Despite the predominance of English in their training data, English-centric Large Language Models (LLMs) like GPT-3 and LLaMA display a remarkable ability to perform multilingual tasks, raising questions about the depth and nature of their cross-lingual capabilities. This paper introduces the decomposed prompting approach to probe the linguistic structure understanding of these LLMs in sequence labeling tasks. Diverging from the single text-to-text prompt, our method generates for each token of the input sentence an individual prompt which asks for its linguistic label. We assess our method on the Universal Dependencies part-of-speech tagging dataset for 38 languages, utilizing both English-centric and multilingual LLMs. Our findings show that decomposed prompting surpasses the iterative prompting baseline in efficacy and efficiency under zero- and few-shot settings. Further analysis reveals the influence of evaluation methods and the use of instructions in prompts. Our multilingual investigation shows that English-centric language models perform better on average than multilingual models. Our study offers insights into the multilingual transferability of English-centric LLMs, contributing to the understanding of their multilingual linguistic knowledge.
The open-ended nature of language generation makes the evaluation of autoregressive large language models (LLMs) challenging. One common evaluation approach uses multiple-choice questions (MCQ) to limit the response space. The model is then evaluated by ranking the candidate answers by the log probability of the first token prediction. However, first-tokens may not consistently reflect the final response output, due to model's diverse response styles such as starting with "Sure" or refusing to answer. Consequently, MCQ evaluation is not indicative of model behaviour when interacting with users. But by how much? We evaluate how aligned first-token evaluation is with the text output along several dimensions, namely final option choice, refusal rate, choice distribution and robustness under prompt perturbation. Our results show that the two approaches are severely misaligned on all dimensions, reaching mismatch rates over 60%. Models heavily fine-tuned on conversational or safety data are especially impacted. Crucially, models remain misaligned even when we increasingly constrain prompts, i.e., force them to start with an option letter or example template. Our findings i) underscore the importance of inspecting the text output, too and ii) caution against relying solely on first-token evaluation.
Prompt-based methods have been successfully applied to multilingual pretrained language models for zero-shot cross-lingual understanding. However, most previous studies primarily focused on sentence-level classification tasks, and only a few considered token-level labeling tasks such as Named Entity Recognition (NER) and Part-of-Speech (POS) tagging. In this paper, we propose Token-Level Prompt Decomposition (ToPro), which facilitates the prompt-based method for token-level sequence labeling tasks. The ToPro method decomposes an input sentence into single tokens and applies one prompt template to each token. Our experiments on multilingual NER and POS tagging datasets demonstrate that ToPro-based fine-tuning outperforms Vanilla fine-tuning and Prompt-Tuning in zero-shot cross-lingual transfer, especially for languages that are typologically different from the source language English. Our method also attains state-of-the-art performance when employed with the mT5 model. Besides, our exploratory study in multilingual large language models shows that ToPro performs much better than the current in-context learning method. Overall, the performance improvements show that ToPro could potentially serve as a novel and simple benchmarking method for sequence labeling tasks.
When training data are collected from human annotators, the design of the annotation instrument, the instructions given to annotators, the characteristics of the annotators, and their interactions can impact training data. This study demonstrates that design choices made when creating an annotation instrument also impact the models trained on the resulting annotations. We introduce the term annotation sensitivity to refer to the impact of annotation data collection methods on the annotations themselves and on downstream model performance and predictions. We collect annotations of hate speech and offensive language in five experimental conditions of an annotation instrument, randomly assigning annotators to conditions. We then fine-tune BERT models on each of the five resulting datasets and evaluate model performance on a holdout portion of each condition. We find considerable differences between the conditions for 1) the share of hate speech/offensive language annotations, 2) model performance, 3) model predictions, and 4) model learning curves. Our results emphasize the crucial role played by the annotation instrument which has received little attention in the machine learning literature. We call for additional research into how and why the instrument impacts the annotations to inform the development of best practices in instrument design.
Machine Learning (ML) systems are becoming instrumental in the public sector, with applications spanning areas like criminal justice, social welfare, financial fraud detection, and public health. While these systems offer great potential benefits to institutional decision-making processes, such as improved efficiency and reliability, they still face the challenge of aligning intricate and nuanced policy objectives with the precise formalization requirements necessitated by ML models. In this paper, we aim to bridge the gap between ML and public sector decision-making by presenting a comprehensive overview of key technical challenges where disjunctions between policy goals and ML models commonly arise. We concentrate on pivotal points of the ML pipeline that connect the model to its operational environment, delving into the significance of representative training data and highlighting the importance of a model setup that facilitates effective decision-making. Additionally, we link these challenges with emerging methodological advancements, encompassing causal ML, domain adaptation, uncertainty quantification, and multi-objective optimization, illustrating the path forward for harmonizing ML and public sector objectives.
