Croatia

We report our efforts in identifying a set of previous human evaluations in NLP that would be suitable for a coordinated study examining what makes human evaluations in NLP more/less reproducible. We present our results and findings, which include that just 13% of papers had (i) sufficiently low barriers to reproduction, and (ii) enough obtainable information, to be considered for reproduction and that all but one of the experiments we selected for reproduction was discovered to have flaws that made the meaningfulness of conducting a reproduction questionable. As a result, we had to change our coordinated study design from a reproduce approach to a standardise-then-reproduce-twice approach. Our overall (negative) finding that the great majority of human evaluations in NLP is not repeatable and/or not reproducible and/or too flawed to justify reproduction, paints a dire picture, but presents an opportunity for a rethink about how to design and report human evaluations in NLP.

EACL 2023

Missing Information, Unresponsive Authors, Experimental Flaws: The Impossibility of Assessing the Reproducibility of Previous Human

experimental flaws

reproducibility

Truth and Responsibility
“Truth and Responsibility” is a call to reimagine anthropology to meet the demands of the present moment. The imperative to bear witness, take action, and be held accountable to the truths we write and circulate invites us to reflect on our responsibility in reckoning with disciplinary histories, harms, and possibilities. To whom are we giving evidence and toward what ends? For whom are we writing? To whom are we accountable, and in what ways? We are thrilled to announce the theme of the 2021 AAA Annual Meeting to be held in Baltimore, MD: Truth and Responsibility.

Many consider anthropology to be a holistic social science dedicated to cross-cultural understanding as a way to reduce bigotry and more fully comprehend “what it means to be human.” Others push against this readily circulated origin story, arguing that “human” remains a highly contested, political category—evident as the presumed set of rights and protections associated with the “human” continue to be stripped away violently in the communities in which some of us live and work. Which violent histories and epistemologies must anthropologists take responsibility for in telling truth(s) about the discipline’s development? How do the dual histories of settler colonialism and slavery continue to influence anthropological thought and practices? In these political times, how can anthropologists throughout the globe work to secure a capacious, progressive vision of the human? How might movements for anti-racism, decoloniality, queer liberation, and healing such as #BlackLivesMatter, #ProtectMaunaKea, and #MeToo push a future anthropology out of the ashes of the anthropological past? Finally, what are the limits and possibilities of the anthropological imagination?

Across subfields, we find truths in patterns of human behavior, language, evolution, and cultural worlds. Industry-positioned anthropologists do this by extending the boundaries of the discipline not as an administrative solution to job shortages but as a source of intervention and knowledge production. Increasingly, anthropologists who make use of the scientific method question how status and identity help to decide which facts come to represent truth. How might quantitative researchers use our methods to identify patterned truths, and what responsibility do we hold in challenging conventional wisdom about these patterns? What responsibility do we have to both our colleagues and larger publics, to push toward increased transparency and ethical grounding in our data collection, analysis, and presentation?

Pedagogy is a key medium for the communication of anthropological truths. Which (and whose) truths are foregrounded in our curricula? What are the possibilities for a liberatory pedagogy in a precarious, “post-truth” era? Conferences, museums, organizing spaces, and classrooms are sites where pedagogy and various forms of communication can help us interrogate the performativity of truth. How might anthropologists conceive of truth performances? Might visual, aural, and choreographic performances during AAA sessions offer anthropologists alternative genres for truth-telling and truth-making that provoke discussions about responsibility?

In accordance with the urgency of the year’s theme, the questions posed in this call for papers do not have easy answers. The hope is that those who submit proposals will take up these questions and engage in collective thinking and imagining about the truths we hold, the truths we challenge, and the responsibilities we bear in co-creating a free and more equitable world. All panel proposals should have a statement about how the panel has incorporated the goals of equity, diversity, and inclusion, and/or analysis of power.

Executive Session proposals have already been selected and speak directly to the theme “Truth and Responsibility.” To practice the critical reflexivity this call for papers invites, we asked that Executive Session submissions explicitly address the diversity of session participants, such as organizational/institutional affiliation, career stage, race, ethnicity, disability, sexuality, class, religion, national identity, and gender. Proposals for Executive Sessions that demonstrated a commitment to equity and inclusion and/or a clear analysis of power (i.e. anti-imperialist, anti-white supremacist, decolonial, anti-ableist, feminist, anti-transphobic, anti-xenophobic, etc) were given priority. We are excited to be in conversation with residents of the Baltimore region and look forward to the discussions, performances, and actions generated by the dynamic group of anthropologists attending the 2021 AAA Annual Meeting.

