Lea Frermann

Xudong Han, Aili Shen, Yitong Li et al.

This paper presents fairlib, an open-source framework for assessing and improving classification fairness. It provides a systematic framework for quickly reproducing existing baseline models, developing new methods, evaluating models with different metrics, and visualizing their results. Its modularity and extensibility enable the framework to be used for diverse types of inputs, including natural language, images, and audio. In detail, we implement 14 debiasing methods, including pre-processing, at-training-time, and post-processing approaches. The built-in metrics cover the most commonly used fairness criterion and can be further generalized and customized for fairness evaluation.

Uri Berger, Gabriel Stanovsky, Omri Abend et al.

Recent advances in self-supervised modeling of text and images open new opportunities for computational models of child language acquisition, which is believed to rely heavily on cross-modal signals. However, prior studies have been limited by their reliance on vision models trained on large image datasets annotated with a pre-defined set of depicted object categories. This is (a) not faithful to the information children receive and (b) prohibits the evaluation of such models with respect to category learning tasks, due to the pre-imposed category structure. We address this gap, and present a cognitively-inspired, multimodal acquisition model, trained from image-caption pairs on naturalistic data using cross-modal self-supervision. We show that the model learns word categories and object recognition abilities, and presents trends reminiscent of those reported in the developmental literature. We make our code and trained models public for future reference and use.

Aili Shen, Xudong Han, Trevor Cohn et al.

Real-world datasets often encode stereotypes and societal biases. Such biases can be implicitly captured by trained models, leading to biased predictions and exacerbating existing societal preconceptions. Existing debiasing methods, such as adversarial training and removing protected information from representations, have been shown to reduce bias. However, a disconnect between fairness criteria and training objectives makes it difficult to reason theoretically about the effectiveness of different techniques. In this work, we propose two novel training objectives which directly optimise for the widely-used criterion of {\it equal opportunity}, and show that they are effective in reducing bias while maintaining high performance over two classification tasks.

Uri Berger, Gabriel Stanovsky, Omri Abend et al.

The task of image captioning has recently been gaining popularity, and with it the complex task of evaluating the quality of image captioning models. In this work, we present the first survey and taxonomy of over 70 different image captioning metrics and their usage in hundreds of papers, specifically designed to help users select the most suitable metric for their needs. We find that despite the diversity of proposed metrics, the vast majority of studies rely on only five popular metrics, which we show to be weakly correlated with human ratings. We hypothesize that combining a diverse set of metrics can enhance correlation with human ratings. As an initial step, we demonstrate that a linear regression-based ensemble method, which we call EnsembEval, trained on one human ratings dataset, achieves improved correlation across five additional datasets, showing there is a lot of room for improvement by leveraging a diverse set of metrics.

Uri Berger, Lea Frermann, Gabriel Stanovsky et al.

We present a large, multilingual study into how vision constrains linguistic choice, covering four languages and five linguistic properties, such as verb transitivity or use of numerals. We propose a novel method that leverages existing corpora of images with captions written by native speakers, and apply it to nine corpora, comprising 600k images and 3M captions. We study the relation between visual input and linguistic choices by training classifiers to predict the probability of expressing a property from raw images, and find evidence supporting the claim that linguistic properties are constrained by visual context across languages. We complement this investigation with a corpus study, taking the test case of numerals. Specifically, we use existing annotations (number or type of objects) to investigate the effect of different visual conditions on the use of numeral expressions in captions, and show that similar patterns emerge across languages. Our methods and findings both confirm and extend existing research in the cognitive literature. We additionally discuss possible applications for language generation.

Jinrui Yang, Sheilla Njoto, Marc Cheong et al.

Gender discrimination in hiring is a pertinent and persistent bias in society, and a common motivating example for exploring bias in NLP. However, the manifestation of gendered language in application materials has received limited attention. This paper investigates the framing of skills and background in CVs of self-identified men and women. We introduce a data set of 1.8K authentic, English-language, CVs from the US, covering 16 occupations, allowing us to partially control for the confound occupation-specific gender base rates. We find that (1) women use more verbs evoking impressions of low power; and (2) classifiers capture gender signal even after data balancing and removal of pronouns and named entities, and this holds for both transformer-based and linear classifiers.

Lea Frermann, Jiatong Li, Shima Khanehzar et al.

