Patient Information Leaflets (PILs) provide essential information about medication usage, side effects, precautions, and interactions, making them a valuable resource for Question Answering (QA) systems in healthcare. However, no dedicated benchmark currently exists to evaluate QA systems specifically on PILs, limiting progress in this domain. To address this gap, we introduce a fact-supported synthetic benchmark composed of multiple-choice questions and answers generated from real PILs. We construct the benchmark using a fully automated pipeline that leverages multiple Large Language Models (LLMs) to generate diverse, realistic, and contextually relevant question-answer pairs. The benchmark is publicly released as a standardized evaluation framework for assessing the ability of LLMs to process and reason over PIL content. To validate its effectiveness, we conduct an initial evaluation with state-of-the-art LLMs, showing that the benchmark presents a realistic and challenging task, making it a valuable resource for advancing QA research in the healthcare domain.

Evaluating the truthfulness of online content is critical for combating misinformation. This study examines the efficiency and effectiveness of crowdsourced truthfulness assessments through a comparative analysis of two approaches: one involving full-length webpages as evidence for each claim, and another using summaries for each evidence document generated with an LLM. Using an A/B testing setting, we engage a diverse pool of participants tasked with evaluating the truthfulness of statements under these conditions.

Assessing the truthfulness of information is a critical task in fact-checking, and is typically performed using binary or coarse ordinal scales (2-6 levels), though fine-grained scales (e.g., 100 levels) have also been explored. Magnitude Estimation (ME) takes this approach further by allowing assessors to assign any value in the range (0, + ∞). However, it introduces challenges, including the need for aggregation of assessments from individuals with different interpretations of the scale. Despite these, its successful applications in other domains suggest its potential suitability for truthfulness assessment. We conduct a crowdsourcing study by collecting assessments on claims sourced from the PolitiFact fact-checking organization using ME. To the best of our knowledge, this is the first systematic investigation of ME in the context of truthfulness assessment. Our results show that while aggregation methods significantly impact assessment quality, optimal aggregation strategies yield accuracy and reliability comparable to traditional scales. More importantly, ME allows capturing subtle differences in truthfulness, offering richer insights than conventional coarse-grained scales.

This study challenges the prevailing approach of measuring political leanings in Large Language Models (LLMs) through direct questioning. By extensively testing LLMs with original, positively and negatively paraphrased Political Compass questions we demonstrate that LLMs do not consistently reveal their political biases in response to standard questions. Our findings indicate that LLMs' political orientations are elusive, easily influenced by subtle changes in phrasing and context. This study underscores the limitations of direct questioning in accurately measuring the political biases of LLMs and emphasizes the necessity for more refined and effective approaches to understand their true political stances.

We leverage generative NLP-based models, specifically Transformer-Based models, for multi-horizon univariate and multivariate power consumption forecasting. We apply our approach to various datasets, focusing on short-term (1 day) and long-term (1 week) forecasts. We test several lag configurations with and without additional contextual information and achieve promising results. We evaluate the forecasts' effectiveness using a range of metrics, and aggregate the results on a monthly basis for a comprehensive understanding of the performance throughout the year.

Research on misinformation detection has primarily focused either on furthering Artificial Intelligence (AI) for automated detection or on studying humans' ability to deliver an effective crowdsourced solution. Each of these directions however shows different benefits. This motivates our work to study hybrid human-AI approaches jointly leveraging the potential of large language models and crowdsourcing, which is understudied to date. We propose novel combination strategies Model First, Worker First, and Meta Vote, which we evaluate along with baseline methods such as mean, median, hard- and soft-voting. Using 120 statements from the PolitiFact dataset, and a combination of state-of-the-art AI models and crowdsourced assessments, we evaluate the effectiveness of these combination strategies. Results suggest that the effectiveness varies with scales granularity, and that combining AI and human judgments enhances truthfulness assessments' effectiveness and robustness.