AAA Annual Meeting 2021

“Truth and Responsibility” is a call to reimagine anthropology to meet the demands of the present moment. The imperative to bear witness, take action, and be held accountable to the truths we write and circulate invites us to reflect on our responsibility in reckoning with disciplinary histories, harms, and possibilities. To whom are we giving evidence and toward what ends? For whom are we writing? To whom are we accountable, and in what ways? We are thrilled to announce the theme of the 2021 AAA Annual Meeting to be held in Baltimore, MD: Truth and Responsibility.

workshop paper

### 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.

Existing language and vision models achieve impressive performance in image-text understanding. Yet, it is an open question to what extent they can be used for language understanding in 3D environments and whether they implicitly acquire 3D object knowledge, e.g. about different views of an object.
In this paper, we investigate whether a state-of-the-art language and vision model, CLIP, is able to ground perspective descriptions of a 3D object and identify canonical views of common objects based on text queries.
We present an evaluation framework that uses a circling camera around a 3D object to generate images from different viewpoints and evaluate them in terms of their similarity to natural language descriptions.
We find that a pre-trained CLIP model performs poorly on most canonical views and that fine-tuning using hard negative sampling and random contrasting yields good results even under conditions with little available training data.

Paparazzi: A Deep Dive into the Capabilities of Language and Vision Models for Grounding Viewpoint Descriptions

Dialect Identification is a crucial task for localizing various Large Language Models. This paper outlines our approach to the VarDial 2023 DSL-TL shared task. Here we have to identify three or two dialects from three languages each which results in a 9-way classification for Track-1 and 6-way classification for Track-2 respectively. Our proposed approach consists of a two-stage system and outperforms other participants’ systems and previous works in this domain. We achieve a score of 58.54% for Track-1 and 85.61% for Track-2. Our codebase is available publicly.

Two-stage Pipeline for Multilingual Dialect Detection

In this work, we induce character-level noise in various forms when fine-tuning BERT to enable zero-shot cross-lingual transfer to unseen dialects and languages. We fine-tune BERT on three sentence-level classification tasks and evaluate our approach on an assortment of unseen dialects and languages. We find that character-level noise can be an extremely effective agent of cross-lingual transfer under certain conditions, while it is not as helpful in others. Specifically, we explore these differences in terms of the nature of the task and the relationships between source and target languages, finding that introduction of character-level noise during fine-tuning is particularly helpful when a task draws on surface level cues and the source-target cross-lingual pair has a relatively high lexical overlap with shorter (i.e., less meaningful) unseen tokens on average.

Fine-Tuning BERT with Character-Level Noise for Zero-Shot Transfer to Dialects and Closely-Related Languages

Hate speech detection in online platforms has been widely studied in the past. Most of these works were conducted in English and a few rich-resource languages. Recent approaches tailored for low-resource languages have explored the interests of zero shot cross-lingual transfer learning models in resource-scarce scenarios. However, languages variations between geolects such as American English and British English, Latin-American Spanish, and European Spanish is still a problem for NLP models that often relies on (latent) lexical information for their classification tasks. More importantly, the cultural aspect, crucial for hate speech detection, is often overlooked.

In this work, we present the results of a thorough analysis of hate speech detection models performance on different variants of Spanish, including a new hate speech toward immigrants Twitter data set we built to cover these variants. Using mBERT and Beto, a monolingual Spanish Bert-based language model, as the basis of our transfer learning architecture, our results indicate that hate speech detection models for a given Spanish variant are affected when different variations of such language are not considered. Hate speech expressions could vary from region to region where the same language is spoken.

Analyzing Zero-Shot transfer Scenarios across Spanish variants for Hate Speech Detection

Pre-trained models usually come with a pre-defined tokenization and little flexibility as to what subword tokens can be used in downstream tasks. This problem concerns especially multilingual NLP and low-resource languages, which are typically processed using cross-lingual transfer. In this paper, we aim to find out if the right granularity of tokenization is helpful for a text classification task, namely dialect classification. Aiming at generalizations beyond the studied cases, we look for the optimal granularity in four dialect datasets, two with relatively consistent writing (one Arabic and one Indo-Aryan set) and two with con- siderably inconsistent writing (one Arabic and one Swiss German set). To gain more control over subword tokenization and ensure direct comparability in the experimental settings, we train a CNN classifier from scratch comparing two subword tokenization methods (Unigram model and BPE). For reference, we compare the results obtained in our analysis to the state of the art achieved by fine-tuning pre-trained models. We show that models trained from scratch with an optimal tokenization level perform better than fine-tuned classifiers in the case of highly inconsistent writing. In the case of relatively consistent writing, fine-tuned mod- els remain better regardless of the tokenization level.

Optimizing the Size of Subword Vocabularies in Dialect Classification

Eye-tracking data in Chinese languages present unique challenges due to the non-alphabetic and unspaced nature of the Chinese writing systems. This paper introduces the first deeply-annotated joint Mandarin-Cantonese eye-tracking dataset, from which we achieve a unified eye-tracking prediction system for both language varieties. In addition to the commonly studied first fixation duration and the total fixation duration, this dataset also includes the second fixation duration, expressing fixation patterns that are more relevant to higher-level, structural processing. 

