# Pre-Trained and Attention-Based Neural Networks for Building Noetic Task-Oriented Dialogue Systems

Source: [Pre-Trained and Attention-Based Neural Networks for Building Noetic Task-Oriented Dialogue Systems](<https://www.sota2.com/research/paper/68fa67501e209131cea1556b>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-04T03:15:55.044Z

Authors: Jia-Chen Gu, Tianda Li, Quan Liu, Xiaodan Zhu, Zhen-Hua Ling, Yu-Ping Ruan

Published: 2020-04-04T14:14:43.000Z

## Abstract

The NOESIS II challenge, as the Track 2 of the 8th Dialogue System Technology Challenges (DSTC 8), is the extension of DSTC 7. This track incorporates new elements that are vital for the creation of a deployed task-oriented dialogue system. This paper describes our systems that are evaluated on all subtasks under this challenge. We study the problem of employing pre-trained attention-based network for multi-turn dialogue systems. Meanwhile, several adaptation methods are proposed to adapt the pre-trained language models for multi-turn dialogue systems, in order to keep the intrinsic property of dialogue systems. In the released evaluation results of Track 2 of DSTC 8, our proposed models ranked fourth in subtask 1, third in subtask 2, and first in subtask 3 and subtask 4 respectively.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Dialogue Thread Reconstruction (Subtask 4) on Ubuntu Track 2 DSTC 8 (test)](<https://www.sota2.com/research/sota/dialogue-thread-reconstruction-subtask-4-on-ubuntu-track-2-dstc-8-test.md>): Dialogue Thread Reconstruction (Subtask 4); Ubuntu Track 2 DSTC 8 (test); Precision: 44.3.

- [Response Selection (Subtask 1) on Ubuntu Track 2 DSTC 8 (test)](<https://www.sota2.com/research/sota/response-selection-subtask-1-on-ubuntu-track-2-dstc-8-test.md>): Response Selection (Subtask 1); Ubuntu Track 2 DSTC 8 (test); Recall@1: 64.9.

- [Response Selection (Subtask 1) on Advising Track 2 DSTC 8 (test)](<https://www.sota2.com/research/sota/response-selection-subtask-1-on-advising-track-2-dstc-8-test.md>): Response Selection (Subtask 1); Advising Track 2 DSTC 8 (test); Recall@1: 22.4.

- [Response Selection (Subtask 3) on Advising Track 2 hidden DSTC 8 (test)](<https://www.sota2.com/research/sota/response-selection-subtask-3-on-advising-track-2-hidden-dstc-8-test.md>): Response Selection with Multiple Correct Responses (Subtask 3); Advising Track 2 hidden DSTC 8 (test); Accuracy: 0.802.

- [Response Selection with No Correct Response (Subtask 2) on Ubuntu Track 2 DSTC 8 (test)](<https://www.sota2.com/research/sota/response-selection-with-no-correct-response-subtask-2-on-ubuntu-track-2-dstc-8-test.md>): Response Selection with No Correct Response (Subtask 2); Ubuntu Track 2 DSTC 8 (test); Recall@1: 50.6.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2004.01940>)

[PDF](<https://arxiv.org/pdf/2004.01940.pdf>)