Scalable Sentiment for Sequence-to-sequence Chatbot Response with Performance Analysis

TL;DR

This paper introduces five models to incorporate sentiment control into seq2seq chatbot responses and evaluates their effectiveness using new and existing metrics, highlighting reinforcement learning and cycleGAN as particularly promising.

Contribution

The paper proposes five novel models for sentiment adjustment in chatbot responses and develops evaluation metrics that correlate well with human judgment.

Findings

Reinforcement learning and cycleGAN outperform other models.

Evaluation metrics align closely with human assessments.

Models effectively modify sentiment without losing response relevance.

Abstract

Conventional seq2seq chatbot models only try to find the sentences with the highest probabilities conditioned on the input sequences, without considering the sentiment of the output sentences. Some research works trying to modify the sentiment of the output sequences were reported. In this paper, we propose five models to scale or adjust the sentiment of the chatbot response: persona-based model, reinforcement learning, plug and play model, sentiment transformation network and cycleGAN, all based on the conventional seq2seq model. We also develop two evaluation metrics to estimate if the responses are reasonable given the input. These metrics together with other two popularly used metrics were used to analyze the performance of the five proposed models on different aspects, and reinforcement learning and cycleGAN were shown to be very attractive. The evaluation metrics were also found to be well correlated with human evaluation.