There is an important ongoing effort aimed to tackle misinformation and to perform reliable fact-checking by employing human assessors at scale, with a crowdsourcing-based approach. Previous studies on the feasibility of employing crowdsourcing for the task of misinformation detection have provided inconsistent results: some of them seem to confirm the effectiveness of crowdsourcing for assessing the truthfulness of statements and claims, whereas others fail to reach an effectiveness level higher than automatic machine learning approaches, which are still unsatisfactory. In this paper, we aim at addressing such inconsistency and understand if truthfulness assessment can indeed be crowdsourced effectively. To do so, we build on top of previous studies; we select some of those reporting low effectiveness levels, we highlight their potential limitations, and we then reproduce their work attempting to improve their setup to address those limitations. We employ various approaches, data quality levels, and agreement measures to assess the reliability of crowd workers when assessing the truthfulness of (mis)information. Furthermore, we explore different worker features and compare the results obtained with different crowds. According to our findings, crowdsourcing can be used as an effective methodology to tackle misinformation at scale. When compared to previous studies, our results indicate that a significantly higher agreement between crowd workers and experts can be obtained by using a different, higher-quality, crowdsourcing platform and by improving the design of the crowdsourcing task. Also, we find differences concerning task and worker features and how workers provide truthfulness assessments.

The increase of the amount of misinformation spread every day online is a huge threat to the society. Organizations and researchers are working to contrast this misinformation plague. In this setting, human assessors are indispensable to correctly identify, assess and/or revise the truthfulness of information items, i.e., to perform the fact-checking activity. Assessors, as humans, are subject to systematic errors that might interfere with their fact-checking activity. Among such errors, cognitive biases are those due to the limits of human cognition. Although biases help to minimize the cost of making mistakes, they skew assessments away from an objective perception of information. Cognitive biases, hence, are particularly frequent and critical, and can cause errors that have a huge potential impact as they propagate not only in the community, but also in the datasets used to train automatic and semi-automatic machine learning models to fight misinformation. In this work, we present a review of the cognitive biases which might occur during the fact-checking process. In more detail, inspired by PRISMA – a methodology used for systematic literature reviews – we manually derive a list of 221 cognitive biases that may affect human assessors. Then, we select the 39 biases that might manifest during the fact-checking process, we group them into categories, and we provide a description. Finally, we present a list of 11 countermeasures that can be adopted by researchers, practitioners, and organizations to limit the effect of the identified cognitive biases on the fact-checking activity.

To scale the size of Information Retrieval collections, crowdsourcing has become a common way to collect relevance judgments at scale. Crowdsourcing experiments usually employ 100–10,000 workers, but such a number is often decided in a heuristic way. The downside is that the resulting dataset does not have any guarantee of meeting predefined statistical requirements as, for example, have enough statistical power to be able to distinguish in a statistically significant way between the relevance of two documents. We propose a methodology adapted from literature on sound topic set size design, based on t-test and ANOVA, which aims at guaranteeing the resulting dataset to meet a predefined set of statistical requirements. We validate our approach on several public datasets. Our results show that we can reliably estimate the recommended number of workers needed to achieve statistical power, and that such estimation is dependent on the topic, while the effect of the relevance scale is limited. Furthermore, we found that such estimation is dependent on worker features such as agreement. Finally, we describe a set of practical estimation strategies that can be used to estimate the worker set size, and we also provide results on the estimation of document set sizes.

Due to their relatively low cost and ability to scale, crowdsourcing based approaches are widely used to collect a large amount of human annotated data. To this aim, multiple crowdsourcing platforms exist, where requesters can upload tasks and workers can carry them out and obtain payment in return. Such platforms share a task design and deploy workflow that is often counter-intuitive and cumbersome. To address this issue, we propose Crowd\_Frame, a simple and complete framework which allows to develop and deploy diverse types of complex crowdsourcing tasks in an easy and customizable way. We show the abilities of the proposed framework and we make it available to researchers and practitioners.

