Dialogue summarization involves summarizing long conversations while preserving the most salient information. Real-life dialogues often involve naturally occurring variations (e.g., repetitions, hesitations). In this study, we systematically investigate the impact of such variations on state-of-the-art open dialogue summarization models whose details are publicly known (e.g., architectures, weights, and training corpora). To simulate real-life variations, we introduce two types of perturbations: utterance-level perturbations that modify individual utterances with errors and language variations, and dialogue-level perturbations that add non-informative exchanges (e.g., repetitions, greetings). We perform our analysis along three dimensions of robustness: consistency, saliency, and faithfulness, which aim to capture different aspects of performance of a summarization model. We find that both fine-tuned and instruction-tuned models are affected by input variations, with the latter being more susceptible, particularly to dialogue-level perturbations. We also validate our findings via human evaluation. Finally, we investigate whether the robustness of fine-tuned models can be improved by training them with a fraction of perturbed data. We find that this approach does not yield consistent performance gains, warranting further research. Overall, our work highlights robustness challenges in current open encoder-decoder summarization models and provides insights for future research.

An emergent research trend explores the use of Large Language Models (LLMs) as the backbone of agentic systems (e.g., SWE-Bench, Agent-Bench). To fulfill LLMs’ potential as autonomous agents, they must be able to identify, call, and interact with a variety of external tools and application program interfaces (APIs). This capability of LLMs, commonly termed function calling, leads to a myriad of advantages such as access to current and domain-specific information in databases and the outsourcing of tasks that can be reliably performed by tools. In this work, we introduce Granite-20B-FunctionCalling, a model trained using a multi-task training approach on seven fundamental tasks encompassed in function calling. Our comprehensive evaluation on multiple out-of-domain datasets, which compares Granite-20B-FunctionCalling to more than 15 other best proprietary and open models, shows that Granite-20B-FunctionCalling has better generalizability on multiple tasks across seven different evaluation benchmarks. Moreover, Granite-20B-FunctionCalling shows the best performance among all open models and ranks among the top on the Berkeley Function Calling Leaderboard (BFCL).

Applying language models to natural language processing tasks typically relies on the representations in the final model layer, as intermediate hidden layer representations are presumed to be less informative. In this work, we argue that due to the gradual improvement across model layers, additional information can be gleaned from the contrast between higher and lower layers during inference. Specifically, in choosing between the probable next token predictions of a generative model, the predictions of lower layers can be used to highlight which candidates are best avoided. We propose a novel approach that utilizes the contrast between layers to improve text generation outputs, and show that it mitigates degenerative behaviors of the model in open-ended generation, significantly improving the quality of generated texts. Furthermore, our results indicate that contrasting between model layers at inference time can yield substantial benefits to certain aspects of general language model capabilities, more effectively extracting knowledge during inference from a given set of model parameters.

With the growing interest in large language models, the need for evaluating the quality of machine text compared to reference (typically human-generated) text has become focal attention. Most recent works focus either on task-specific evaluation metrics or study the properties of machine-generated text captured by the existing metrics. In this work, we propose a new evaluation scheme to model human judgments in 7 NLP tasks, based on the fine-grained mismatches between a pair of texts. Inspired by the recent efforts in several NLP tasks for fine-grained evaluation, we introduce a set of 13 mismatch error types such as spatial/geographic errors, entity errors, etc, to guide the model for better prediction of human judgments. We propose a neural framework for evaluating machine texts that uses these mismatch error types as auxiliary tasks and re-purposes the existing single-number evaluation metrics as additional scalar features, in addition to textual features extracted from the machine and reference texts. Our experiments reveal key insights about the existing metrics via the mismatch errors. We show that the mismatch errors between the sentence pairs on the held-out datasets from 7 NLP tasks align well with the human evaluation.

In this paper we explore the task of modeling semi-structured object sequences; in particular, we focus our attention on the problem of developing a structure-aware input representation for such sequences. Examples of such data include user activity on websites, machine logs, and many others. This type of data is often represented as a sequence of sets of key-value pairs over time and can present modeling challenges due to an ever-increasing sequence length. We propose a two-part approach, which first considers each key independently and encodes a representation of its values over time; we then self-attend over these value-aware key representations to accomplish a downstream task. This allows us to operate on longer object sequences than existing methods. We introduce a novel shared-attention-head architecture between the two modules and present an innovative training schedule that interleaves the training of both modules with shared weights for some attention heads. Our experiments on multiple prediction tasks using real-world data demonstrate that our approach outperforms a unified network with hierarchical encoding, as well as other methods including a record-centric representation and a flattened representation of the sequence.

