ISParity

Repository of work and experiments in XAI and fairness Machine Learning.
⭐ Data Science & Artificial Intelligence ⭐

---

Authors & contributors:

Alessandro Castelnovo, Riccardo Crupi, Daniele Regoli

---

To know more about the research work of our team of ⭐ Data Science & Artificial Intelligence ⭐ , please refer to our publications

Journal papers & Conference proceedings

Fairness

• Bias on Demand: A Modelling Framework That Generates Synthetic Data With Bias
  Joachim Baumann, Alessandro Castelnovo, Riccardo Crupi, Nicole Inverardi, Daniele Regoli
  FAccT '23: Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency June 2023, Pages 1002–1013 (video presentation at FAccT)

  Nowadays, Machine Learning (ML) systems are widely used in various businesses and are increasingly being adopted to make decisions that can significantly impact people’s lives. However, these decision-making systems rely on data-driven learning, which poses a risk of propagating the bias embedded in the data. Despite various attempts by the algorithmic fairness community to outline different types of bias in data and algorithms, there is still a limited understanding of how these biases relate to the fairness of ML-based decision-making systems. In addition, efforts to mitigate bias and unfairness are often agnostic to the specific type(s) of bias present in the data. This paper explores the nature of fundamental types of bias, discussing their relationship to moral and technical frameworks. To prevent harmful consequences, it is essential to comprehend how and where bias is introduced throughout the entire modelling pipeline and possibly how to mitigate it. Our primary contribution is a framework for generating synthetic datasets with different forms of biases. We use our proposed synthetic data generator to perform experiments on different scenarios to showcase the interconnection between biases and their effect on performance and fairness evaluations. Furthermore, we provide initial insights into mitigating specific types of bias through post-processing techniques. The implementation of the synthetic data generator and experiments can be found at https://github.com/rcrupiISP/BiasOnDemand.

  • An Open-Source Toolkit to Generate Biased Datasets
    Joachim Baumann, Alessandro Castelnovo, Riccardo Crupi, Nicole Inverardi, Daniele Regoli
    EWAF’23: European Workshop on Algorithmic Fairness, June 07–09, 2023, Winterthur, Switzerland

    Many different types of bias are discussed in the algorithmic fairness community. A clear understanding of those biases and their relation to fairness metrics and mitigation techniques is still missing. We introduce Bias on Demand: a modelling framework to generate synthetic datasets that contain various types of bias. Furthermore, we clarify the effect of those biases on the accuracy and fairness of ML systems and provide insights into the trade-offs that emerge when trying to mitigate them. We believe that our open-source package will enable researchers and practitioners to better understand and mitigate different types of biases throughout the ML pipeline. The package can be installed via pip and the experiments are available at https://github.com/rcrupiISP/BiasOnDemand. We encourage readers to consult the full paper.

• A clarification of the nuances in the fairness metrics landscape
  Alessandro Castelnovo, Riccardo Crupi, Greta Greco, Daniele Regoli, Ilaria Penco, Andrea Cosentini
  Scientific Reports 12, 4209 (2022)

  In recent years, the problem of addressing fairness in machine learning (ML) and automatic decision making has attracted a lot of attention in the scientific communities dealing with artificial intelligence. A plethora of different definitions of fairness in ML have been proposed, that consider different notions of what is a “fair decision” in situations impacting individuals in the population. The precise differences, implications and “orthogonality” between these notions have not yet been fully analyzed in the literature. In this work, we try to make some order out of this zoo of definitions.

• Towards Responsible AI: A Design Space Exploration of Human-Centered Artificial Intelligence User Interfaces to Investigate Fairness
  Yuri Nakao, Lorenzo Strappelli, Simone Stumpf, Aisha Naseer, Daniele Regoli, Giulia Del Gamba
  International Journal of Human–Computer Interaction 2022

  With Artificial intelligence (AI) to aid or automate decision-making advancing rapidly, a particular concern is its fairness. In order to create reliable, safe and trustworthy systems through human-centred artificial intelligence (HCAI) design, recent efforts have produced user interfaces (UIs) for AI experts to investigate the fairness of AI models. In this work, we provide a design space exploration that supports not only data scientists but also domain experts to investigate AI fairness. Using loan applications as an example, we held a series of workshops with loan officers and data scientists to elicit their requirements. We instantiated these requirements into FairHIL, a UI to support human-in-the-loop fairness investigations, and describe how this UI could be generalized to other use cases. We evaluated FairHIL through a think-aloud user study. Our work contributes better designs to investigate an AI model’s fairness—and move closer towards responsible AI.

