Croatia

There has been a surge of interest regarding the alignment of large-scale language models with human language comprehension behaviour. The majority of this research investigates comprehension behaviours from reading isolated, written sentences. We propose studying the perception of dialogue, focusing on an intrinsic form of language use: spoken conversations. 

Using the task of predicting upcoming dialogue turns, we ask whether turn plausibility scores produced by state-of-the-art language models correlate with human judgements. We find a strong correlation for some but not all models: masked language models produce stronger correlations than auto-regressive models. In doing so, we quantify human performance on the response selection task for open-domain spoken conversation. To the best of our knowledge, this is the first such quantification.

We find that response selection performance can be used as a coarse proxy for the strength of correlation with human judgements, however humans and models make different response selection mistakes. The model which produces the strongest correlation also outperforms human response selection performance. Through ablation studies, we show that pre-trained language models provide a useful basis for turn representations; however, fine-grained contextualisation, inclusion of dialogue structure information, and fine-tuning towards response selection all boost response selection accuracy by over 30 absolute points.

EACL 2023

Do dialogue representations align with perception? An empirical study

psychometric predictive power

dialogue acceptability

perception

dialogue

poster

### Welcome to the 17th Conference of the European Chapter of the Association for Computational Linguistics (EACL) 
Continuing its mission of expanding and involving the science community of all European countries, EACL had selected the Ukrainian community for the 16th EACL, which took place online due to the COVID pandemic. Unfortunately, the ongoing war made the organisation in Kyiv impossible. Considering the importance of physical interaction among researchers, especially after the restrictions imposed by the COVID pandemic, in addition to an online mode, the conference will be held in Dubrovnik, Croatia, from 2 to 6 of May, 2023. The original aim of strengthening the connection with the Ukrainian community will not change as our program will feature a dedicated session and a workshop to highlight work on Ukrainian language technologies. As the flagship European conference in the field of computational linguistics, EACL welcomes European and international researchers covering a broad spectrum of research areas that are concerned with computational approaches to natural language.

You need to register for the conference in order to access this site. Please visit https://2023.eacl.org/registration for more information.

As the flagship European conference in the field of computational linguistics, EACL welcomes European and international researchers covering a broad spectrum of research areas that are concerned with computational approaches to natural language.

Large pre-trained language models have achieved impressive results on various style classification tasks, but they often learn spurious domain-specific words to make predictions (Hayati et al., 2021). While human explanation highlights stylistic tokens as important features for this task, we observe that model explanations often do not align with them. To tackle this issue, we introduce StyLEx, a model that learns from human annotated explanations of stylistic features and jointly learns to perform the task and predict these features as model explanations. Our experiments show that StyLEx can provide human like stylistic lexical explanations without sacrificing the performance of sentence-level style prediction on both in-domain and out-of-domain datasets. Explanations from StyLEx show significant improvements in explanation metrics (sufficiency, plausibility) and when evaluated with human annotations. They are also more understandable by human judges compared to the widely-used saliency-based explanation baseline.

StyLEx: Explaining Style Using Human Lexical Annotations

Temporal concept drift refers to the problem of data changing over time. In the field of NLP, that would entail that language (e.g. new expressions, meaning shifts) and factual knowledge (e.g. new concepts, updated facts) evolve over time. Focusing on the latter, we benchmark 11 pretrained masked language models (MLMs) on a series of tests designed to evaluate the effect of temporal concept drift, as it is crucial that widely used language models remain up-to-date with the ever-evolving factual updates of the real world. Specifically, we provide a holistic framework that (1) dynamically creates temporal test sets of any time granularity (e.g. month, quarter, year) of factual data from Wikidata, (2) constructs fine-grained splits of tests (e.g. updated, new, unchanged facts) to ensure comprehensive analysis, and (3) evaluates MLMs in three distinct ways (single-token probing, multi-token generation, MLM scoring). In contrast to prior work, our framework aims to unveil how robust an MLM is over time and thus to provide a signal in case it has become outdated, by leveraging multiple views of evaluation.

Dynamic Benchmarking of Masked Language Models on Temporal Concept Drift with Multiple Views

Recent advances in discourse parsing performance create the impression that, as in other NLP tasks, performance for high-resource languages such as English is finally becoming reliable. In this paper we demonstrate that this is not the case, and thoroughly investigate the impact of data diversity on RST parsing stability. We show that state-of-the-art architectures trained on the standard English newswire benchmark do not generalize well, even within the news domain. Using the two largest RST corpora of English with text from multiple genres, we quantify the impact of genre diversity in training data for achieving generalization to text types unseen during training. Our results show that a heterogeneous training regime is critical for stable and generalizable models, across parser architectures. We also provide error analyses of model outputs and out-of-domain performance. To our knowledge, this study is the first to fully evaluate cross-corpus RST parsing generalizability on complete trees, examine between-genre degradation within an RST corpus, and investigate the impact of genre diversity in training data composition.

Why Can't Discourse Parsing Generalize? A Thorough Investigation of the Impact of Data Diversity

Multi-hop reasoning, a prevalent approach for query answering, aims at inferring new facts along reasonable paths over a knowledge graph.
Reinforcement learning methods can be adopted by formulating the problem into a Markov decision process.
However, common suffering within RL-based reasoning models is that the agent can be biased to spurious paths which coincidentally lead to the correct answer with poor explanation.
In this work, we take a deep dive into this phenomenon and define a metric named Path Spuriousness (PS), to quantitatively estimate to what extent a path is spurious.
Guided by the definition of PS, we design a model with a new reward that considers both answer accuracy and path reasonableness.
We test our method on four datasets and experiments reveal that our method considerably enhances the agent's capacity to prevent spurious paths while keeping comparable to state-of-the-art performance.

