Memory-Augmented Generative Adversarial Transformers
This addresses the issue of low factual accuracy in Transformer-based conversational AI for applications like goal-oriented dialogues and style adaptation, though it is incremental as it builds on existing architectures.
The paper tackles the problem of conversational AI systems struggling to integrate external data with generated language by proposing a memory-augmented Generative Adversarial Transformer architecture, which improves factual question answering and enables style adaptation in dialogues.
Conversational AI systems that rely on Large Language Models, like Transformers, have difficulty interweaving external data (like facts) with the language they generate. Vanilla Transformer architectures are not designed for answering factual questions with high accuracy. This paper investigates a possible route for addressing this problem. We propose to extend the standard Transformer architecture with an additional memory bank holding extra information (such as facts drawn from a knowledge base), and an extra attention layer for addressing this memory. We add this augmented memory to a Generative Adversarial Network-inspired Transformer architecture. This setup allows for implementing arbitrary felicity conditions on the generated language of the Transformer. We first demonstrate how this machinery can be deployed for handling factual questions in goal-oriented dialogues. Secondly, we demonstrate that our approach can be useful for applications like {\it style adaptation} as well: the adaptation of utterances according to certain stylistic (external) constraints, like social properties of human interlocutors in dialogues.