• FFTree: A flexible tree to handle multiple fairness criteria
  Alessandro Castelnovo, Andrea Cosentini, Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica
  Information Processing & Management 2022

  The demand for transparency and fairness in AI-based decision-making systems is constantly growing. Organisations need to be assured that their applications, based on these technologies, behave fairly, without introducing negative social implications in relation to sensitive attributes such as gender or race. Since the notion of fairness is context dependent and not uniquely defined, studies in the literature have proposed various formalisation. In this work, we propose a novel, flexible, discrimination-aware decision-tree that allows the user to employ different fairness criteria depending on the application domain. Our approach enhances decision-tree classifiers to provide transparent and fair rules to final users.

• BeFair: Addressing Fairness in the Banking Sector
  Alessandro Castelnovo, Riccardo Crupi, Giulia Del Gamba, Greta Greco, Aisha Naseer, Daniele Regoli, Beatriz San Miguel Gonzalez
  2020 IEEE International Conference on Big Data (Big Data)
  (preprint version)

  Algorithmic bias mitigation has been one of the most difficult conundrums for the data science community and Machine Learning (ML) experts. Over several years, there have appeared enormous efforts in the field of fairness in ML. Despite the progress toward identifying biases and designing fair algorithms, translating them into the industry remains a major challenge. In this paper, we present the initial results of an industrial open innovation project in the banking sector: we propose a general roadmap for fairness in ML and the implementation of a toolkit called BeFair that helps to identify and mitigate bias. Results show that training a model without explicit constraints may lead to bias exacerbation in the predictions.

• Fundamental Rights and Artificial Intelligence Impact Assessment: A New Quantitative Methodology in the Upcoming Era of Ai Act
  Bertaina, Samuele and Biganzoli, Ilaria and Desiante, Rachele and Fontanella, Dario and Inverardi, Nicole and Penco, Ilaria Giuseppina and Cosentini, Andrea.
  Fundamental Rights and Artificial Intelligence Impact Assessment: A New Quantitative Methodology in the Upcoming Era of Ai Act.

  The upcoming Artificial Intelligence Act should require that operators of Artificial Intelligence (AI) systems perform a Fundamental Rights Impact Assessment (FRIA) when such systems are classified as high-risk as defined in Annex III of the proposed regulation. The aim of this work is to offer a comprehensive framework, specifically thought for private businesses, to assess the impact of AI systems on Fundamental Rights (FR) – also named human rights – of individuals. In a nutshell, the assessment approach that we propose consists of two stages: (1) an open-ended survey that helps gather the contextual information and the technical features, in order to properly identify potential threats for FR, and (2) a quantitative matrix that considers each right guaranteed by the European Charter of Fundamentals Rights and tries to measure the potential impacts with a traceable and robust procedure. In light of an increasingly pervasive use of AI systems and considering the specificity of such technologies, we believe that a structured and quantitative process for assessing the impact on FR of individuals is still lacking and could be of great importance in discovering and remedying possible violations. Indeed, the proposed framework could allow to: (1) be accountable and transparent in assessing the risks of implementing AI systems that affect people; (2) gain insights to understand if any right is threatened or any group of people is more vulnerable; (3) put in place, if necessary, remediation strategies before the deployment of AI systems through demonstrable mitigative actions, with the aim of being compliant with the regulation and limiting reputational damage.

• Preserving Utility in Fair Top-k Ranking with Intersectional Bias
  Nicola Alimonda, Alessandro Castelnovo, Riccardo Crupi, Fabio Mercorio, Mario Mezzanzanica
  International Workshop on Algorithmic Bias in Search and Recommendation, 59-73

  Ranking is required for many real applications, such as search, personalisation, recommendation, and filtering. Recent research has focused on developing reliable ranking algorithms that maintain fairness in their outcomes. However, only a few consider multiple protected groups since this extension introduces significant challenges. While useful in the research sector, considering only one binary sensitive feature for handling fairness is inappropriate when the algorithm must be deployed responsibly in real-world applications. Our work is built on top of Multinomial FAIR, a Fair Top-k ranking with multiple protected groups, which we extend to provide users the option to balance fairness and utility, adapting the final ranking accordingly. Our experimental results show that alternative better solutions overlooked by Multinomial FAIR may be found through our approach without violating fairness boundaries. The code of the implemented solution and the experiments are publicly available to the community as a GitHub repository.

