Jorge Pérez

Cinthia Sánchez, Hernan Sarmiento, Andres Abeliuk et al.

Social media data has emerged as a useful source of timely information about real-world crisis events. One of the main tasks related to the use of social media for disaster management is the automatic identification of crisis-related messages. Most of the studies on this topic have focused on the analysis of data for a particular type of event in a specific language. This limits the possibility of generalizing existing approaches because models cannot be directly applied to new types of events or other languages. In this work, we study the task of automatically classifying messages that are related to crisis events by leveraging cross-language and cross-domain labeled data. Our goal is to make use of labeled data from high-resource languages to classify messages from other (low-resource) languages and/or of new (previously unseen) types of crisis situations. For our study we consolidated from the literature a large unified dataset containing multiple crisis events and languages. Our empirical findings show that it is indeed possible to leverage data from crisis events in English to classify the same type of event in other languages, such as Spanish and Italian (80.0% F1-score). Furthermore, we achieve good performance for the cross-domain task (80.0% F1-score) in a cross-lingual setting. Overall, our work contributes to improving the data scarcity problem that is so important for multilingual crisis classification. In particular, mitigating cold-start situations in emergency events, when time is of essence.

José Cañete, Gabriel Chaperon, Rodrigo Fuentes et al.

The Spanish language is one of the top 5 spoken languages in the world. Nevertheless, finding resources to train or evaluate Spanish language models is not an easy task. In this paper we help bridge this gap by presenting a BERT-based language model pre-trained exclusively on Spanish data. As a second contribution, we also compiled several tasks specifically for the Spanish language in a single repository much in the spirit of the GLUE benchmark. By fine-tuning our pre-trained Spanish model, we obtain better results compared to other BERT-based models pre-trained on multilingual corpora for most of the tasks, even achieving a new state-of-the-art on some of them. We have publicly released our model, the pre-training data, and the compilation of the Spanish benchmarks.

Francisco Plana, Andrés Abeliuk, Jorge Pérez

We present a simple and quick method to approximate network centrality indexes. Our approach, called QuickCent, is inspired by so-called fast and frugal heuristics, which are heuristics initially proposed to model some human decision and inference processes. The centrality index that we estimate is the harmonic centrality, which is a measure based on shortest-path distances, so infeasible to compute on large networks. We compare QuickCent with known machine learning algorithms on synthetic data generated with preferential attachment, and some empirical networks. Our experiments show that QuickCent is able to make estimates that are competitive in accuracy with the best alternative methods tested, either on synthetic scale-free networks or empirical networks. QuickCent has the feature of achieving low error variance estimates, even with a small training set. Moreover, QuickCent is comparable in efficiency -- accuracy and time cost -- to those produced by more complex methods. We discuss and provide some insight into how QuickCent exploits the fact that in some networks, such as those generated by preferential attachment, local density measures such as the in-degree, can be a proxy for the size of the network region to which a node has access, opening up the possibility of approximating centrality indices based on size such as the harmonic centrality. Our initial results show that simple heuristics and biologically inspired computational methods are a promising line of research in the context of network measure estimations.

Diego Gomez-Zara, Hernán Valdivieso, Jorge Pérez et al.

Codebooks are central to framing research, providing theoretically grounded criteria for analyzing news content. While traditionally codebooks are built from theoretical frameworks and researchers' knowledge, applying these codebooks to large news corpora often exposes ambiguities, borderline cases, and underspecified rules that are difficult to resolve through theory alone. Moreover, news corpora evolve over time and differ across cultures, necessitating that researchers revisit the theoretical frameworks underlying these codebooks. In this article, we propose a workflow that uses Large Language Models (LLMs) to augment the creation and refinement of framing codebooks by combining theoretical frameworks with data-driven exploration. Rather than treating LLMs as automated classifiers, this approach positions them as analytic collaborators that help externalize decision rules, surface latent dimensions, and support iterative revisions of codebooks through dialogues between researchers and their data. We illustrate this workflow using a dataset of Latin American news coverage, demonstrating how the application of LLMs' capabilities has led to the surfacing of latent patterns, the generation of frame distinctions, and the adaptation of frameworks to new contexts. This method provides an LLM-assisted strategy that supports methodology creativity while preserving researchers' interpretative authority.

Marcelo Arenas, Daniel Baez, Pablo Barceló et al.

Several queries and scores have recently been proposed to explain individual predictions over ML models. Given the need for flexible, reliable, and easy-to-apply interpretability methods for ML models, we foresee the need for developing declarative languages to naturally specify different explainability queries. We do this in a principled way by rooting such a language in a logic, called FOIL, that allows for expressing many simple but important explainability queries, and might serve as a core for more expressive interpretability languages. We study the computational complexity of FOIL queries over two classes of ML models often deemed to be easily interpretable: decision trees and OBDDs. Since the number of possible inputs for an ML model is exponential in its dimension, the tractability of the FOIL evaluation problem is delicate but can be achieved by either restricting the structure of the models or the fragment of FOIL being evaluated. We also present a prototype implementation of FOIL wrapped in a high-level declarative language and perform experiments showing that such a language can be used in practice.

