Dominican Republic

Transformer-based pre-training techniques of text and layout have proven effective in a number of document understanding tasks. Despite this success, multimodal pre-training models suffer from very high computational and memory costs. Motivated by human reading strategies, this paper presents Skim-Attention, a new attention mechanism that takes advantage of the structure of the document and its layout. Skim-Attention only attends to the 2-dimensional position of the words in a document. Our experiments show that Skim-Attention obtains a lower perplexity than prior works, while being more computationally efficient. Skim-Attention can be further combined with long-range Transformers to efficiently process long documents. We also show how Skim-Attention can be used off-the-shelf as a mask for any Pre-trained Language Model, allowing to improve their performance while restricting attention. Finally, we show the emergence of a document structure representation in Skim-Attention.

EMNLP 2021

Skim-Attention: Learning to Focus via Document Layout

EMNLP 2021 is planned to be a hybrid event in Punta Cana, Dominican Republic, with both on-site and fully virtual participation possible. The experience for on-site participants would closely approximate a normal pre-COVID *ACL conference, with 5-6 thematically organized parallel sessions and live Q/A and interactive discussion immediately after the talks. Presentations by virtual participants will be equitably interleaved with those of on-site participants, projected on the auditorium screens as if on-site, and also followed immediately by live Q/A and interactive discussion at a time during reasonable waking hours for the virtual presenter. For all participants, on-site and virtual, who are unable to attend a session due to either time-zone issues or because they are participating in another session live, talk recordings and slides will be available online at a minimum after the live presentation (and in many cases before as well), and questions may be submitted in advance on session-specific discussion boards and answered live in session with the usual visual aids if desired.

<iframe style="width:700px;height:400px" src="https://online.fliphtml5.com/ebtyf/ceby/"  seamless="seamless" scrolling="no" frameborder="0" allowtransparency="true" allowfullscreen="true" ></iframe>

Please Note: The EMNLP registration system is not currently connected to the underline site as we are still in the process of building out EMNLP 2021. You will receive access instructions from underline the week of November 1st. 

Access is given only to EMNLP upon registration, if you have not registered please do so [here](https://2021.emnlp.org/registration).

Registered attendees will receive access the week of November 1st.

EMNLP 2021 is planned to be a hybrid event in Punta Cana, Dominican Republic, with both on-site and fully virtual participation possible.

technical paper

Self-supervised learning has recently attracted considerable attention in the NLP community for its ability to learn discriminative features using a contrastive objective. This paper investigates whether contrastive learning can be extended to Transfomer attention to tackling the Winograd Schema Challenge. To this end, we propose a novel self-supervised framework, leveraging a contrastive loss directly at the level of self-attention. Experimental analysis of our attention-based models on multiple datasets demonstrates superior commonsense reasoning capabilities. The proposed approach outperforms all comparable unsupervised approaches while occasionally surpassing supervised ones.

Attention-based Contrastive Learning for Winograd Schemas

In this paper, we present a new method for training a writing improvement model adapted to the writer’s first language (L1) that goes beyond grammatical error correction (GEC). Without using annotated training data, we rely solely on pre-trained language models fine-tuned with parallel corpora of reference translation aligned with machine translation. We evaluate our model with corpora of academic papers written in English by L1 Portuguese and L1 Spanish scholars and a reference corpus of expert academic English. We show that our model is able to address specific L1-influenced writing and more complex linguistic phenomena than existing methods, outperforming what a state-of-the-art GEC system can achieve in this regard. Our code and data are open to other researchers.

Beyond Grammatical Error Correction: Improving L1-influenced research writing in English using pre-trained encoder-decoder models

Online tenant reviews of multifamily residential properties present a unique source of information for commercial real estate investing and research. Real estate professionals frequently read tenant reviews to uncover property-related issues that are otherwise difficult to detect, a process that is both biased and time-consuming. Using this as motivation, we asked whether a text classification-based approach can automate the detection of four carefully defined, major quality-of-life issues: severe crime, noise nuisance, pest burden, and parking difficulties. We aggregate 5.5 million tenant reviews from five sources and use two-stage crowdsourced labeling on 0.1% of the data to produce high-quality labels for subsequent text classification. Following fine-tuning of pretrained language models on millions of reviews, we train a multi-label reviews classifier that achieves a mean AUROC of 0.965 on these labels. We next use the model to reveal temporal and spatial patterns among tens of thousands of multifamily properties. Collectively, these results highlight the feasibility of automated analysis of housing trends and investment opportunities using tenant-perspective data.

Classification and Geotemporal Analysis of Quality-of-Life Issues in Tenant Reviews

Inspired by Curriculum Learning, we propose a consecutive (i.e., image-to-text-to-text) generation framework where we divide the problem of radiology report generation into two steps. Contrary to generating the full radiology report from the image at once, the model generates global concepts from the image in the first step and then reforms them into finer and coherent texts using transformer-based architecture. We follow the transformer-based sequence-to-sequence paradigm at each step. We improve upon the state-of-the-art on two benchmark datasets.

Progressive Transformer-Based Generation of Radiology Reports

Natural language models often fall short when understanding and generating mathematical notation. 
What is not clear is whether these shortcomings are due to fundamental limitations of the models, or the absence of appropriate tasks. In this paper, we explore the extent to which natural language models can learn semantics between mathematical notation and their surrounding text. We propose two notation prediction tasks, and train a model that selectively masks notation tokens and encodes left and/or right sentences as context. Compared to baseline models trained by masked language modeling, our method achieved significantly better performance at the two tasks, showing that this approach is a good first step towards modeling mathematical texts. However, the current models rarely predict unseen symbols correctly, and token-level predictions are more accurate than symbol-level predictions, indicating more work is needed to represent structural patterns. Based on the results, we suggest future works toward modeling mathematical texts.