• Marrying LLMs with Domain Expert Validation for Causal Graph Generation
  Alessandro Castelnovo, Riccardo Crupi, Fabio Mercorio, Mario Mezzanzanica, Daniele Potertì, Daniele Regoli
  3rd Italian Workshop on Artificial Intelligence and Applications for Business and Industries - AIABI co-located with AI*IA 2023

  In the era of rapid growth and transformation driven by artificial intelligence across various sectors, which is catalyzing the fourth industrial revolution, this research is directed toward harnessing its potential to enhance the efficiency of decision-making processes within organizations. When constructing machine learning-based decision models, a fundamental step involves the conversion of domain knowledge into causal-effect relationships that are represented in causal graphs. This process is also notably advantageous for constructing explanation models. We present a method for generating causal graphs that integrates the strengths of Large Language Models (LLMs) with traditional causal theory algorithms. Our method seeks to bridge the gap between AI’s theoretical potential and practical applications. In contrast to recent related works that seek to exclude the involvement of domain experts, our method places them at the forefront of the process. We present a novel pipeline that streamlines and enhances domain-expert validation by providing robust causal graph proposals. These proposals are enriched with transparent reports that blend foundational causal theory reasoning with explanations from LLMs.

• Fair Enough? A map of the current limitations of the requirements to have" fair''algorithms
  Alessandro Castelnovo, Nicole Inverardi, Gabriele Nanino, Ilaria Giuseppina Penco, Daniele Regoli

  In the recent years, the raise in the usage and efficiency of Artificial Intelligence and, more in general, of Automated Decision-Making systems has brought with it an increasing and welcome awareness of the risks associated with such systems. One of such risks is that of perpetuating or even amplifying bias and unjust disparities present in the data from which many of these systems learn to adjust and optimise their decisions. This awareness has on one side encouraged several scientific communities to come up with more and more appropriate ways and methods to assess, quantify, and possibly mitigate such biases and disparities. On the other hand, it has prompted more and more layers of society, including policy makers, to call for ``fair'' algorithms. We believe that while a lot of excellent and multidisciplinary research is currently being conducted, what is still fundamentally missing is the awareness that having "fair" algorithms is per s'e a nearly meaningless requirement, that needs to be complemented with a lot of additional societal choices to become actionable. Namely, there is a hiatus between what the society is demanding from Automated Decision-Making systems, and what this demand actually means in real-world scenarios. In this work, we outline the key features of such a hiatus, and pinpoint a list of fundamental ambiguities and attention points that we as a society must address in order to give a concrete meaning to the increasing demand of fairness in Automated Decision-Making systems.

• Leveraging Group Contrastive Explanations for Handling Fairness
  Alessandro Castelnovo, Nicole Inverardi, Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica & Andrea Seveso

  With the increasing adoption of Artificial Intelligence (AI) for decision-making processes by companies, developing systems that behave fairly and do not discriminate against specific groups of people becomes crucial. Reaching this objective requires a multidisciplinary approach that includes domain experts, data scientists, philosophers, and legal experts, to ensure complete accountability for algorithmic decisions. In such a context, Explainable AI (XAI) plays a key role in enabling professionals from different backgrounds to comprehend the functioning of automatized decision-making processes and, consequently, being able to identify the presence of fairness issues. This paper presents FairX, an innovative approach that uses Group-Contrastive (G-contrast) explanations to estimate whether different decision criteria apply among distinct population subgroups. FairX provides actionable insights through a comprehensive explanation of the decision-making process, enabling businesses to: detect the presence of direct discrimination on the target variable and choose the most appropriate fairness framework.

eXplainable AI

• Counterfactual explanations as interventions in latent space
  Riccardo Crupi, Alessandro Castelnovo, Daniele Regoli, Beatriz San Miguel Gonzalez
  Data Mining and Knowledge Discovery (2022)
  (prerpint version)