Despite increasing interest in the automatic detection of media frames in NLP, the problem is typically simplified as single-label classification and adopts a topic-like view on frames, evading modelling the broader document-level narrative. In this work, we revisit a widely used conceptualization of framing from the communication sciences which explicitly captures elements of narratives, including conflict and its resolution, and integrate it with the narrative framing of key entities in the story as heroes, victims or villains. We adapt an effective annotation paradigm that breaks a complex annotation task into a series of simpler binary questions, and present an annotated data set of English news articles, and a case study on the framing of climate change in articles from news outlets across the political spectrum. Finally, we explore automatic multi-label prediction of our frames with supervised and semi-supervised approaches, and present a novel retrieval-based method which is both effective and transparent in its predictions. We conclude with a discussion of opportunities and challenges for future work on document-level models of narrative framing.

Xudong Han, Aili Shen, Trevor Cohn et al.

Mitigating bias in training on biased datasets is an important open problem. Several techniques have been proposed, however the typical evaluation regime is very limited, considering very narrow data conditions. For instance, the effect of target class imbalance and stereotyping is under-studied. To address this gap, we examine the performance of various debiasing methods across multiple tasks, spanning binary classification (Twitter sentiment), multi-class classification (profession prediction), and regression (valence prediction). Through extensive experimentation, we find that data conditions have a strong influence on relative model performance, and that general conclusions cannot be drawn about method efficacy when evaluating only on standard datasets, as is current practice in fairness research.

Gisela Vallejo, Timothy Baldwin, Lea Frermann

The manifestation and effect of bias in news reporting have been central topics in the social sciences for decades, and have received increasing attention in the NLP community recently. While NLP can help to scale up analyses or contribute automatic procedures to investigate the impact of biased news in society, we argue that methodologies that are currently dominant fall short of addressing the complex questions and effects addressed in theoretical media studies. In this survey paper, we review social science approaches and draw a comparison with typical task formulations, methods, and evaluation metrics used in the analysis of media bias in NLP. We discuss open questions and suggest possible directions to close identified gaps between theory and predictive models, and their evaluation. These include model transparency, considering document-external information, and cross-document reasoning rather than single-label assignment.

Kemal Kurniawan, Lea Frermann, Philip Schulz et al.

Cross-lingual transfer is a leading technique for parsing low-resource languages in the absence of explicit supervision. Simple `direct transfer' of a learned model based on a multilingual input encoding has provided a strong benchmark. This paper presents a method for unsupervised cross-lingual transfer that improves over direct transfer systems by using their output as implicit supervision as part of self-training on unlabelled text in the target language. The method assumes minimal resources and provides maximal flexibility by (a) accepting any pre-trained arc-factored dependency parser; (b) assuming no access to source language data; (c) supporting both projective and non-projective parsing; and (d) supporting multi-source transfer. With English as the source language, we show significant improvements over state-of-the-art transfer models on both distant and nearby languages, despite our conceptually simpler approach. We provide analyses of the choice of source languages for multi-source transfer, and the advantage of non-projective parsing. Our code is available online.

Yilin Geng, Haonan Li, Honglin Mu et al.

Large language models (LLMs) are increasingly deployed with hierarchical instruction schemes, where certain instructions (e.g., system-level directives) are expected to take precedence over others (e.g., user messages). Yet, we lack a systematic understanding of how effectively these hierarchical control mechanisms work. We introduce a systematic evaluation framework based on constraint prioritization to assess how well LLMs enforce instruction hierarchies. Our experiments across six state-of-the-art LLMs reveal that models struggle with consistent instruction prioritization, even for simple formatting conflicts. We find that the widely-adopted system/user prompt separation fails to establish a reliable instruction hierarchy, and models exhibit strong inherent biases toward certain constraint types regardless of their priority designation. We find that LLMs more reliably obey constraints framed through natural social hierarchies (e.g., authority, expertise, consensus) than system/user roles, which suggests that pretraining-derived social structures act as latent control priors, with potentially stronger influence than post-training guardrails.

Uri Berger, Omri Abend, Lea Frermann et al.

Incorporating automatically predicted human feedback into the process of training generative models has attracted substantial recent interest, while feedback at inference time has received less attention. The typical feedback at training time, i.e., preferences of choice given two samples, does not naturally transfer to the inference phase. We introduce a novel type of feedback -- caption reformulations -- and train models to mimic reformulation feedback based on human annotations. Our method does not require training the image captioning model itself, thereby demanding substantially less computational effort. We experiment with two types of reformulation feedback: first, we collect a dataset of human reformulations that correct errors in the generated captions. We find that incorporating reformulation models trained on this data into the inference phase of existing image captioning models results in improved captions, especially when the original captions are of low quality. We apply our method to non-English image captioning, a domain where robust models are less prevalent, and gain substantial improvement. Second, we apply reformulations to style transfer. Quantitative evaluations reveal state-of-the-art performance on German image captioning and English style transfer, while human validation with a detailed comparative framework exposes the specific axes of improvement.