In Information Retrieval (IR) evaluation, preference judgments are collected by presenting to the assessors a pair of documents and asking them to select which of the two, if any, is the most relevant. This is an alternative to the classic relevance judgment approach, in which human assessors judge the relevance of a single document on a scale; such an alternative allows to make relative rather than absolute judgments of relevance. While preference judgments are easier for human assessors to perform, the number of possible document pairs to be judged is usually so high that it makes it unfeasible to judge them all. Thus, following a similar idea to pooling strategies for single document relevance judgments where the goal is to sample the most useful documents to be judged, in this work we focus on analyzing alternative ways to sample document pairs to judge, in order to maximize the value of a fixed number of preference judgments that can feasibly be collected. Such value is defined as how well we can evaluate IR systems given a budget, that is, a fixed number of human preference judgments that may be collected. By relying on several datasets featuring relevance judgments gathered by means of experts and crowdsourcing, we experimentally compare alternative strategies to select document pairs and show how different strategies lead to different IR evaluation result quality levels. Our results show that, by using the appropriate procedure, it is possible to achieve good IR evaluation results with a limited number of preference judgments, thus confirming the feasibility of using preference judgments to create IR evaluation collections.

Crowdsourcing has become a standard methodology to collect manually annotated data such as relevance judgments at scale. On crowdsourcing platforms like Amazon MTurk or FigureEight, crowd workers select tasks to work on based on different dimensions such as task reward and requester reputation. Requesters then receive the judgments of workers who self-selected into the tasks and completed them successfully. Several crowd workers, however, preview tasks, begin working on them, reaching varying stages of task completion without finally submitting their work. Such behavior results in unrewarded effort which remains invisible to requesters. In this paper, we conduct an investigation of the phenomenon of task abandonment, the act of workers previewing or beginning a task and deciding not to complete it. We follow a three-fold methodology which includes 1) investigating the prevalence and causes of task abandonment by means of a survey over different crowdsourcing platforms, 2) data-driven analysis of logs collected during a large-scale relevance judgment experiment, and 3) controlled experiments measuring the effect of different dimensions on abandonment. Our results show that task abandonment is a widely spread phenomenon. Apart from accounting for a considerable amount of wasted human effort, this bears important implications on the hourly wages of workers as they are not rewarded for tasks that they do not complete. We also show how task abandonment may have strong implications on the use of collected data (for example, on the evaluation of Information Retrieval systems).

Misinformation is an ever increasing problem that is difficult to solve for the research community and has a negative impact on the society at large. Very recently, the problem has been addressed with a crowdsourcing-based approach to scale up labeling efforts: to assess the truthfulness of a statement, instead of relying on a few experts, a crowd of (non-expert) judges is exploited. We follow the same approach to study whether crowdsourcing is an effective and reliable method to assess statements truthfulness during a pandemic. We specifically target statements related to the COVID-19 health emergency, that is still ongoing at the time of the study and has arguably caused an increase of the amount of misinformation that is spreading online (a phenomenon for which the term "infodemic" has been used). By doing so, we are able to address (mis)information that is both related to a sensitive and personal issue like health and very recent as compared to when the judgment is done: two issues that have not been analyzed in related work.\n\nIn our experiment, crowd workers are asked to assess the truthfulness of statements, as well as to provide evidence for the assessments as a URL and a text justification. Besides showing that the crowd is able to accurately judge the truthfulness of the statements, we also report results on many different aspects, including: agreement among workers, the effect of different aggregation functions, of scales transformations, and of workers background / bias. We also analyze workers behavior, in terms of queries submitted, URLs found / selected, text justifications, and other behavioral data like clicks and mouse actions collected by means of an ad hoc logger.

