United States

Inference tasks such as answer sentence selection (AS2) or fact verification are typically solved by fine-tuning transformer-based models as individual sentence-pair classifiers. Recent studies show that these tasks benefit from modeling dependencies across multiple candidate sentences jointly. In this paper, we first show that popular pre-trained transformers perform poorly when used for fine-tuning on multi-candidate inference tasks. We then propose a new pre-training objective that models the paragraph-level semantics across multiple input sentences. Our evaluation on three AS2 and one fact verification datasets demonstrates the superiority of our pre-training technique over the traditional ones for transformers used as joint models for multi-candidate inference tasks, as well as when used as cross-encoders for sentence-pair formulations of these tasks.

NAACL 2022

Paragraph-based Transformer Pre-training for Multi-Sentence Inference

## Welcome to NAACL 2022!
The Annual Meeting of the North American Chapter of the Association for Computational Linguistics (NAACL) is taking place July 10-15, 2022 as a hybrid event, in Seattle, WA and online. We are happy to welcome all of you to conference! 
 
The main conference program features oral presentations, in-person and virtual posters and demo sessions, a plenary session for our best paper presentations and awards, keynote presentations, a plenary panel on the place of linguistics and symbolic structures in NLP, and an Industrial Track panel on Careers in NLP.
 
Posters (including Findings of NAACL 2022) and demos are grouped by areas for both the in-person and the virtual sessions. For the virtual component, the talks will be on Zoom and the posters and the demos will be in GatherTown. The Student Research Workshop and will have an oral session and a poster session.
 
The program also features 6 Tutorials and 26 Workshops.
 
 We wish you a wonderful conference!
 
[The NAACL 2022 Organizing Committee](https://2022.naacl.org/committees/organization/)

2022 Annual Conference of the North American Chapter of the Association for Computational Linguistics

poster

A document can be summarized in a number of ways. Reference-based evaluation of summarization has been criticized for its inflexibility. The more sufficient the number of abstracts, the more accurate the evaluation results. However, it is difficult to collect sufficient reference summaries. In this paper, we propose a new automatic reference-free evaluation metric that compares semantic distribution between source document and summary by pretrained language models and considers summary compression ratio. The experiments show that this metric is more consistent with human evaluation in terms of coherence, consistency, relevance and fluency.

Reference-free Summarization Evaluation via Semantic Correlation and Compression Ratio

Active learning, which effectively collects informative unlabeled data for annotation, reduces the demand for labeled data. In this work, we propose to retrieve unlabeled samples with a local sensitivity and hardness-aware acquisition function. The proposed method generates data copies through local perturbations and selects data points whose predictive likelihoods diverge the most from their copies. We further empower our acquisition function by injecting the select-worst case perturbation. Our method achieves consistent gains over the commonly used active learning strategies in various classification tasks. Furthermore, we observe consistent improvements over the baselines on the study of prompt selection in prompt-based few-shot learning. These experiments demonstrate that our acquisition guided by local sensitivity and hardness can be effective and beneficial for many NLP tasks.

ALLSH: Active Learning Guided by Local Sensitivity and Hardness

Distantly-supervised named entity recognition (NER) locates and classifies entities using only knowledge bases and unlabeled corpus to mitigate the reliance on human-annotated labels. The distantly annotated data suffer from the noise in labels, and previous works on DSNER have proved the importance of pre-refining distant labels with hand-crafted rules and extra existing semantic information. In this work, we explore the way to directly learn the distant label refinement knowledge by imitating annotations of different qualities and comparing these annotations in contrastive learning frameworks. the proposed distant label refinement model can give modified suggestions on distant data without additional supervised labels, and thus reduces the requirement on the quality of the knowledge bases. We perform extensive experiments and observe that recent and state-of-the-art DSNER methods gain evident benefits with our method.

Label Refinement via Contrastive Learning for Distantly-Supervised Named Entity Recognition

Multimodal named entity recognition and relation extraction (MNER and MRE) is a fundamental and crucial branch in information extraction. 
However, existing approaches for MNER and MRE usually suffer from error sensitivity when irrelevant object images incorporated in texts.
To deal with these issues, we propose a novel Hierarchical Visual Prefix fusion NeTwork (HVPNeT) for visual-enhanced entity and relation extraction, aiming to achieve more effective and robust performance. Specifically, we regard visual representation as pluggable visual prefix to guide  the textual representation for error insensitive forecasting decision. We further propose a dynamic gated aggregation strategy to achieve hierarchical multi-scaled visual features as visual prefix for fusion. Extensive  experiments on three benchmark datasets demonstrate the effectiveness of our method, and achieve  state-of-the-art performance.

Good Visual Guidance Make A Better Extractor: Hierarchical Visual Prefix for Multimodal Entity and Relation Extraction

Visual storytelling (VST) is the task of generating a story paragraph that describes a given image sequence. Most existing storytelling approaches have evaluated their models using traditional natural language generation metrics like BLEU or CIDEr. However, such metrics based on $n$-gram matching tend to have poor correlation with human evaluation scores and do not explicitly consider other criteria necessary for storytelling such as sentence structure or topic coherence. Moreover, a single score is not enough to assess a story as it does not inform us about what specific errors were made by the model. In this paper, we propose 3 evaluation metrics sets that analyses which aspects we would look for in a good story: 1) visual grounding, 2) coherence, and 3) non-redundancy. We measure the reliability of our metric sets by analysing its correlation with human judgement scores on a sample of machine stories obtained from 4 state-of-the-arts models trained on the Visual Storytelling Dataset (VIST). Our metric sets outperforms other metrics on human correlation, and could be served as a learning based evaluation metric set that is complementary to existing rule-based metrics.