William De Michele, Abel Armas Cervantes, Lea Frermann

Business processes are fundamental to organizational operations, yet their optimization remains challenging due to the timeconsuming nature of manual process analysis. Our paper harnesses Large Language Models (LLMs) to automate value-added analysis, a qualitative process analysis technique that aims to identify steps in the process that do not deliver value. To date, this technique is predominantly manual, time-consuming, and subjective. Our method offers a more principled approach which operates in two phases: first, decomposing high-level activities into detailed steps to enable granular analysis, and second, performing a value-added analysis to classify each step according to Lean principles. This approach enables systematic identification of waste while maintaining the semantic understanding necessary for qualitative analysis. We develop our approach using 50 business process models, for which we collect and publish manual ground-truth labels. Our evaluation, comparing zero-shot baselines with more structured prompts reveals (a) a consistent benefit of structured prompting and (b) promising performance for both tasks. We discuss the potential for LLMs to augment human expertise in qualitative process analysis while reducing the time and subjectivity inherent in manual approaches.

Rena Gao, Ming-Bin Chen, Lea Frermann et al.

English as a Second Language (ESL) speakers often struggle to engage in group discussions due to language barriers. While moderators can facilitate participation, few studies assess conversational engagement and evaluate moderation effectiveness. To address this gap, we develop a dataset comprising 17 sessions from an online ESL conversation club, which includes both moderated and non-moderated discussions. We then introduce an approach that integrates automatic ESL dialogue assessment and a framework that categorizes moderation strategies. Our findings indicate that moderators help improve the flow of topics and start/end a conversation. Interestingly, we find active acknowledgement and encouragement to be the most effective moderation strategy, while excessive information and opinion sharing by moderators has a negative impact. Ultimately, our study paves the way for analyzing ESL group discussions and the role of moderators in non-native conversation settings.

Aotao Xu, Charles Kemp, Lea Frermann et al.

A key function of the lexicon is to express novel concepts as they emerge over time through a process known as lexicalization. The most common lexicalization strategies are the reuse and combination of existing words, but they have typically been studied separately in the areas of word meaning extension and word formation. Here we offer an information-theoretic account of how both strategies are constrained by a fundamental tradeoff between competing communicative pressures: word reuse tends to preserve the average length of word forms at the cost of less precision, while word combination tends to produce more informative words at the expense of greater word length. We test our proposal against a large dataset of reuse items and compounds that appeared in English, French and Finnish over the past century. We find that these historically emerging items achieve higher levels of communicative efficiency than hypothetical ways of constructing the lexicon, and both literal reuse items and compounds tend to be more efficient than their non-literal counterparts. These results suggest that reuse and combination are both consistent with a unified account of lexicalization grounded in the theory of efficient communication.

Ming-Bin Chen, Lea Frermann, Jey Han Lau

We propose WHoW, an evaluation framework for analyzing the facilitation strategies of moderators across different domains/scenarios by examining their motives (Why), dialogue acts (How) and target speaker (Who). Using this framework, we annotated 5,657 moderation sentences with human judges and 15,494 sentences with GPT-4o from two domains: TV debates and radio panel discussions. Comparative analysis demonstrates the framework's cross-domain generalisability and reveals distinct moderation strategies: debate moderators emphasise coordination and facilitate interaction through questions and instructions, while panel discussion moderators prioritize information provision and actively participate in discussions. Our analytical framework works for different moderation scenarios, enhances our understanding of moderation behaviour through automatic large-scale analysis, and facilitates the development of moderator agents.

Hannah Bansal, Kemal Kurniawan, Lea Frermann

Recent trends in NLP aim to improve reasoning capabilities in Large Language Models (LLMs), with key focus on generalization and robustness to variations in tasks. Counterfactual task variants introduce minimal but semantically meaningful changes to otherwise valid first-order logic (FOL) problem instances altering a single predicate or swapping roles of constants to probe whether a reasoning system can maintain logical consistency under perturbation. Previous studies showed that LLMs becomes brittle on counterfactual variations, suggesting that they often rely on spurious surface patterns to generate responses. In this work, we explore if a neurosymbolic (NS) approach that integrates an LLM and a symbolic logical solver could mitigate this problem. Experiments across LLMs of varying sizes show that NS methods are more robust but perform worse overall that purely neural methods. We then propose NSCoT that combines an NS method and Chain-of-Thought (CoT) prompting and demonstrate that while it improves performance, NSCoT still lags behind standard CoT. Our analysis opens research directions for future work.

Kemal Kurniawan, Lea Frermann, Philip Schulz et al.

