Ireland

A long-term ambition of information seeking QA systems is to reason over multi-modal contexts and generate natural answers to user queries. Today, memory intensive pre-trained language models are adapted to downstream tasks such as QA by fine-tuning the model on QA data in a specific modality like unstructured text or structured tables. To avoid training such memory-hungry models while utilizing a uniform architecture for each modality, parameter-efficient adapters add and train small task-specific bottle-neck layers between transformer layers. In this work, we study parameter-efficient abstractive QA in encoder-decoder models over structured tabular data and unstructured textual data using only 1.5% additional parameters for each modality. We also ablate over adapter layers in both encoder and decoder modules to study the efficiency-performance trade-off and demonstrate that reducing additional trainable parameters down to 0.7%-1.0% leads to comparable results. Our models out-perform current state-of-the-art models on tabular QA datasets such as Tablesum and FeTaQA, and achieve comparable performance on a textual QA dataset such as NarrativeQA using significantly less trainable parameters than fine-tuning.

ACL 2022

Parameter-Efficient Abstractive Question Answering over Tables or Text

textqa

tabularqa

adapters

language generation

# Welcome everyone to ACL 2022!

The 60th Annual Meeting of the Association for Computational Linguistics is taking place May 22-27, 2022 as a hybrid event, in Dublin and online. We are happy to welcome all of you to this anniversary edition with an almost 50-50 in-person and virtual participation. 
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, three amazing keynote events and two new initiatives of invited talks: Spotlight Talks for Young Rising Stars (STIRS) and The Next Big Idea Talks. Posters (including Findings of ACL 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 will have an oral session and a poster session as part of Poster Session 1. The program also features eight Tutorials and 28 Workshops. 

 
We wish you a wonderful conference! 
[**The ACL 2022 Organizing Committee**](https://www.2022.aclweb.org/organisers)
 
[**Conference Handbook**](https://drive.google.com/file/d/1_BUCMfhMVrjG9E2e71aHdHeE28KSje0l/view?usp=sharing) 
[**Mini Handbook**](https://drive.google.com/file/d/1qlBKl0wzmlVF1oCeMQl3BahLd9nLP5Ce/view?usp=sharing) 
[**Posters and Demo guides**](https://drive.google.com/file/d/1UucMAoCNncIOaH1rMMDa0owuG9qgvJTG/view?usp=sharing)

The Association for Computational Linguistics (ACL) is the premier international scientific and professional society for people working on computational problems involving human language, a field often referred to as either computational linguistics or natural language processing (NLP). 

workshop paper

When multiple conversations occur simultaneously, a listener must decide which conversation each utterance is part of in order to interpret and respond to it appropriately. 
This task is referred as dialogue disentanglement.
A significant drawback of previous studies on disentanglement lies in that they only focus on pair-wise relationships between utterances while neglecting the conversation structure which is important for conversation structure modeling.
In this paper, we propose a hierarchical model, named Dialogue BERT (DialBERT), which integrates the local and global semantics in the context range by using BERT to encode each message-pair and using BiLSTM to aggregate the chronological context information into the output of BERT.
In order to integrate the conversation structure information into the model, two types of loss of conversation-structure loss and tree-structure loss are designed. 
In this way, our model can implicitly learn and leverage the conversation structures without being restricted to the lack of explicit access to such structures during the inference stage. 
Experimental results on two large datasets show that our method outperforms previous methods by substantial margins, achieving great performance on dialogue disentanglement.

Conversation- and Tree-Structure Losses for Dialogue Disentanglement

We deal with the scenario of conversational search, where user queries are under-specified or ambiguous. This calls for a mixed-initiative setup. User-asks (queries) and system-answers, as well as system-asks (clarification questions) and user response, in order to clarify her information needs. 
We focus on the task of selecting the next clarification question, given the conversation context. Our method leverages passage retrieval from a background content to fine-tune two deep-learning models for ranking candidate clarification questions. We evaluated our method on two different use-cases. The first is an open domain conversational search in a large web collection. The second is a task-oriented customer-support setup.
We show that our method performs well on both use-cases.

Conversational Search with Mixed-Initiative - Asking Good Clarification Questions backed-up by Passage Retrieval

Coreference resolution such as for anaphora has been an essential challenge that is commonly found in conversational machine reading comprehension (CMRC). This task aims to determine the referential entity to which a pronoun refers on the basis of contextual information. Existing approaches based on pre-trained language models (PLMs) mainly rely on an end-to-end method, which still has limitations in clarifying referential dependency. In this study, a novel graph-based approach is proposed to integrate the coreference of given text into graph structures (called coreference graphs), which can pinpoint a pronoun's referential entity. We propose two graph-combined methods, evidence-enhanced and the fusion model, for CMRC to integrate coreference graphs from different levels of the PLM architecture. Evidence-enhanced refers to textual level methods that include an evidence generator (for generating new text to elaborate a pronoun) and enhanced question (for rewriting a pronoun in a question) as PLM input. The fusion model is a structural level method that combines the PLM with a graph neural network. We evaluated these approaches on a CoQA pronoun-containing dataset and the whole CoQA dataset. The result showed that our methods can outperform baseline PLM methods with BERT and RoBERTa.

