PPubMed1 Apr 2021Cited 16×

A dynamic goal adapted task oriented dialogue agent A dynamic goal adapted task oriented dialogue agent

Abhisek Tiwari, Tulika Saha, Sriparna Saha, Shubhashis Sengupta, Anutosh Maitra, Roshni Ramnani, Pushpak Bhattacharyya, Weinan Zhang

Abstract

This paper presents a dynamic goal adapted task-oriented dialogue agent that can adapt to goal deviations and serve user goals dynamically. The dialogue policy learning task is formulated as a Partially Observable Markov Decision Process (POMDP) with a unique state representation and a novel reward model. We created and annotated a dialogue corpus, DevVA, that contains conversation pertaining to user goal deviation. The agent utilizes user’s sentiment in dialogue policy learning as immediate feedback for identifying goal deviation/discrepancy and making the VA user-adaptive. The negative senti

§ The Valyu brief

The paper, in five minutes

Read from the full text via Valyu Contents

The DGDVA handles dynamic user goals in dialogue, achieving high user satisfaction and a 0.88 task success rate.

What’s new

i.Created the DevVA dataset with annotated intents, slots, and sentiment labels
ii.Developed a Dynamic Goal Driven Dialogue Agent (DGDVA) using a Goal Driven Module on top of deep reinforcement learning
iii.Integrated user sentiment as feedback to detect goal discrepancies and adapt the dialogue

How they did it

The authors first built the DevVA conversational dataset containing dynamic goal deviations. They then trained a deep reinforcement‑learning dialogue manager augmented with a Goal Driven Module that uses negative user sentiment to identify goal discrepancies and update the user’s goal in real time.

What they found

Task success rate of 0.88

Caveats

✕Comparison with traditional agents may be unfair because they were not trained for dynamic‑goal scenarios
✕Ensuring optimal dialogue policy is difficult without a large supervised training corpus
✕Reinforcement‑learning agents require significant human effort to tune the reward model

§ The Valyu brief

Reading the full paper and taking notes. This takes a few seconds…

§ Ask this paper

Ask a question about this paper

Valyu reads the full text and answers from what the paper actually says.

§ Related work in other archives

A dynamic goal adapted task oriented dialogue agent A dynamic goal adapted task oriented dialogue agent

The paper, in five minutes

Ask a question about this paper

Similar papers in all four archives

Similar papers in all four archives

The paper, in five minutes