Jessica Díaz, Jorge Pérez, Carolina Gallardo et al.

In recent years, the qualitative research on empirical software engineering that applies Grounded Theory is increasing. Grounded Theory (GT) is a technique for developing theory inductively e iteratively from qualitative data based on theoretical sampling, coding, constant comparison, memoing, and saturation, as main characteristics. Large or controversial GT studies may involve multiple researchers in collaborative coding, which requires a kind of rigor and consensus that an individual coder does not. Although many qualitative researchers reject quantitative measures in favor of other qualitative criteria, many others are committed to measuring consensus through Inter-Rater Reliability (IRR) and/or Inter-Rater Agreement (IRA) techniques to develop a shared understanding of the phenomenon being studied. However, there are no specific guidelines about how and when to apply IRR/IRA during the iterative process of GT, so researchers have been using ad hoc methods for years. This paper presents a process for systematically applying IRR/IRA in GT studies that meets the iterative nature of this qualitative research method, which is supported by a previous systematic literature review on applying IRR/RA in GT studies in software engineering. This process allows researchers to incrementally generate a theory while ensuring consensus on the constructs that support it and, thus, improving the rigor of qualitative research. This formalization helps researchers to apply IRR/IRA to GT studies when various raters are involved in coding. Measuring consensus among raters promotes communicability, transparency, reflexivity, replicability, and trustworthiness of the research.

Aymé Arango, Jorge Pérez, Barbara Poblete

Automatic hate speech detection in online social networks is an important open problem in Natural Language Processing (NLP). Hate speech is a multidimensional issue, strongly dependant on language and cultural factors. Despite its relevance, research on this topic has been almost exclusively devoted to English. Most supervised learning resources, such as labeled datasets and NLP tools, have been created for this same language. Considering that a large portion of users worldwide speak in languages other than English, there is an important need for creating efficient approaches for multilingual hate speech detection. In this work we propose to address the problem of multilingual hate speech detection from the perspective of transfer learning. Our goal is to determine if knowledge from one particular language can be used to classify other language, and to determine effective ways to achieve this. We propose a hate specific data representation and evaluate its effectiveness against general-purpose universal representations most of which, unlike our proposed model, have been trained on massive amounts of data. We focus on a cross-lingual setting, in which one needs to classify hate speech in one language without having access to any labeled data for that language. We show that the use of our simple yet specific multilingual hate representations improves classification results. We explain this with a qualitative analysis showing that our specific representation is able to capture some common patterns in how hate speech presents itself in different languages. Our proposal constitutes, to the best of our knowledge, the first attempt for constructing multilingual specific-task representations. Despite its simplicity, our model outperformed the previous approaches for most of the experimental setups. Our findings can orient future solutions toward the use of domain-specific representations.

Jesus Perez-Martin, Benjamin Bustos, Silvio Jamil F. Guimarães et al.

Research in the Vision and Language area encompasses challenging topics that seek to connect visual and textual information. When the visual information is related to videos, this takes us into Video-Text Research, which includes several challenging tasks such as video question answering, video summarization with natural language, and video-to-text and text-to-video conversion. This paper reviews the video-to-text problem, in which the goal is to associate an input video with its textual description. This association can be mainly made by retrieving the most relevant descriptions from a corpus or generating a new one given a context video. These two ways represent essential tasks for Computer Vision and Natural Language Processing communities, called text retrieval from video task and video captioning/description task. These two tasks are substantially more complex than predicting or retrieving a single sentence from an image. The spatiotemporal information present in videos introduces diversity and complexity regarding the visual content and the structure of associated language descriptions. This review categorizes and describes the state-of-the-art techniques for the video-to-text problem. It covers the main video-to-text methods and the ways to evaluate their performance. We analyze twenty-six benchmark datasets, showing their drawbacks and strengths for the problem requirements. We also show the progress that researchers have made on each dataset, we cover the challenges in the field, and we discuss future research directions.

Daniel López-Fernández, Jessica Díaz, Javier García et al.