Path Spuriousness-aware Reinforcement Learning for Multi-Hop Knowledge Graph Reasoning

Recent advances in the capacity of large language models to generate human-like text have resulted in their increased adoption in user-facing settings. In parallel, these improvements have prompted a heated discourse around the risks of societal harms they introduce, whether inadvertent or malicious. Several studies have explored these harms and called for their mitigation via development of safer, fairer models. Going beyond enumerating the risks of harms, this work provides a survey of practical methods for addressing potential threats and societal harms from language generation models. We draw on several prior works' taxonomies of language model risks to present a structured overview of strategies for detecting and ameliorating different kinds of risks/harms of language generators. Bridging diverse strands of research, this survey aims to serve as a practical guide for both LM researchers and practitioners, with explanations of different strategies' motivations, their limitations, and open problems for future research.

Language Generation Models Can Cause Harm: So What Can We Do About It? An Actionable Survey

Self-attention weights and their transformed variants have been the main source of information for analyzing token-to-token interactions in Transformer-based models. But despite their ease of interpretation, these weights are not faithful to the models' decisions as they are only one part of an encoder, and other components in the encoder layer can have considerable impact on information mixing in the output representations. In this work, by expanding the scope of analysis to the whole encoder block, we propose Value Zeroing, a novel context mixing score customized for Transformers that provides us with a deeper understanding of how information is mixed at each encoder layer. We demonstrate the superiority of our context mixing score over other analysis methods through a series of complementary evaluations with different viewpoints based on linguistically informed rationales, probing, and faithfulness analysis.

Quantifying Context Mixing in Transformers

A promising approach for knowledge-based Word Sense Disambiguation (WSD) is to select the sense whose contextualized embeddings computed for its definition sentence are closest to those computed for a target word in a given sentence. This approach relies on the similarity of the sense and context embeddings computed by a pre-trained language model. We propose a semantic specialization for WSD where contextualized embeddings are adapted to the WSD task using solely lexical knowledge. The key idea is, for a given sense, to bring semantically related senses and contexts closer and send different/unrelated senses farther away. We realize this idea as the joint optimization of the Attract-Repel objective for sense pairs and the self-training objective for context-sense pairs while controlling deviations from the original embeddings. The proposed method outperformed previous studies that adapt contextualized embeddings. It achieved state-of-the-art performance on knowledge-based WSD when combined with the reranking heuristic that uses the sense inventory. We found that the similarity characteristics of specialized embeddings conform to the key idea. We also found that the (dis)similarity of embeddings between the related/different/unrelated senses correlates well with the performance of WSD.

Semantic Specialization for Knowledge-based Word Sense Disambiguation

The recent explosion of question-answering (QA) datasets and models has increased the interest in the generalization of models across multiple domains and formats by either training on multiple datasets or combining multiple models. Despite the promising results of multi-dataset models, some domains or QA formats may require specific architectures, and thus the adaptability of these models might be limited. In addition, current approaches for combining models disregard cues such as question-answer compatibility. In this work, we propose to combine expert agents with a novel, flexible, and training-efficient architecture that considers questions, answer predictions, and answer-prediction confidence scores to select the best answer among a list of answer predictions. Through quantitative and qualitative experiments, we show that our model i) creates a collaboration between agents that outperforms previous multi-agent and multi-dataset approaches, ii) is highly data-efficient to train, and iii) can be adapted to any QA format. We release our code and a dataset of answer predictions from expert agents for 16 QA datasets to foster future research of multi-agent systems.

MetaQA: Combining Expert Agents for Multi-Skill Question Answering

Creating an abridged version of a text involves shortening it while maintaining its linguistic qualities. In this paper, we examine this task from an NLP perspective for the first time. We present a new resource, AbLit, which is derived from abridged versions of English literature books. The dataset captures passage-level alignments between the original and abridged texts. We characterize the linguistic relations of these alignments, and create automated models to predict these relations as well as to generate abridgements for new texts. Our findings establish abridgement as a challenging task, motivating future resources and research. The dataset is available at github.com/roemmele/AbLit.

AbLit: A Resource for Analyzing and Generating Abridged Versions of English Literature

Providing natural language instructions in prompts is a useful new paradigm for improving task performance of large language models in a zero-shot setting. Recent work has aimed to improve such prompts via manual rewriting or gradient-based tuning. However, manual rewriting is time-consuming and requires subjective interpretation, while gradient-based tuning can be extremely computationally demanding for large models and may not be feasible for API-based models. In this work, we introduce Gradient-free Instructional Prompt Search (GrIPS), a gradient-free, edit-based search approach for improving task instructions for large language models. GrIPS takes in instructions designed for humans and automatically returns an improved, edited prompt, while allowing for API-based tuning. With InstructGPT models, GrIPS improves the average task performance by up to 4.30 percentage points on eight classification tasks from the Natural Instructions dataset (with similar improvements for OPT, BLOOM, and FLAN-T5). We see improvements for both instruction-only prompts and instruction + k-shot examples prompts. Notably, GrIPS outperforms manual rewriting and purely example-based prompts while controlling for the available compute and data budget. Further, performance of GrIPS is comparable to select gradient-based tuning approaches. Qualitatively, we show our edits can simplify instructions and at times make them incoherent but nonetheless improve accuracy.

Downloads

Next from EACL 2023

StyLEx: Explaining Style Using Human Lexical Annotations