Truthfulness judgments are a fundamental step in the process of fighting misinformation, as they are crucial to train and evaluate classifiers that automatically distinguish true and false statements. Usually such judgments are made by experts, like journalists for political statements or medical doctors for medical statements. In this paper, we follow a different approach and rely on (non-expert) crowd workers. This of course leads to the following research question: Can crowdsourcing be reliably used to assess the truthfulness of information and to create large-scale labeled collections for information credibility systems? To address this issue, we present the results of an extensive study based on crowdsourcing: we collect thousands of truthfulness assessments over two datasets, and we compare expert judgments with crowd judgments, expressed on scales with various granularity levels. We also measure the political bias and the cognitive background of the workers, and quantify their effect on the reliability of the data provided by the crowd.

Crowdsourcing is a popular technique to collect large amounts of human-generated labels, such as relevance judgments used to create information retrieval (IR) evaluation collections. Previous research has shown how collecting high quality labels from a crowdsourcing platform can be challenging. Existing quality assurance techniques focus on answer aggregation or on the use of gold questions where ground-truth data allows to check for the quality of the responses.

Information Retrieval (IR) researchers have often used existing IR evaluation collections and transformed the relevance scale in which judgments have been collected, e.g., to use metrics that assume binary judgments like Mean Average Precision. Such scale transformations are often arbitrary (e.g., 0,1 mapped to 0 and 2,3 mapped to 1) and it is assumed that they have no impact on the results of IR evaluation. Moreover, the use of crowdsourcing to collect relevance judgments has become a standard methodology. When designing the crowdsourcing relevance judgment task, one of the decision to be made is the how granular the relevance scale used to collect judgments should be. Such decision has then repercussions on the metrics used to measure IR system effectiveness. In this paper we look at the effect of scale transformations in a systematic way. We perform extensive experiments to study the transformation of judgments from fine-grained to coarse-grained. We use different relevance judgments expressed on different relevance scales and either expressed by expert annotators or collected by means of crowdsourcing. The objective is to understand the impact of relevance scale transformations on IR evaluation outcomes and to draw conclusions on how to best transform judgments into a different scale, when necessary.

Recently proposed methods allow the generation of simulated scores representing the values of an effectiveness metric, but they do not investigate the generation of the actual lists of retrieved documents. In this paper we address this limitation: we present an approach that exploits an evolutionary algorithm and, given a metric score, creates a simulated relevance profile (i.e., a ranked list of relevance values) that produces that score. We show how the simulated relevance profiles are realistic under various analyses.

In a test collection setting, topic difficulty can be defined as the average effectiveness of a set of systems for a topic. In this paper we study the effects on the topic difficulty of: (i) the set of retrieval systems; (ii) the underlying document corpus; and (iii) the system components. By generalizing methods recently proposed to study system component factor analysis, we perform a comprehensive analysis on topic difficulty and the relative effects of systems, corpora, and component interactions. Our findings show that corpora have the most significant effect on topic difficulty.

Crowdsourcing has become a standard methodology to collect manually annotated data such as relevance judgments at scale. On crowdsourcing platforms like Amazon MTurk or FigureEight, crowd workers select tasks to work on based on different dimensions such as task reward and requester reputation. Requesters then receive the judgments of workers who self-selected into the tasks and completed them successfully. Several crowd workers, however, preview tasks, begin working on them, reaching varying stages of task completion without finally submitting their work. Such behavior results in unrewarded effort which remains invisible to requesters. In this paper, we conduct the first investigation into the phenomenon of task abandonment, the act of workers previewing or beginning a task and deciding not to complete it. We follow a three-fold methodology which includes 1) investigating the prevalence and causes of task abandonment by means of a survey over different crowdsourcing platforms, 2) data-driven analyses of logs collected during a large-scale relevance judgment experiment, and 3) controlled experiments measuring the effect of different dimensions on abandonment. Our results show that task abandonment is a widely spread phenomenon. Apart from accounting for a considerable amount of wasted human effort, this bears important implications on the hourly wages of workers as they are not rewarded for tasks that they do not complete. We also show how task abandonment may have strong implications on the use of collected data (for example, on the evaluation of IR systems).