Modeling Mathematical Notation Semantics in Academic Papers

Pre-trained language models (PLMs) like BERT have made great progress in NLP. News articles usually contain rich textual information, and PLMs have the potentials to enhance news text modeling for various intelligent news applications like news recommendation and retrieval. However, most existing PLMs are in huge size with hundreds of millions of parameters. Many online news applications need to serve millions of users with low latency tolerance, which poses great challenges to incorporating PLMs in these scenarios. Knowledge distillation techniques can compress a large PLM into a much smaller one and meanwhile keeps good performance. However, existing language models are pre-trained and distilled on general corpus like Wikipedia, which has gaps with the news domain and may be suboptimal for news intelligence. In this paper, we propose NewsBERT, which can distill PLMs for efficient and effective news intelligence. In our approach, we design a teacher-student joint learning and distillation framework to collaboratively learn both teacher and student models, where the student model can learn from the learning experience of the teacher model. In addition, we propose a momentum distillation method by incorporating the gradients of teacher model into the update of student model to better transfer the knowledge learned by the teacher model. Thorough experiments on two real-world datasets with three tasks show that NewsBERT can empower various intelligent news applications with much smaller models.

NewsBERT: Distilling Pre-trained Language Model for Intelligent News Application

This paper proposes a transformer over transformer framework, called Transformer^2, to perform neural text segmentation. It consists of two components: bottom-level sentence encoders using pre-trained transformers, and an upper-level transformer-based segmentation model based on the sentence embeddings. The bottom-level component transfers the pre-trained knowledge learnt from large external corpora under both single and pair-wise supervised NLP tasks to model the sentence embeddings for the documents. Given the sentence embeddings, the upper-level transformer is trained to recover the segmentation boundaries as well as the topic labels of each sentence. Equipped with a multi-task loss and the pre-trained knowledge, Transformer^2 can better capture the semantic coherence within the same segments. Our experiments show that (1) Transformer^2$manages to surpass state-of-the-art text segmentation models in terms of a commonly-used semantic coherence measure; (2) in most cases, both single and pair-wise pre-trained knowledge contribute to the model performance; (3) bottom-level sentence encoders pre-trained on specific languages yield better performance than those pre-trained on specific domains.

Transformer over Pre-trained Transformer for Neural Text Segmentation with Enhanced Topic Coherence

Adverse Events (AE) are harmful events resulting from the use of medical products. Although social media may be crucial for early AE detection, the sheer scale of this data makes it logistically intractable to analyze using human agents, with NLP representing the only low-cost and scalable alternative.

In this paper, we frame AE Detection and Extraction as a sequence-to-sequence problem using the T5 model architecture and achieve strong performance improvements over the baselines on several English benchmarks (F1 = 0.71, 12.7% relative improvement for AE Detection; Strict F1 = 0.713, 12.4% relative improvement for AE Extraction). Motivated by the strong commonalities between AE tasks, the class imbalance in AE benchmarks, and the linguistic and structural variety typical of social media texts, we propose a new strategy for multi-task training that accounts, at the same time, for task and dataset characteristics. Our approach increases model robustness, leading to further performance gains. Finally, our framework shows some language transfer capabilities, obtaining higher performance than Multilingual BERT in zero-shot learning on French data.

Exploring a Unified {S}equence-{T}o-{S}equence {T}ransformer for Medical Product Safety Monitoring in Social Media

The relationships that exist between entities can be a reliable indicator for classifying sensitive information, such as commercially sensitive information. For example, the relation person-IsDirectorOf-company can indicate whether an individual’s salary should be considered as sensitive personal information. Representations of such relations are often learned using a knowledge graph to produce embeddings for relation types, generalised across different entity-pairs. However, a relation type may or may not correspond to a sensitivity depending on the entities that participate to the relation. Therefore, generalised relation embeddings are typically insufficient for classifying sensitive information. In this work, we propose a novel method for representing entities and relations within a single embedding to better capture the relationship between the entities. Moreover, we show that our proposed entity-relation-entity embedding approach can significantly improve (McNemar’s test, p <0.05) the effectiveness of sensitivity classification, compared to classification approaches that leverage relation embedding approaches from the literature. (0.426 F1 vs 0.413 F1)

RelDiff: Enriching Knowledge Graph Relation Representations for Sensitivity Classification

Empathy is the link between self and others. Detecting and understanding empathy is a key element for improving human-machine interaction. However, annotating data for detecting empathy at a large scale is a challenging task. This paper employs multi-task training with knowledge distillation to incorporate knowledge from available resources (emotion and sentiment) to detect empathy from the natural language in different domains. This approach yields better results on an existing news-related empathy dataset compared to strong baselines. In addition, we build a new dataset for empathy prediction with fine-grained empathy direction, seeking or providing empathy, from Twitter. We release our dataset for research purposes.

Skim-Attention: Learning to Focus via Document Layout

Downloads

Next from EMNLP 2021

Attention-based Contrastive Learning for Winograd Schemas

Similar lecture

HittER: Hierarchical Transformers for Knowledge Graph Embeddings

Skim-Attention: Learning to Focus via Document Layout

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from EMNLP 2021

Attention-based Contrastive Learning for Winograd Schemas

Similar lecture

HittER: Hierarchical Transformers for Knowledge Graph Embeddings

Downloads