Graph-combined Coreference Resolution Methods on Conversational Machine Reading Comprehension with Pre-trained Language Model

We work on a recommendation dialogue system to help a user understand the appealing points of some target (e.g., a movie). In such dialogues, the recommendation system needs to utilize structured external knowledge to make informative and detailed recommendations. However, there is no dialogue dataset with structured external knowledge designed to make detailed recommendations for the target. Therefore, we construct a dialogue dataset, Japanese Movie Recommendation Dialogue (JMRD), in which the recommender recommends one movie in a long dialogue (23 turns on average). The external knowledge used in this dataset is hierarchically structured, including title, casts, reviews, and plots. Every recommender’s utterance is associated with the external knowledge related to the utterance. We then create a movie recommendation dialogue system that considers the structure of the external knowledge and the history of the knowledge used. Experimental results show that the proposed model is superior in knowledge selection to the baseline models.

Construction of Hierarchical Structured Knowledge-based Recommendation Dialogue Dataset and Dialogue System

To diversify and enrich generated dialogue responses, knowledge-grounded dialogue has been investigated in recent years. The existing methods tackle the knowledge grounding challenge by retrieving the relevant sentences over a large corpus and augmenting the dialogues with explicit extra information. Despite their success, however, the existing works have drawbacks in inference efficiency. This paper proposes KnowExpert, a framework to bypass the explicit retrieval process and inject knowledge into the pre-trained language models with lightweight adapters and adapt to the knowledge-grounded dialogue task. To the best of our knowledge, this is the first attempt to tackle this challenge without retrieval in this task under an open-domain chit-chat scenario. The experimental results show that Knowexpert performs comparably with some retrieval-based baselines while being time-efficient in inference, demonstrating the effectiveness of our proposed method.

Retrieval-Free Knowledge-Grounded Dialogue Response Generation with Adapters

This work presents our contribution from the Text-to-Knowledge team of Ghent University (UGent-T2K) to the MultiDoc2Dial shared task on modeling dialogs grounded in multiple documents. We propose a pipeline system, comprising (1) document retrieval, (2) passage retrieval, and (3) response generation. We engineered these individual components mainly by, for (1)-(2), combining multiple ranking models and adding a final LambdaMART reranker, and, for (3), by adopting a Fusion-in-Decoder (FiD) model. We thus significantly boost the baseline system’s performance (over +10 points for both F1 and SacreBLEU). Further, error analysis reveals two major failure cases, to be addressed in future work: (i) in case of topic shift within the dialog, retrieval often fails to select the correct grounding document(s), and (ii) generation sometimes fails to use the correctly retrieved grounding passage.

UGent-T2K at the 2nd DialDoc Shared Task: A Retrieval-Focused Dialog System Grounded in Multiple Documents

MultiDoc2Dial presents an important challenge on modeling dialogues grounded with multiple documents. This paper proposes a pipeline system of "retrieve, re-rank, and generate", where each component is individually optimized. This enables the passage re-ranker and response generator to fully exploit training with ground-truth data. Furthermore, we use a deep cross-encoder trained with localized hard negative passages from the retriever. For the response generator, we use grounding span prediction as an auxiliary task to be jointly trained with the main task of response generation. We also adopt a passage dropout and regularization technique to improve response generation performance. Experimental results indicate that the system clearly surpasses the competitive baseline and our team CPII-NLP ranked 1st on the seen test and 2nd on the unseen test of this challenge.

Grounded Dialogue Generation with Cross-encoding Re-ranker, Grounding Span Prediction, and Passage Dropout

This presentation discusses our proposed approach, Docalog, for the DialDoc-22 (Multi-Doc2Dial) shared task. Docalog identifies themost relevant knowledge in the associated document, in a multi-document setting. Docalog, is a three-stage pipeline consisting of (1) a document retriever model (DR. TEIT), (2) an answer span prediction model, and (3) an ultimate span picker deciding on the most likely answer span, out of all predicted spans. In the test phase of MultiDoc2Dial 2022, Docalog achieved f1-scores of 36.07% and 28.44% and SacreBLEU scores of 23.70% and 20.52%, respectively on the MDD-SEEN and MDD-UNSEEN folds.

Docalog: Multi-document Dialogue System using Transformer-based Span Retrieval

In this presentation we describe the speech corpus for the Siberian Ingrian Finnish language. The speech corpus includes: audio data, annotations, software tools for data-processing, two databases and a web application. We have published part of the audio data and annotations. The software tool for parsing annotation files and feeding a relational database is developed and published under a free license. A web application is developed and available. At this moment, about 300 words and 200 phrases can be displayed using this web application.

Development of the Siberian Ingrian Finnish Speech Corpus

Focus on language-specific properties with insights from formal minimalist syntax can improve universal dependency (UD) parsing. Such improvements are especially sensitive for low-resource African languages, like Wolof, which have fewer UD treebanks in number and amount of annotations, and fewer con- tributing annotators. For two different UD parser pipelines, one parser model was trained on the original Wolof treebank, and one was trained on an edited treebank. For each parser pipeline, the accuracy of the edited treebank was higher than the original for both the de- pendency relations and dependency labels. Ac- curacy for universal dependency relations im- proved as much as 2.90%, while accuracy for universal dependency labels increased as much as 3.38%. An annotation scheme that better fits a language’s distinct syntax results in better parsing accuracy.

Parameter-Efficient Abstractive Question Answering over Tables or Text

Downloads

Next from ACL 2022

Conversation- and Tree-Structure Losses for Dialogue Disentanglement

Similar lecture

AraT5: Text-to-Text Transformers for Arabic Language Generation

Parameter-Efficient Abstractive Question Answering over Tables or Text

.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 ACL 2022

Conversation- and Tree-Structure Losses for Dialogue Disentanglement

Similar lecture

AraT5: Text-to-Text Transformers for Arabic Language Generation

Downloads