Towards Fairness Through Time

Castelnovo, Alessandro; Malandri, Lorenzo; Mercorio, Fabio; Mezzanzanica, Mario; Cosentini, Andrea

doi:10.1007/978-3-030-93736-2_46

Cited by 8 publications

(4 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An interesting example of application in such sense is the paper Towards Fairness Through Time (Castelnovo et al, 2021), presented at the 2nd Workshop on Bias and Fairness in AI (BIAS) 6 at ECML-PKDD, which uses ContrXT to observe the evolution of a ML model for credit lending over time. Understanding the changing of the gaps between different population subgroups, like gender or race, allows observing whether the mitigation strategies in place are bringing benefits to society, favoring the convergence between individual and group fairness.…”

Section: Impact Statement and Ethical Considerationsmentioning

confidence: 99%

Contrastive Explanations of Text Classifiers as a Service

Malandri¹,

Mercorio²,

Mezzanzanica³

et al. 2022

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

Self Cite

View full text Add to dashboard Cite

The recent growth of black-box machinelearning methods in data analysis has increased the demand for explanation methods and tools to understand their behaviour and assist human-ML model cooperation. In this paper, we demonstrate ContrXT , a novel approach that uses natural language explanations to help users to comprehend how a back-box model works. ContrXT provides time contrastive (tcontrast) explanations by computing the differences in the classification logic of two different trained models and then reasoning on their symbolic representations through Binary Decision Diagrams. ContrXT is publicly available at ContrXT.ai as a python pip package.

show abstract

Section: Impact Statement and Ethical Considerationsmentioning

confidence: 99%

Contrastive Explanations of Text Classifiers as a Service

Malandri¹,

Mercorio²,

Mezzanzanica³

et al. 2022

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

Self Cite

View full text Add to dashboard Cite

show abstract

“…The practical application of XAI to fairness is explored by Castelnovo et al, who examine the limitations of conventional group fairness mitigation techniques in dynamic financial environments and offer a novel strategy fusing human behavioral insights for ongoing model refinement [128]. Stanley et al highlight the role of XAI in uncovering subgroup bias in medical imaging, where disparities in XAI outputs provide clues to underlying biases [129].…”

Section: Algorithmic Fairnessmentioning

confidence: 99%

“…Studies reviewed leverage a wide array of complex multidimensional or multi-modal datasets for XAI-enabled analysis. For societal applications, data types include demographic, financial, medical imaging, behavioral, and educational datasets, helping to address issues such as algorithmic fairness [122], [123], [128], digital ethics [131], and system accountability [137]. In the realm of scientific discovery, data types range from genomic sequences [160] to quantum system properties [158], passing through molecular structures for drug discovery [270] and clinical imaging data for healthcare applications [152].…”

Section: Data Collection and Acquisitionmentioning

confidence: 99%

“…Prior to XAI-enabled analysis, the data must undergo rigorous pre-processing. In societal contexts, pre-processing steps include standardization to mitigate bias, ensure fairness and transparency [129], and transform inconsistent or missing financial data for dynamic settings [128]. For scientific exploration, studies can employ normalization, feature extraction and selection techniques to hone the datasets relevant to the domain, such as chemical attributes for drug discovery [148] or physiological and environmental factors for medical diagnostics [154].…”

Section: Data Preparation and Transformationmentioning

confidence: 99%

See 1 more Smart Citation

Classification of Kinematic Data Using Explainable Artificial Intelligence (XAI) for Smart Motion

Yong¹

2021

Explainable Artificial Intelligence for Smart Cities

View full text Add to dashboard Cite

Given the complexity and lack of transparency in deep neural networks (DNNs), extensive efforts have been made to make these systems more interpretable or explain their behaviors in accessible terms. Unlike most reviews, which focus on algorithmic and model-centric perspectives, this work takes a "data-centric" view, examining how data collection, processing, and analysis contribute to explainable AI (XAI). We categorize existing work into three categories subject to their purposes: interpretations of deep models, referring to feature attributions and reasoning processes that correlate data points with model outputs; influences of training data, examining the impact of training data nuances, such as data valuation and sample anomalies, on decision-making processes; and insights of domain knowledge, discovering latent patterns and fostering new knowledge from data and models to advance social values and scientific discovery. Specifically, we distill XAI methodologies into data mining operations on training and testing data across modalities, such as images, text, and tabular data, as well as on training logs, checkpoints, models and other DNN behavior descriptors. In this way, our study offers a comprehensive, data-centric examination of XAI from a lens of data mining methods and applications.

show abstract