A basic comparison of the features and measurements in our dataset revealed variation between Mandarin and Cantonese on fixation patterns related to word class and word position. The test of feature usefulness suggested that traditional features are less powerful in predicting the second-pass fixation, to which the linear distance to root makes a leading contribution in Mandarin. In contrast, Cantonese eye-movement behavior relies more on word position and part of speech.

Comparing and Predicting Eye-tracking Data of Mandarin and Cantonese

One of the challenges with finetuning pretrained language models (PLMs) is that their tokenizer is optimized for the language(s) it was pretrained on, but brittle when it comes to previously unseen variations in the data. This can for instance be observed when finetuning PLMs on one language and evaluating them on data in a closely related language variety with no standardized orthography. Despite the high linguistic similarity, tokenization no longer corresponds to meaningful representations of the target data, leading to low performance in, e.g., part-of-speech tagging. 

In this work, we finetune PLMs on seven languages from three different families and analyze their zero-shot performance on closely related, non-standardized varieties. 
We consider different measures for the divergence in the tokenization of the source and target data, and the way they can be adjusted by manipulating the tokenization during the finetuning step. Overall, we find that the similarity between the percentage of words that get split into subwords in the source and target data (the split word ratio difference) is the strongest predictor for model performance on target data.

Does Manipulating Tokenization Aid Cross-Lingual Transfer? A Study on POS Tag- ging for Non-Standardized Languages

This paper presents Murreviikko, a dataset of dialectal Finnish tweets which have been dialectologically annotated and manually normalized to a standard form. The dataset can be used as a test set for dialect identification and dialect-to-standard normalization, for instance. 
We evaluate the dataset on the normalization task, comparing an existing normalization model built on a spoken dialect corpus and three newly trained models with different architectures. We find that there are significant differences in normalization difficulty between the dialects, and that a character-level statistical machine translation model performs best on the Murreviikko tweet dataset.

Murreviikko - A Dialectologically Annotated and Normalized Dataset of Finnish Tweets

Multilingual pretraining approaches in Neural Machine Translation (NMT) have shown that training models to denoise synthetic code-switched data can yield impressive performance gains — owing to better multilingual semantic representations and transfer learning. However, they generated the synthetic code-switched data using non-contextual, one-to-one word translations obtained from lexicons - which can lead to significant noise in a variety of cases, including the poor handling of polysemes and multi-word expressions, violation of linguistic agreement and inability to scale to agglutinative languages. To overcome these limitations, we propose an approach called Contextual Code-Switching (CCS), where contextual, many-to-many word translations are generated using a `base' NMT model. We conduct experiments on 3 different language families - Romance, Uralic, and Indo-Aryan - and show significant improvements (by up to 5.5 spBLEU points) over the previous lexicon-based SOTA approaches. We also observe that small CCS models can perform comparably or better than massive models like mBART50 and mRASP2, depending on the size of data provided. Lastly, through ablation studies, we highlight the major code-switching aspects (including context, many-to-many substitutions, code-switching language count etc.) that contribute to the enhanced pretraining of multilingual NMT models.

Exploring Enhanced Code-Switched Noising for Pretraining in Neural Machine Translation

The digitisation of historical texts has provided new horizons for NLP research, but such data also presents a set of challenges, including scarcity and inconsistency. The lack of editorial standard during digitisation exacerbates these difficulties.

This study explores the potential for temporal domain adaptation in Early Modern Irish and pre-reform Modern Irish data. We describe two experiments carried out on the book subcorpus of the Historical Irish Corpus, which includes Early Modern Irish and pre-reform Modern Irish texts from 1581 to 1926. We also propose a simple orthographic normalisation method for historical Irish that reduces the type-token ratio by 21.43% on average in our data.

The results demonstrate that the use of out-of-domain data significantly improves a language model's performance. Providing a model with additional input from another historical stage of the language improves its quality by 12.49% on average on non-normalised texts and by 27.02% on average on normalised (demutated) texts. Most notably, using only out-of-domain data for both pre-training and training stages allowed for up to 86.81% of the baseline model quality on non-normalised texts and up to 95.68% on normalised texts without any target domain data. 

Additionally, we investigate the effect of temporal distance between the training and test data. The hypothesis that there is a positive correlation between performance and temporal proximity of training and test data has been validated, which manifests best in normalised data. Expanding this approach even further back, to Middle and Old Irish, and testing it on other languages is a further research direction.

Downloads

Next from EACL 2023

Paparazzi: A Deep Dive into the Capabilities of Language and Vision Models for Grounding Viewpoint Descriptions

Similar lecture

Social Reproduction: Theory and Ethnography