Some methods have been developed for automatic effectiveness evaluation without relevance judgments. We propose to use those methods, and their combination based on a machine learning approach, for query performance prediction. Moreover, since predicting average precision as it is usually done in query performance prediction literature is sensitive to the reference system that is chosen, we focus on predicting the average of average precision values over several systems. Results of an extensive experimental evaluation on ten TREC collections show that our proposed methods outperform state-of-the-art query performance predictors.

In test collection based evaluation of retrieval effectiveness, many research investigated different directions for an economical and a semi-automatic evaluation of retrieval systems. Although several methods have been proposed and experimentally evaluated, their accuracy seems still limited. In this paper we present our proposal for a more engineered approach to information retrieval evaluation.

In Information Retrieval evaluation, the classical approach of adopting binary relevance judgments has been replaced by multi-level relevance judgments and by gain-based metrics leveraging such multi-level judgment scales. Recent work has also proposed and evaluated unbounded relevance scales by means of Magnitude Estimation (ME) and compared them with multi-level scales. While ME brings advantages like the ability for assessors to always judge the next document as having higher or lower relevance than any of the documents they have judged so far, it also comes with some drawbacks. For example, it is not a natural approach for human assessors to judge items as they are used to do on the Web (e.g., 5-star rating). In this work, we propose and experimentally evaluate a bounded and fine-grained relevance scale having many of the advantages and dealing with some of the issues of ME. We collect relevance judgments over a 100-level relevance scale (S100) by means of a large-scale crowdsourcing experiment and compare the results with other relevance scales (binary, 4-level, and ME) showing the benefit of fine-grained scales over both coarse-grained and unbounded scales as well as highlighting some new results on ME. Our results show that S100 maintains the flexibility of unbounded scales like ME in providing assessors with ample choice when judging document relevance (i.e., assessors can fit relevance judgments in between of previously given judgments). It also allows assessors to judge on a more familiar scale (e.g., on 10 levels) and to perform efficiently since the very first judging task.

We propose IRevalOO, a flexible Object Oriented framework that (i) can be used as-is as a replacement of the widely adopted trec\_eval software, and (ii) can be easily extended (or "instantiated'', in framework terminology) to implement different scenarios of test collection based retrieval evaluation. Instances of IRevalOO can provide a usable and convenient alternative to the state-of-the-art software commonly used by different initiatives (TREC, NTCIR, CLEF, FIRE, etc.). Also, those instances can be easily adapted to satisfy future customization needs of researchers, as: implementing and experimenting with new metrics, even based on new notions of relevance; using different formats for system output and "qrels''; and in general visualizing, comparing, and managing retrieval evaluation results.

Several researchers have proposed to reduce the number of topics used in TREC-like initiatives. One research direction that has been pursued is what is the optimal topic subset of a given cardinality that evaluates the systems/runs in the most accurate way. Such a research direction has been so far mainly theoretical, with almost no indication on how to select the few good topics in practice. We propose such a practical criterion for topic selection: we rely on the methods for automatic system evaluation without relevance judgments, and by running some experiments on several TREC collections we show that the topics selected on the basis of those evaluations are indeed more informative than random topics.

In the context of micro-task crowdsourcing, each task is usually performed by several workers. This allows researchers to leverage measures of the agreement among workers on the same task, to estimate the reliability of collected data and to better understand answering behaviors of the participants. While many measures of agreement between annotators have been proposed, they are known for suffering from many problems and abnormalities. In this paper, we identify the main limits of the existing agreement measures in the crowdsourcing context, both by means of toy examples as well as with real-world crowdsourcing data, and propose a novel agreement measure based on probabilistic parameter estimation which overcomes such limits. We validate our new agreement measure and show its flexibility as compared to the existing agreement measures.