Abstractive summarization models often produce factually inconsistent summaries that are not supported by the original article. Recently, a number of fact-consistent evaluation techniques have been proposed to address this issue; however, a detailed analysis of how these metrics agree with one another has yet to be conducted. In this paper, we present X-FACTOR, a cross-evaluation of three high-performing fact-aware abstractive summarization methods. First, we show that summarization models are often fine-tuned on datasets that contain factually inconsistent summaries and propose a fact-aware filtering mechanism that improves the quality of training data and, consequently, the factuality of these models. Second, we propose a corrector module that can be used to improve the factual consistency of generated summaries. Third, we present a re-ranking technique that samples summary instances from the output distribution of a summarization model and re-ranks the sampled instances based on their factuality. Finally, we provide a detailed cross-metric agreement analysis that shows how tuning a model to output summaries based on a particular factuality metric influences factuality as determined by the other metrics. Our goal in this work is to facilitate research that improves the factuality and faithfulness of abstractive summarization models.

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.

In a typical customer service chat scenario, customers contact a support center to ask for help or raise complaints, and human agents try to solve the issues. In most cases, at the end of the conversation, agents are asked to write a short summary emphasizing the problem and the proposed solution, usually for the benefit of other agents that may have to deal with the same customer or issue. The goal of the present article is advancing the automation of this task. We introduce the first large scale, high quality, customer care dialog summarization dataset with close to 6500 human annotated summaries. The data is based on real-world customer support dialogs and includes both extractive and abstractive summaries. We also introduce a new unsupervised, extractive summarization method specific to dialogs.

Neural abstractive summarization models have drastically improved in the recent years. However, the summaries generated by these models generally suffer from issues such as: not capturing the critical facts in source documents, and containing facts that are inconsistent with the source documents. In this work, we present a general framework to train abstractive summarization models to alleviate such issues. We first train a sequence-to-sequence model to summarize documents, and then further train this model in a Reinforcement Learning setting with question-answering based rewards. We evaluate the summaries generated by the this framework using multiple automatic measures and human judgements. The experimental results show that the question-answering rewards can be used as a general framework to improve neural abstractive summarization. Particularly, the results from human evaluations show that the summaries generated by our approach is preferred over 30% of the time over the summaries generated by general abstractive summarization models.

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.

A frequent pattern in customer care conversations is the agents responding with appropriate webpage URLs that address users’ needs. We study the task of predicting the documents that customer care agents can use to facilitate users’ needs. We also introduce a new public dataset which supports the aforementioned problem. Using this dataset and two others, we investigate state-of-the art deep learning (DL) and information retrieval (IR) models for the task. Additionally, we analyze the practicality of such systems in terms of inference time complexity. Our show that an hybrid IR+DL approach provides the best of both worlds.

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.

Customer support agents play a crucial role as an interface between an organization and its end-users. We propose CAIRAA: Conversational Approach to Information Retrieval for Agent Assistance, to reduce the cognitive workload of support agents who engage with users through conversation systems. CAIRAA monitors an evolving conversation and recommends both responses and URLs of documents the agent can use in replies to their client. We combine traditional information retrieval (IR) approaches with more recent Deep Learning (DL) models to ensure high accuracy and efficient run-time performance in the deployed system. Here, we describe the CAIRAA system and demonstrate its effectiveness in a pilot study via a short video.

Disentangling conversations mixed together in a single stream of messages is a difficult task, made harder by the lack of large manually annotated datasets. We created a new dataset of 77,563 messages manually annotated with reply-structure graphs that both disentangle conversations and define internal conversation structure. Our data is 16 times larger than all previously released datasets combined, the first to include adjudication of annotation disagreements, and the first to include context. We use our data to re-examine prior work, in particular, finding that 89% of conversations in a widely used dialogue corpus are either missing messages or contain extra messages. Our manually-annotated data presents an opportunity to develop robust data-driven methods for conversation disentanglement, which will help advance dialogue research.

Goal-oriented dialogue in complex domains is an extremely challenging problem and there are relatively few datasets. This task provided two new resources that presented different challenges: one was focused but small, while the other was large but diverse. We also considered several new variations on the next utterance selection problem: (1) increasing the number of candidates, (2) including paraphrases, and (3) not including a correct option in the candidate set. Twenty teams participated, developing a range of neural network models, including some that successfully incorporated external data to boost performance. Both datasets have been publicly released, enabling future work to build on these results, working towards robust goal-oriented dialogue systems.