Context: DevOps can be defined as a cultural movement to improve and accelerate the delivery of business value by making the collaboration between development and operations effective. Objective: This paper aims to help practitioners and researchers to better understand the organizational structure and characteristics of teams adopting DevOps. Method: We conducted an exploratory study by leveraging in depth, semi-structured interviews to relevant stakeholders of 31 multinational software-intensive companies, together with industrial workshops and observations at organizations' facilities that supported triangulation. We used Grounded Theory as qualitative research method to explore the structure and characteristics of teams, and statistical analysis to discover their implications in software delivery performance. Results: We describe a taxonomy of team structure patterns that shows emerging, stable and consolidated product teams that are classified according to six variables, such as collaboration frequency, product ownership sharing, autonomy, among others, as well as their implications on software delivery performance. These teams are often supported by horizontal teams (DevOps platform teams, Centers of Excellence, and chapters) that provide them with platform technical capability, mentoring and evangelization, and even temporarily facilitate human resources. Conclusion: This study aims to strengthen evidence and support practitioners in making better informed about organizational team structures by analyzing their main characteristics and implications in software delivery performance.

Pablo Barceló, Mikaël Monet, Jorge Pérez et al.

In spite of several claims stating that some models are more interpretable than others -- e.g., "linear models are more interpretable than deep neural networks" -- we still lack a principled notion of interpretability to formally compare among different classes of models. We make a step towards such a notion by studying whether folklore interpretability claims have a correlate in terms of computational complexity theory. We focus on local post-hoc explainability queries that, intuitively, attempt to answer why individual inputs are classified in a certain way by a given model. In a nutshell, we say that a class $\mathcal{C}_1$ of models is more interpretable than another class $\mathcal{C}_2$, if the computational complexity of answering post-hoc queries for models in $\mathcal{C}_2$ is higher than for those in $\mathcal{C}_1$. We prove that this notion provides a good theoretical counterpart to current beliefs on the interpretability of models; in particular, we show that under our definition and assuming standard complexity-theoretical assumptions (such as P$\neq$NP), both linear and tree-based models are strictly more interpretable than neural networks. Our complexity analysis, however, does not provide a clear-cut difference between linear and tree-based models, as we obtain different results depending on the particular post-hoc explanations considered. Finally, by applying a finer complexity analysis based on parameterized complexity, we are able to prove a theoretical result suggesting that shallow neural networks are more interpretable than deeper ones.

Jorge Pérez, Jessica Díaz, Javier Garcia-Martin et al.

Context: Systematic literature reviews (SLRs) rely on a rigorous and auditable methodology for minimizing biases and ensuring reliability. A common kind of bias arises when selecting studies using a set of inclusion/exclusion criteria. This bias can be decreased through dual revision, which makes the selection process more time-consuming and remains prone to generating bias depending on how each researcher interprets the inclusion/exclusion criteria. Objective: To reduce the bias and time spent in the study selection process, this paper presents a process for selecting studies based on the use of Cohen's Kappa statistic. We have defined an iterative process based on the use of this statistic during which the criteria are refined until obtain almost perfect agreement (k>0.8). At this point, the two researchers interpret the selection criteria in the same way, and thus, the bias is reduced. Starting from this agreement, dual review can be eliminated; consequently, the time spent is drastically shortened. Method: The feasibility of this iterative process for selecting studies is demonstrated through a tertiary study in the area of software engineering on works that were published from 2005 to 2018. Results: The time saved in the study selection process was 28% (for 152 studies) and if the number of studies is sufficiently large, the time saved tend asymptotically to 50%. Conclusions: Researchers and students may take advantage of this iterative process for selecting studies when conducting SLRs to reduce bias in the interpretation of inclusion and exclusion criteria. It is especially useful for research with few resources.

Constanza Fierro, Jorge Pérez, Javier Mora

Deep learning techniques have been successfully applied to predict unplanned readmissions of patients in medical centers. The training data for these models is usually based on historical medical records that contain a significant amount of free-text from admission reports, referrals, exam notes, etc. Most of the models proposed so far are tailored to English text data and assume that electronic medical records follow standards common in developed countries. These two characteristics make them difficult to apply in developing countries that do not necessarily follow international standards for registering patient information, or that store text information in languages other than English. In this paper we propose a deep learning architecture for predicting unplanned readmissions that consumes data that is significantly less structured compared with previous models in the literature. We use it to present the first results for this task in a large clinical dataset that mainly contains Spanish text data. The dataset is composed of almost 10 years of records in a Chilean medical center. On this dataset, our model achieves results that are comparable to some of the most recent results obtained in US medical centers for the same task (0.76 AUROC).

Jorge Pérez, Javier Marinković, Pablo Barceló

Alternatives to recurrent neural networks, in particular, architectures based on attention or convolutions, have been gaining momentum for processing input sequences. In spite of their relevance, the computational properties of these alternatives have not yet been fully explored. We study the computational power of two of the most paradigmatic architectures exemplifying these mechanisms: the Transformer (Vaswani et al., 2017) and the Neural GPU (Kaiser & Sutskever, 2016). We show both models to be Turing complete exclusively based on their capacity to compute and access internal dense representations of the data. In particular, neither the Transformer nor the Neural GPU requires access to an external memory to become Turing complete. Our study also reveals some minimal sets of elements needed to obtain these completeness results.