RoViST: Learning Robust Metrics for Visual Storytelling

Matching model is essential for Image-Text Retrieval framework. Existing research usually train the model with a triplet loss and explore various strategy to retrieve hard negative sentences in the dataset. We argue that current retrieval-based negative sample construction approach is limited in the scale of the dataset thus fail to identify negative sample of high difficulty for every image. We propose our TAiloring neGative Sentences with Discrimination and Correction (TAGS-DC) to generate synthetic sentences automatically as negative samples. TAGS-DC is composed of masking and refilling to generate synthetic negative sentences with higher difficulty. To keep the difficulty during training, we mutually improve the retrieval and generation through parameter sharing. To further utilize fine-grained semantic of mismatch in the negative sentence, we propose two auxiliary tasks, namely word discrimination and word correction to improve the training. In experiments, we verify the effectiveness of our model on MS-COCO and Flickr30K compared with current state-of-the-art models and demonstrates its robustness and faithfulness in the further analysis.

Negative Sample is Negative in Its Own Way: Tailoring Negative Sentences for Image-Text Retrieval

Referring resolution is the task of identifying the referent of a natural language expression, for example "the woman behind the other woman getting a massage''. In this paper we investigate which are the kinds of referring expressions on which current transformer based models fail. Motivated by this analysis we identify the weakening of the spatial natural constraints as one of its causes and propose a model that aims to restore it. We evaluate our proposed model on different datasets for the task showing improved performance on the most challenging kinds of referring expressions. Finally we present a thorough analysis of the kinds errors that are improved by the new model and those that are not and remain future challenges for the task.

What kinds of errors do reference resolution models make and what can we learn from them?

Vision-and-Language Navigation (VLN) tasks require an agent to navigate through the environment based on language instructions. In this paper, we aim to solve two key challenges in this task: utilizing multilingual instructions for improved instruction-path grounding and navigating through new environments that are unseen during training. To address these challenges, first, our agent learns a shared and visually-aligned cross-lingual language representation for the three languages (English, Hindi and Telugu) in the Room-Across-Room dataset. Our language representation learning is guided by text pairs that are aligned by visual information. Second, our agent learns an environment-agnostic visual representation by maximizing the similarity between semantically-aligned image pairs (with constraints on object-matching) from different environments. Our environment agnostic visual representation can mitigate the environment bias induced by low-level visual information. Empirically, on the Room-Across-Room dataset, we show that our multi-lingual agent gets large improvements in all metrics over the strong baseline model when generalizing to unseen environments with the cross-lingual language representation and the environment-agnostic visual representation. Furthermore, we show that our learned language and visual representations can be successfully transferred to the Room-to-Room and Cooperative Vision-and-Dialogue Navigation task, and present detailed qualitative and quantitative generalization and grounding analysis.

CLEAR: Improving Vision-Language Navigation with Cross-Lingual, Environment-Agnostic Representations

Modern image captioning models are usually trained with text similarity objectives. However, since reference captions in public datasets often describe the most salient common objects, models trained with the text similarity objectives tend to ignore specific and detailed aspects of an image that distinguish it from others. Towards more descriptive and distinctive caption generation, we propose to use CLIP, a multimodal encoder trained on huge image-text pairs from the web, to calculate multi-modal similarity and use it as a reward function. We also propose a simple finetuning strategy of CLIP text encoder to improve grammar that does not require extra text annotation. This completely eliminates the need for reference captions during the reward computation. To comprehensively evaluate descriptive captions, we introduce FineCapEval, a new dataset for caption evaluation with fine-grained criteria: overall, background, object, relations. In our experiments on text-to-image retrieval and FineCapEval, the proposed CLIP-guided model generates more distinctive captions than the CIDEroptimized model. We also show that our unsupervised grammar finetuning of the CLIP text encoder alleviates the degeneration problem of the naive CLIP reward. Lastly, we show human analysis where the annotators strongly prefer CLIP reward to CIDEr and MLE objectives on diverse criteria.

Fine-grained Image Captioning with CLIP Reward

Automatic extraction of narrative elements from text, combining narrative theories with computational models, has been receiving increasing attention over the last few years. Previous works have utilized the oral narrative theory by Labov and Waletzky to identify various narrative elements in personal stories texts. Instead, we direct our focus to informational texts, specifically news stories. 

We introduce NEAT (Narrative Elements AnnoTation) -- a novel NLP task for detecting narrative elements in raw text. For this purpose, we designed a new multi-label narrative annotation scheme, better suited for informational text (e.g. news media), by adapting elements from the narrative theory of Labov and Waletzky (Complication and Resolution) and adding a new narrative element of our own (Success). We then used this scheme to annotate a new dataset of 2,209 sentences, compiled from 46 news articles from various category domains. We trained a number of supervised models in several different setups over the annotated dataset to identify the different narrative elements, achieving an average $F_1$ score of up to 0.77. The results demonstrate the holistic nature of our annotation scheme as well as its robustness to domain category.


Paragraph-based Transformer Pre-training for Multi-Sentence Inference

Downloads

Next from NAACL 2022

Reference-free Summarization Evaluation via Semantic Correlation and Compression Ratio

Similar lecture

CoSe-Co: Text Conditioned Generative CommonSense Contextualizer