  Explainable Artificial Intelligence (XAI) is a set of techniques that allows the understanding of both technical and non-technical aspects of Artificial Intelligence (AI) systems. XAI is crucial to help satisfying the increasingly important demand of trustworthy Artificial Intelligence, characterized by fundamental aspects such as respect of human autonomy, prevention of harm, transparency, accountability, etc. Within XAI techniques, counterfactual explanations aim to provide to end users a set of features (and their corresponding values) that need to be changed in order to achieve a desired outcome. Current approaches rarely take into account the feasibility of actions needed to achieve the proposed explanations, and in particular, they fall short of considering the causal impact of such actions. In this paper, we present Counterfactual Explanations as Interventions in Latent Space (CEILS), a methodology to generate counterfactual explanations capturing by design the underlying causal relations from the data, and at the same time to provide feasible recommendations to reach the proposed profile. Moreover, our methodology has the advantage that it can be set on top of existing counterfactuals generator algorithms, thus minimising the complexity of imposing additional causal constrains. We demonstrate the effectiveness of our approach with a set of different experiments using synthetic and real datasets (including a proprietary dataset of the financial domain).

• Leveraging Causal Relations to Provide Counterfactual Explanations and Feasible Recommendations to End Users
  Riccardo Crupi, Alessandro San Miguel González, Beatriz: Castelnovo, Daniele Regoli
  Proceedings of the 14th International Conference on Agents and Artificial Intelligence (ICAART 2022) - Volume 2, pages 24-32

  Over the last years, there has been a growing debate on the ethical issues of Artificial Intelligence (AI). Explainable Artificial Intelligence (XAI) has appeared as a key element to enhance trust of AI systems from both technological and human-understandable perspectives. In this sense, counterfactual explanations are becoming a de facto solution for end users to assist them in acting to achieve a desired outcome. In this paper, we present a new method called Counterfactual Explanations as Interventions in Latent Space (CEILS) to generate explanations focused on the production of feasible user actions. The main features of CEILS are: it takes into account the underlying causal relations by design, and can be set on top of an arbitrary counterfactual explanation generator. We demonstrate how CEILS succeeds through its evaluation on a real dataset of the financial domain.

• DTOR: Decision Tree Outlier Regressor to explain anomalies
  Riccardo Crupi, Daniele Regoli, Alessandro Damiano Sabatino, Immacolata Marano, Massimiliano Brinis, Luca Albertazzi, Andrea Cirillo, Andrea Claudio Cosentini
  GitHub

  Explaining outliers occurrence and mechanism of their occurrence can be extremely important in a variety of domains. Malfunctions, frauds, threats, in addition to being correctly identified, oftentimes need a valid explanation in order to effectively perform actionable counteracts. The ever more widespread use of sophisticated Machine Learning approach to identify anomalies make such explanations more challenging. We present the Decision Tree Outlier Regressor (DTOR), a technique for producing rule-based explanations for individual data points by estimating anomaly scores generated by an anomaly detection model. This is accomplished by first applying a Decision Tree Regressor, which computes the estimation score, and then extracting the relative path associated with the data point score. Our results demonstrate the robustness of DTOR even in datasets with a large number of features. Additionally, in contrast to other rule-based approaches, the generated rules are consistently satisfied by the points to be explained. Furthermore, our evaluation metrics indicate comparable performance to Anchors in outlier explanation tasks, with reduced execution time.

• Quantifying credit portfolio sensitivity to asset correlations with interpretable generative neural networks
  Sergio Caprioli, Emanuele Cagliero and Riccardo Crupi
  Journal of Risk Model Validation - ISSN: 1753-9579 (print) 1753-9587 (online)
  GitHub

  We propose a novel approach for quantifying the sensitivity of credit portfolio value-at-risk to asset correlations with the use of synthetic financial correlation matrixes generated with deep learning models. In previous work, generative adversarial networks (GANs) were employed to demonstrate the generation of plausible correlation matrixes that capture the essential characteristics observed in empirical correlation matrixes estimated on asset returns. Instead of GANs, we employ variational autoencoders (VAEs) to achieve a more interpretable latent space representation and to obtain a generator of plausible correlation matrixes by sampling the VAE’s latent space. Through our analysis, we reveal that the VAE’s latent space can be a useful tool to capture the crucial factors impacting portfolio diversification, particularly in relation to the sensitivity of credit portfolios to changes in asset correlations. A VAE trained on the historical time series of correlation matrixes is used to generate synthetic correlation matrixes that satisfy a set of expected financial properties. Our analysis provides clear indications that the capacity for realistic data augmentation provided by VAEs, combined with the ability to obtain model interpretability, can prove useful for risk management, enhancing the resilience and accuracy of models when backtesting, as past data may exhibit biases and might not contain the essential high-stress events required for evaluating diverse risk scenarios.

