Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userTaming the Beast: Learning to Control Neural Conversational Models
88
0
0.0
(
0
)
Added by
Andrea Madotto Mr
Publication date
2021
fields
Informatics Engineering
and research's language is
English
Authors
Andrea Madotto
Ask ChatGPT about the research
No Arabic abstract
This thesis investigates the controllability of deep learning-based, end-to-end, generative dialogue systems in both task-oriented and chit-chat scenarios. In particular, we study the different aspects of controlling generative dialogue systems, including controlling styles and topics and continuously adding and combining dialogue skills. In the three decades since the first dialogue system was commercialized, the basic architecture of such systems has remained substantially unchanged, consisting of four pipelined basic components, namely, natural language understanding (NLU), dialogue state tracking (DST), a dialogue manager (DM) and natural language generation (NLG). The dialogue manager, which is the critical component of the modularized system, controls the response content and style. This module is usually programmed by rules and is designed to be highly controllable and easily extendable. With the emergence of powerful deep learning architectures, end-to-end generative dialogue systems have been proposed to optimize overall system performance and simplify training. However, these systems cannot be easily controlled and extended as the modularized dialogue manager can. This is because a single neural system is used, which is usually a large pre-trained language model (e.g., GPT-2), and thus it is hard to surgically change desirable attributes (e.g., style, topics, etc.). More importantly, uncontrollable dialogue systems can generate offensive and even toxic responses. Therefore, in this thesis, we study controllable methods for end-to-end generative dialogue systems in task-oriented and chit-chat scenarios. Throughout the chapters, we describe 1) how to control the style and topics of chit-chat models, 2) how to continuously control and extend task-oriented dialogue systems, and 3) how to compose and control multi-skill dialogue models.
rate research
Read More
This paper presents an empirical study of conversational question reformulation (CQR) with sequence-to-sequence architectures and pretrained language models (PLMs). We leverage PLMs to address the strong token-to-token independence assumption made in the common objective, maximum likelihood estimation, for the CQR task. In CQR benchmarks of task-oriented dialogue systems, we evaluate fine-tuned PLMs on the recently-introduced CANARD dataset as an in-domain task and validate the models using data from the TREC 2019 CAsT Track as an out-domain task. Examining a variety of architectures with different numbers of parameters, we demonstrate that the recent text-to-text transfer transformer (T5) achieves the best results both on CANARD and CAsT with fewer parameters, compared to similar transformer architectures.
In spoken conversational question answering (SCQA), the answer to the corresponding question is generated by retrieving and then analyzing a fixed spoken document, including multi-part conversations. Most SCQA systems have considered only retrieving information from ordered utterances. However, the sequential order of dialogue is important to build a robust spoken conversational question answering system, and the changes of utterances order may severely result in low-quality and incoherent corpora. To this end, we introduce a self-supervised learning approach, including incoherence discrimination, insertion detection, and question prediction, to explicitly capture the coreference resolution and dialogue coherence among spoken documents. Specifically, we design a joint learning framework where the auxiliary self-supervised tasks can enable the pre-trained SCQA systems towards more coherent and meaningful spoken dialogue learning. We also utilize the proposed self-supervised learning tasks to capture intra-sentence coherence. Experimental results demonstrate that our proposed method provides more coherent, meaningful, and appropriate responses, yielding superior performance gains compared to the original pre-trained language models. Our method achieves state-of-the-art results on the Spoken-CoQA dataset.
There has been considerable progress made towards conversational models that generate coherent and fluent responses; however, this often involves training large language models on large dialogue datasets, such as Reddit. These large conversational models provide little control over the generated responses, and this control is further limited in the absence of annotated conversational datasets for attribute specific generation that can be used for fine-tuning the model. In this paper, we first propose and evaluate plug-and-play methods for controllable response generation, which does not require dialogue specific datasets and does not rely on fine-tuning a large model. While effective, the decoding procedure induces considerable computational overhead, rendering the conversational model unsuitable for interactive usage. To overcome this, we introduce an approach that does not require further computation at decoding time, while also does not require any fine-tuning of a large language model. We demonstrate, through extensive automatic and human evaluation, a high degree of control over the generated conversational responses with regard to multiple desired attributes, while being fluent.
Recently, neural network based dialogue systems have become ubiquitous in our increasingly digitalized society. However, due to their inherent opaqueness, some recently raised concerns about using neural models are starting to be taken seriously. In fact, intentional or unintentional behaviors could lead to a dialogue system to generate inappropriate responses. Thus, in this paper, we investigate whether we can learn to craft input sentences that result in a black-box neural dialogue model being manipulated into having its outputs contain target words or match target sentences. We propose a reinforcement learning based model that can generate such desired inputs automatically. Extensive experiments on a popular well-trained state-of-the-art neural dialogue model show that our method can successfully seek out desired inputs that lead to the target outputs in a considerable portion of cases. Consequently, our work reveals the potential of neural dialogue models to be manipulated, which inspires and opens the door towards developing strategies to defend them.
Native speakers can judge whether a sentence is an acceptable instance of their language. Acceptability provides a means of evaluating whether computational language models are processing language in a human-like manner. We test the ability of computational language models, simple language features, and word embeddings to predict native English speakers judgments of acceptability on English-language essays written by non-native speakers. We find that much of the sentence acceptability variance can be captured by a combination of features including misspellings, word order, and word similarity (Pearsons r = 0.494). While predictive neural models fit acceptability judgments well (r = 0.527), we find that a 4-gram model with statistical smoothing is just as good (r = 0.528). Thanks to incorporating a count of misspellings, our 4-gram model surpasses both the previous unsupervised state-of-the art (Lau et al., 2015; r = 0.472), and the average non-expert native speaker (r = 0.46). Our results demonstrate that acceptability is well captured by n-gram statistics and simple language features.
suggested questions
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
