
Cite (Informal):
CESAR: Automatic Induction of Compositional Instructions for Multi-turn Dialogs (Aksu et al., EMNLP 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingest-acl-2023-videos/2023.emnlp-main.717.pdf