• Evaluative Item-Contrastive Explanations in Rankings
  A Castelnovo, R Crupi, N Mombelli, G Nanino, D Regoli

  The remarkable success of Artificial Intelligence in advancing automated decision-making is evident both in academia and industry. Within the plethora of applications, ranking systems hold significant importance in various domains. This paper advocates for the application of a specific form of Explainable AI -- namely, contrastive explanations -- as particularly well-suited for addressing ranking problems. This approach is especially potent when combined with an Evaluative AI methodology, which conscientiously evaluates both positive and negative aspects influencing a potential ranking. Therefore, the present work introduces Evaluative Item-Contrastive Explanations tailored for ranking systems and illustrates its application and characteristics through an experiment conducted on publicly available data.

Other topics

• Disambiguation of company names via deep recurrent networks
  Alessandro Basile, Riccardo Crupi, Michele Grasso, Alessandro Mercanti, Daniele Regoli, Simone Scarsi, Shuyi Yang, Andrea Claudio Cosentini.
  Expert Systems with Applications 238, 122035.
  GitHub

  Name Entity Disambiguation is the Natural Language Processing task of identifying textual records corresponding to the same Named Entity, i.e., real-world entities represented as a list of attributes (names, places, organisations, etc.). In this work, we face the task of disambiguating companies on the basis of their written names. We propose a Siamese LSTM Network approach to extract – via supervised learning – an embedding of company name strings in a (relatively) low dimensional vector space and use this representation to identify pairs of company names that actually represent the same company (i.e., the same Entity). Given that the manual labelling of string pairs is a rather onerous task, we analyse how an Active Learning approach to prioritise the samples to be labelled leads to a more efficient overall learning pipeline. The contributions of this work are: with empirical investigations on real-world industrial data, we show that our proposed Siamese Network outperforms several benchmark approaches based on standard string matching algorithms when enough labelled data are available; moreover, we show that Active Learning prioritisation is indeed helpful when labelling resources are limited, and let the learning models reach the out-of-sample performance saturation with less labelled data with respect to standard (random) data labelling approaches.

• Declarative Encoding of Fairness in Logic Tensor Networks Greta Greco, Federico Alberici, Matteo Palmonari, Andrea Cosentini.

  Algorithms are vulnerable to biases that might render their decisions unfair toward particular groups of individuals. Fairness comes with a range of facets that strongly depend on the application domain and that need to be enforced accordingly. However, most mitigation models embed fairness constraints as fundamental component of the loss function thus requiring code-level adjustments to adapt to specific contexts and domains. Rather than relying on a procedural approach, our model leverages declarative structured knowledge to encode fairness requirements in the form of logic rules capturing unambiguous and precise natural language statements. We propose a neuro-symbolic integration approach based on Logic Tensor Networks that combines data-driven network-based learning with high-level logical knowledge, allowing to perform classification tasks while reducing discrimination. Experimental evidence shows that performance is as good as state-of-the-art (SOTA) thus providing a flexible framework to account for non-discrimination often at a modest cost in terms of accuracy.

Lectures, Tutorials, and others

Tutorial Fairness and Explainability Machine learning Milano - Alessandro Castelnovo 2020

CEILS talk Causal UAI - Riccardo Crupi 2021

Seminar BeFair - Riccardo Crupi 2022

Fairness in AI Deep Learning Italia - Daniele Regoli 2022

Tutorial FFTree - 2022

Toward Fairness Through Time BIAS@ECMLPKDD - Alessandro Castelnovo 2022

Demo BeFair: Addressing Fairness in the Banking Sector - Greta Greco 2022

Preserving fairness in ranking BIAS@ECIR2023 - Nicola Alimonda 2023

BiasOnDemand ACM FAccT - Daniele Regoli 2023

Nexa lunch seminar - Towards responsible AI in banking - Alessandro Castelnovo 2024