Providing technologies to communities or domains where training data is scarce or protected e.g., for privacy reasons, is becoming increasingly important. To that end, we generalise methods for unsupervised transfer from multiple input models for structured prediction. We show that the means of aggregating over the input models is critical, and that multiplying marginal probabilities of substructures to obtain high-probability structures for distant supervision is substantially better than taking the union of such structures over the input models, as done in prior work. Testing on 18 languages, we demonstrate that the method works in a cross-lingual setting, considering both dependency parsing and part-of-speech structured prediction problems. Our analyses show that the proposed method produces less noisy labels for the distant supervision.

Aili Shen, Xudong Han, Trevor Cohn et al.

Trained classification models can unintentionally lead to biased representations and predictions, which can reinforce societal preconceptions and stereotypes. Existing debiasing methods for classification models, such as adversarial training, are often expensive to train and difficult to optimise. In this paper, we propose a method for mitigating bias in classifier training by incorporating contrastive learning, in which instances sharing the same class label are encouraged to have similar representations, while instances sharing a protected attribute are forced further apart. In such a way our method learns representations which capture the task label in focused regions, while ensuring the protected attribute has diverse spread, and thus has limited impact on prediction and thereby results in fairer models. Extensive experimental results across four tasks in NLP and computer vision show (a) that our proposed method can achieve fairer representations and realises bias reductions compared with competitive baselines; and (b) that it can do so without sacrificing main task performance; (c) that it sets a new state-of-the-art performance in one task despite reducing the bias. Finally, our method is conceptually simple and agnostic to network architectures, and incurs minimal additional compute cost.

Shivashankar Subramanian, Afshin Rahimi, Timothy Baldwin et al.

Class imbalance is a common challenge in many NLP tasks, and has clear connections to bias, in that bias in training data often leads to higher accuracy for majority groups at the expense of minority groups. However there has traditionally been a disconnect between research on class-imbalanced learning and mitigating bias, and only recently have the two been looked at through a common lens. In this work we evaluate long-tail learning methods for tweet sentiment and occupation classification, and extend a margin-loss based approach with methods to enforce fairness. We empirically show through controlled experiments that the proposed approaches help mitigate both class imbalance and demographic biases.

Shivashankar Subramanian, Xudong Han, Timothy Baldwin et al.

Bias is pervasive in NLP models, motivating the development of automatic debiasing techniques. Evaluation of NLP debiasing methods has largely been limited to binary attributes in isolation, e.g., debiasing with respect to binary gender or race, however many corpora involve multiple such attributes, possibly with higher cardinality. In this paper we argue that a truly fair model must consider `gerrymandering' groups which comprise not only single attributes, but also intersectional groups. We evaluate a form of bias-constrained model which is new to NLP, as well an extension of the iterative nullspace projection technique which can handle multiple protected attributes.

Chunhua Liu, Trevor Cohn, Lea Frermann

Humans use countless basic, shared facts about the world to efficiently navigate in their environment. This commonsense knowledge is rarely communicated explicitly, however, understanding how commonsense knowledge is represented in different paradigms is important for both deeper understanding of human cognition and for augmenting automatic reasoning systems. This paper presents an in-depth comparison of two large-scale resources of general knowledge: ConcpetNet, an engineered relational database, and SWOW a knowledge graph derived from crowd-sourced word associations. We examine the structure, overlap and differences between the two graphs, as well as the extent to which they encode situational commonsense knowledge. We finally show empirically that both resources improve downstream task performance on commonsense reasoning benchmarks over text-only baselines, suggesting that large-scale word association data, which have been obtained for several languages through crowd-sourcing, can be a valuable complement to curated knowledge graphs

Shima Khanehzar, Trevor Cohn, Gosia Mikolajczak et al.

Understanding how news media frame political issues is important due to its impact on public attitudes, yet hard to automate. Computational approaches have largely focused on classifying the frame of a full news article while framing signals are often subtle and local. Furthermore, automatic news analysis is a sensitive domain, and existing classifiers lack transparency in their predictions. This paper addresses both issues with a novel semi-supervised model, which jointly learns to embed local information about the events and related actors in a news article through an auto-encoding framework, and to leverage this signal for document-level frame classification. Our experiments show that: our model outperforms previous models of frame prediction; we can further improve performance with unlabeled training data leveraging the semi-supervised nature of our model; and the learnt event and actor embeddings intuitively corroborate the document-level predictions, providing a nuanced and interpretable article frame representation.

Pinelopi Papalampidi, Frank Keller, Lea Frermann et al.

