Single document news summarization has seen substantial progress on faithfulness in recent years, driven by research on the evaluation of factual consistency, or hallucinations. We ask whether these advances carry over to other text summarization domains. We propose a new evaluation benchmark on topic-focused dialogue summarization, generated by LLMs of varying sizes. We provide binary sentence- level human annotations of the factual consistency of these summaries along with detailed explanations of factually inconsistent sentences. Our analysis shows that existing LLMs hallucinate significant amounts of factual errors in the dialogue domain, regardless of the model’s size. On the other hand, when LLMs, including GPT-4, serve as binary factual evaluators, they perform poorly and can be outperformed by prevailing state-of-the-art specialized factuality evaluation metrics. Finally, we conducted an analysis of hallucination types with a curated error taxonomy. We find that there are diverse errors and error distributions in model-generated summaries and that non-LLM based metrics can capture all error types better than LLM-based evaluators.

The paper presents the results of the Shared Task hosted by the Second DialDoc Workshop on Document-grounded Dialogue and Conversational Question Answering co-located at ACL 2022. The primary goal of this Shared Task is to build goal-oriented information-seeking conversation systems that are grounded in the domain documents, where each dialogue could correspond to multiple subtasks that are based on different documents. The task is to generate agent responses in natural language given the dialogue and document contexts. There are two task settings and leaderboards based on (1) the same sets of domains (SEEN) and (2) one unseen domain (UNSEEN). There are over 20 teams participating in Dev Phase and 8 teams participating in both Dev and Test Phases. Multiple submissions significantly outperform the baseline. The best-performing system achieves 52.06 F1 and the total of 191.30 on the SEEN task; and 34.65 F1 and the total of 130.79 on the UNSEEN task.

Collecting data for training dialog systems can be extremely expensive due to the involvement of human participants and the need for extensive annotation. Especially in document-grounded dialog systems, human experts need to carefully read the unstructured documents to answer the users’ questions. As a result, existing document-grounded dialog datasets are relatively small-scale and obstruct the effective training of dialogue systems. In this paper, we propose an automatic data augmentation technique grounded on documents through a generative dialogue model. The dialogue model consists of a user bot and agent bot that can synthesize diverse dialogues given an input document, which is then used to train a downstream model. When supplementing the original dataset, our method achieves significant improvement over traditional data augmentation methods. We also achieve great performance in the low-resource setting.

Explaining neural network models is important for increasing their trustworthiness in real-world applications. Most existing methods generate post-hoc explanations for neural network models by identifying individual feature attributions or detecting interactions between adjacent features. However, for models with text pairs as inputs (e.g., paraphrase identification), existing methods are not sufficient to capture feature interactions between two texts and their simple extension of computing all word-pair interactions between two texts is computationally inefficient. In this work, we propose the Group Mask (GMASK) method to implicitly detect word correlations by grouping correlated words from the input text pair together and measure their contribution to the corresponding NLP tasks as a whole. The proposed method is evaluated with two different model architectures (decomposable attention model and BERT) across four datasets, including natural language inference and paraphrase identification tasks. Experiments show the effectiveness of GMASK in providing faithful explanations to these models.

Machine reading comprehension is a challenging task especially for querying documents with deep and interconnected contexts. Transformer-based methods have shown advanced performances on this task; however, most of them still treat documents as a flat sequence of tokens. This work proposes a new Transformer-based method that reads a document as tree slices. It contains two modules for identifying more relevant text passage and the best answer span respectively, which are not only jointly trained but also jointly consulted at inference time. Our evaluation results show that our proposed method outperforms several competitive baseline approaches on two datasets from varied domains.

We present the results of Shared Task at Workshop DialDoc 2021 that is focused on document-grounded dialogue and conversational question answering. The primary goal of this Shared Task is to build goal-oriented information-seeking conversation systems that can identify the most relevant knowledge in the associated document for generating agent responses in natural language. It includes two subtasks on predicting agent responses: the first subtask is to predict the grounding text span in the given document for next agent response; the second subtask is to generate agent response in natural language given the context. Many submissions outperform baseline significantly. For the first task, the best-performing system achieved 67.1 Exact Match and 76.3 F1. For the second subtask, the best system achieved 41.1 SacreBLEU and highest rank by human evaluation.

We propose MultiDoc2Dial, a new task and dataset on modeling goal-oriented dialogues grounded in multiple documents. Most previous works treat document-grounded dialogue modeling as machine reading comprehension task based on a single given document or passage. In this work, we aim to address more realistic scenarios where a goal-oriented information-seeking conversation involves multiple topics, and hence is grounded on different documents. To facilitate such task, we introduce a new dataset that contains dialogues grounded in multiple documents from four different domains. We also explore modeling the dialogue-based and document-based contexts in the dataset. We present strong baseline approaches and various experimental results, aiming to support further research efforts on such a task.

Implicit relation classification on Penn Discourse TreeBank (PDTB) 2.0 is a common benchmark task for evaluating the understanding of discourse relations. However, the lack of consistency in preprocessing and evaluation poses challenges to fair comparison of results in the literature. In this work, we highlight these inconsistencies and propose an improved evaluation protocol. Paired with this protocol, we report strong baseline results from pretrained sentence encoders, which set the new state-of-the-art for PDTB 2.0. Furthermore, this work is the first to explore fine-grained relation classification on PDTB 3.0. We expect our work to serve as a point of comparison for future work, and also as an initiative to discuss models of larger context and possible data augmentations for downstream transferability.

We introduce doc2dial, a new dataset of goal-oriented dialogues that are grounded in the associated documents. Inspired by how the authors compose documents for guiding end users, we first construct dialogue flows based on the content elements that corresponds to higher-level relations across text sections as well as lower-level relations between discourse units within a section. Then we present these dialogue flows to crowd contributors to create conversational utterances. The dataset includes over 4500 annotated conversations with an average of 14 turns that are grounded in over 450 documents from four domains. Compared to the prior document-grounded dialogue datasets, this dataset covers a variety of dialogue scenes in information-seeking conversations. For evaluating the versatility of the dataset, we introduce multiple dialogue modeling tasks and present baseline approaches.

In this paper we introduce Chat-crowd, an interactive environment for visual layout composition via conversational interactions. Chat-crowd supports multiple agents with two conversational roles: agents who play the role of a designer are in charge of placing objects in an editable canvas according to instructions or commands issued by agents with a director role. The system can be integrated with crowdsourcing platforms for both synchronous and asynchronous data collection and is equipped with comprehensive quality controls on the performance of both types of agents. We expect that this system will be useful to build multimodal goal-oriented dialog tasks that require spatial and geometric reasoning.