Although the NLP community has adopted central differential privacy as a go-to framework for privacy-preserving model training or data sharing, the choice and interpretation of the key parameter, privacy budget $\varepsilon$ that governs the strength of privacy protection, remains largely arbitrary. We argue that determining the $\varepsilon$ value should not be solely in the hands of researchers or system developers, but must also take into account the actual people who share their potentially sensitive data. In other words: Would you share your instant messages for $\varepsilon$ of 10? We address this research gap by designing, implementing, and conducting a behavioral experiment (311 lay participants) to study the behavior of people in uncertain decision-making situations with respect to privacy-threatening situations. Framing the risk perception in terms of two realistic NLP scenarios and using a vignette behavioral study help us determine what $\varepsilon$ thresholds would lead lay people to be willing to share sensitive textual data - to our knowledge, the first study of its kind.
Machine Learning and Deep Learning have achieved an impressive standard today, enabling us to answer questions that were inconceivable a few years ago. Besides these successes, it becomes clear, that beyond pure prediction, which is the primary strength of most supervised machine learning algorithms, the quantification of uncertainty is relevant and necessary as well. While first concepts and ideas in this direction have emerged in recent years, this paper adopts a conceptual perspective and examines possible sources of uncertainty. By adopting the viewpoint of a statistician, we discuss the concepts of aleatoric and epistemic uncertainty, which are more commonly associated with machine learning. The paper aims to formalize the two types of uncertainty and demonstrates that sources of uncertainty are miscellaneous and can not always be decomposed into aleatoric and epistemic. Drawing parallels between statistical concepts and uncertainty in machine learning, we also demonstrate the role of data and their influence on uncertainty.
Advanced large language models like ChatGPT have gained considerable attention recently, including among students. However, while the debate on ChatGPT in academia is making waves, more understanding is needed among lecturers and teachers on how students use and perceive ChatGPT. To address this gap, we analyzed the content on ChatGPT available on TikTok in February 2023. TikTok is a rapidly growing social media platform popular among individuals under 30. Specifically, we analyzed the content of the 100 most popular videos in English tagged with #chatgpt, which collectively garnered over 250 million views. Most of the videos we studied promoted the use of ChatGPT for tasks like writing essays or code. In addition, many videos discussed AI detectors, with a focus on how other tools can help to transform ChatGPT output to fool these detectors. This also mirrors the discussion among educators on how to treat ChatGPT as lecturers and teachers in teaching and grading. What is, however, missing from the analyzed clips on TikTok are videos that discuss ChatGPT producing content that is nonsensical or unfaithful to the training data.
We propose new ensemble models for multivariate functional data classification as combinations of semi-metric-based weak learners. Our models extend current semi-metric-type methods from the univariate to the multivariate case, propose new semi-metrics to compute distances between functions, and consider more flexible options for combining weak learners using stacked generalisation methods. We apply these ensemble models to identify respondents' difficulty with survey questions, with the aim to improve survey data quality. As predictors of difficulty, we use mouse movement trajectories from the respondents' interaction with a web survey, in which several questions were manipulated to create two scenarios with different levels of difficulty.
Algorithmic profiling is increasingly used in the public sector as a means to allocate limited public resources effectively and objectively. One example is the prediction-based statistical profiling of job seekers to guide the allocation of support measures by public employment services. However, empirical evaluations of potential side-effects such as unintended discrimination and fairness concerns are rare. In this study, we compare and evaluate statistical models for predicting job seekers' risk of becoming long-term unemployed with respect to prediction performance, fairness metrics, and vulnerabilities to data analysis decisions. Focusing on Germany as a use case, we evaluate profiling models under realistic conditions by utilizing administrative data on job seekers' employment histories that are routinely collected by German public employment services. Besides showing that these data can be used to predict long-term unemployment with competitive levels of accuracy, we highlight that different classification policies have very different fairness implications. We therefore call for rigorous auditing processes before such models are put to practice.