Most general-purpose extractive summarization models are trained on news articles, which are short and present all important information upfront. As a result, such models are biased on position and often perform a smart selection of sentences from the beginning of the document. When summarizing long narratives, which have complex structure and present information piecemeal, simple position heuristics are not sufficient. In this paper, we propose to explicitly incorporate the underlying structure of narratives into general unsupervised and supervised extractive summarization models. We formalize narrative structure in terms of key narrative events (turning points) and treat it as latent in order to summarize screenplays (i.e., extract an optimal sequence of scenes). Experimental results on the CSI corpus of TV screenplays, which we augment with scene-level summarization labels, show that latent turning points correlate with important aspects of a CSI episode and improve summarization performance over general extractive algorithms leading to more complete and diverse summaries.

Lahari Poddar, Gyorgy Szarvas, Lea Frermann

The meaning of a word often varies depending on its usage in different domains. The standard word embedding models struggle to represent this variation, as they learn a single global representation for a word. We propose a method to learn domain-specific word embeddings, from text organized into hierarchical domains, such as reviews in an e-commerce website, where products follow a taxonomy. Our structured probabilistic model allows vector representations for the same word to drift away from each other for distant domains in the taxonomy, to accommodate its domain-specific meanings. By learning sets of domain-specific word representations jointly, our model can leverage domain relationships, and it scales well with the number of domains. Using large real-world review datasets, we demonstrate the effectiveness of our model compared to state-of-the-art approaches, in learning domain-specific word embeddings that are both intuitive to humans and benefit downstream NLP tasks.

Stefanos Angelidis, Lea Frermann, Diego Marcheggiani et al.

We present a system for answering questions based on the full text of books (BookQA), which first selects book passages given a question at hand, and then uses a memory network to reason and predict an answer. To improve generalization, we pretrain our memory network using artificial questions generated from book sentences. We experiment with the recently published NarrativeQA corpus, on the subset of Who questions, which expect book characters as answers. We experimentally show that BERT-based retrieval and pretraining improve over baseline results significantly. At the same time, we confirm that NarrativeQA is a highly challenging data set, and that there is need for novel research in order to achieve high-precision BookQA results. We analyze some of the bottlenecks of the current approach, and we argue that more research is needed on text representation, retrieval of relevant passages, and reasoning, including commonsense knowledge.

Lea Frermann, Mirella Lapata

Categories such as animal or furniture are acquired at an early age and play an important role in processing, organizing, and communicating world knowledge. Categories exist across cultures: they allow to efficiently represent the complexity of the world, and members of a community strongly agree on their nature, revealing a shared mental representation. Models of category learning and representation, however, are typically tested on data from small-scale experiments involving small sets of concepts with artificially restricted features; and experiments predominantly involve participants of selected cultural and socio-economical groups (very often involving western native speakers of English such as U.S. college students) . This work investigates whether models of categorization generalize (a) to rich and noisy data approximating the environment humans live in; and (b) across languages and cultures. We present a Bayesian cognitive model designed to jointly learn categories and their structured representation from natural language text which allows us to (a) evaluate performance on a large scale, and (b) apply our model to a diverse set of languages. We show that meaningful categories comprising hundreds of concepts and richly structured featural representations emerge across languages. Our work illustrates the potential of recent advances in computational modeling and large scale naturalistic datasets for cognitive science research.

Lea Frermann, Shay B. Cohen, Mirella Lapata

In this paper we argue that crime drama exemplified in television programs such as CSI:Crime Scene Investigation is an ideal testbed for approximating real-world natural language understanding and the complex inferences associated with it. We propose to treat crime drama as a new inference task, capitalizing on the fact that each episode poses the same basic question (i.e., who committed the crime) and naturally provides the answer when the perpetrator is revealed. We develop a new dataset based on CSI episodes, formalize perpetrator identification as a sequence labeling problem, and develop an LSTM-based model which learns from multi-modal data. Experimental results show that an incremental inference strategy is key to making accurate guesses as well as learning from representations fusing textual, visual, and acoustic input.

Lea Frermann, Michael C. Frank

The impressive ability of children to acquire language is a widely studied phenomenon, and the factors influencing the pace and patterns of word learning remains a subject of active research. Although many models predicting the age of acquisition of words have been proposed, little emphasis has been directed to the raw input children achieve. In this work we present a comparatively large-scale multi-modal corpus of prosody-text aligned child directed speech. Our corpus contains automatically extracted word-level prosodic features, and we investigate the utility of this information as predictors of age of acquisition. We show that prosody features boost predictive power in a regularized regression, and demonstrate their utility in the context of a multi-modal factorized language models trained and tested on child-directed speech.