Research Papers

A structured regression approach for evaluating model performance across intersectional subgroups

Christine Herlihy, Kimberly Truong, Alexandra Chouldechova +1

Disaggregated evaluation is a central task in AI fairness assessment, where the goal is to measure an AI system's performance across different subgroups defined by combinations of demographic or other sensitive attributes. The standard approach is to stratify the evaluation data across subgroups and compute performance metrics separately for each group. However, even for moderately-sized evaluation datasets, sample sizes quickly get small once considering intersectional subgroups, which greatly limits the extent to which intersectional groups are included in analysis. In this work, we introduce a structured regression approach to disaggregated evaluation that we demonstrate can yield reliable system performance estimates even for very small subgroups. We provide corresponding inference strategies for constructing confidence intervals and explore how goodness-of-fit testing can yield insight into the structure of fairness-related harms experienced by intersectional groups. We evaluate our approach on two publicly available datasets, and several variants of semi-synthetic data. The results show that our method is considerably more accurate than the standard approach, especially for small subgroups, and demonstrate how goodness-of-fit testing helps identify the key factors that drive differences in performance.

There is a growing demand for explainable, transparent, and data-driven models within the domain of fraud detection. Decisions made by fraud detection models need to be explainable in the event of a customer dispute. Additionally, the decision-making process in the model must be transparent to win the trust of regulators and business stakeholders. At the same time, fraud detection solutions can benefit from data due to the noisy, dynamic nature of fraud and the availability of large historical data sets. Finally, fraud detection is notorious for its class imbalance: there are typically several orders of magnitude more legitimate transactions than fraudulent ones. In this paper, we present Deep Symbolic Classification (DSC), an extension of the Deep Symbolic Regression framework to classification problems. DSC casts classification as a search problem in the space of all analytic functions composed of a vocabulary of variables, constants, and operations and optimizes for an arbitrary evaluation metric directly. The search is guided by a deep neural network trained with reinforcement learning. Because the functions are mathematical expressions that are in closed-form and concise, the model is inherently explainable both at the level of a single classification decision and the model's decision process. Furthermore, the class imbalance problem is successfully addressed by optimizing for metrics that are robust to class imbalance such as the F1 score. This eliminates the need for oversampling and undersampling techniques that plague traditional approaches. Finally, the model allows to explicitly balance between the prediction accuracy and the explainability. An evaluation on the PaySim data set demonstrates competitive predictive performance with state-of-the-art models, while surpassing them in terms of explainability. This establishes DSC as a promising model for fraud detection systems.

Responsible AI Research Needs Impact Statements Too

Alexandra Olteanu, Michael Ekstrand, Carlos Castillo +1

All types of research, development, and policy work can have unintended, adverse consequences - work in responsible artificial intelligence (RAI), ethical AI, or ethics in AI is no exception.

This article analyzes AI ethics from a distinct business ethics perspective, i.e., 'ordoliberalism 2.0.' It argues that the ongoing discourse on (generative) AI relies too much on corporate self-regulation and voluntary codes of conduct and thus lacks adequate governance mechanisms. To address these issues, the paper suggests not only introducing hard-law legislation with a more effective oversight structure but also merging already existing AI guidelines with an ordoliberal-inspired regulatory and competition policy. However, this link between AI ethics, regulation, and antitrust is not yet adequately discussed in the academic literature and beyond. The paper thus closes a significant gap in the academic literature and adds to the predominantly legal-political and philosophical discourse on AI governance. The paper's research questions and goals are twofold: First, it identifies ordoliberal-inspired AI ethics principles that could serve as the foundation for a 'digital bill of rights.' Second, it shows how those principles could be implemented at the macro level with the help of ordoliberal competition and regulatory policy.

Towards a Comprehensive Human-Centred Evaluation Framework for Explainable AI

Ivania Donoso-Guzmán, Jeroen Ooge, Denis Parra +1

While research on explainable AI (XAI) is booming and explanation techniques have proven promising in many application domains, standardised human-centred evaluation procedures are still missing. In addition, current evaluation procedures do not assess XAI methods holistically in the sense that they do not treat explanations' effects on humans as a complex user experience. To tackle this challenge, we propose to adapt the User-Centric Evaluation Framework used in recommender systems: we integrate explanation aspects, summarise explanation properties, indicate relations between them, and categorise metrics that measure these properties. With this comprehensive evaluation framework, we hope to contribute to the human-centred standardisation of XAI evaluation.

AI Alignment is often presented as an interaction between a single designer and an artificial agent in which the designer attempts to ensure the agent's behavior is consistent with its purpose, and risks arise solely because of conflicts caused by inadvertent misalignment between the utility function intended by the designer and the resulting internal utility function of the agent. With the advent of agents instantiated with large-language models (LLMs), which are typically pre-trained, we argue this does not capture the essential aspects of AI safety because in the real world there is not a one-to-one correspondence between designer and agent, and the many agents, both artificial and human, have heterogeneous values. Therefore, there is an economic aspect to AI safety and the principal-agent problem is likely to arise. In a principal-agent problem conflict arises because of information asymmetry together with inherent misalignment between the utility of the agent and its principal, and this inherent misalignment cannot be overcome by coercing the agent into adopting a desired utility function through training. We argue the assumptions underlying principal-agent problems are crucial to capturing the essence of safety problems involving pre-trained AI models in real-world situations. Taking an empirical approach to AI safety, we investigate how GPT models respond in principal-agent conflicts. We find that agents based on both GPT-3.5 and GPT-4 override their principal's objectives in a simple online shopping task, showing clear evidence of principal-agent conflict. Surprisingly, the earlier GPT-3.5 model exhibits more nuanced behaviour in response to changes in information asymmetry, whereas the later GPT-4 model is more rigid in adhering to its prior alignment. Our results highlight the importance of incorporating principles from economics into the alignment process.

On the Mechanics of NFT Valuation: AI Ethics and Social Media

Luyao Zhang, Yutong Sun, Yutong Quan +2

As CryptoPunks pioneers the innovation of non-fungible tokens (NFTs) in AI and art, the valuation mechanics of NFTs has become a trending topic. Earlier research identifies the impact of ethics and society on the price prediction of CryptoPunks. Since the booming year of the NFT market in 2021, the discussion of CryptoPunks has propagated on social media. Still, existing literature hasn't considered the social sentiment factors after the historical turning point on NFT valuation. In this paper, we study how sentiments in social media, together with gender and skin tone, contribute to NFT valuations by an empirical analysis of social media, blockchain, and crypto exchange data. We evidence social sentiments as a significant contributor to the price prediction of CryptoPunks. Furthermore, we document structure changes in the valuation mechanics before and after 2021. Although people's attitudes towards Cryptopunks are primarily positive, our findings reflect imbalances in transaction activities and pricing based on gender and skin tone. Our result is consistent and robust, controlling for the rarity of an NFT based on the set of human-readable attributes, including gender and skin tone. Our research contributes to the interdisciplinary study at the intersection of AI, Ethics, and Society, focusing on the ecosystem of decentralized AI or blockchain. We provide our data and code for replicability as open access on GitHub.

Classification and Explanation of Distributed Denial-of-Service (DDoS) Attack Detection using Machine Learning and Shapley Additive Explanation (SHAP) Methods

Yuanyuan Wei, Julian Jang-Jaccard, Amardeep Singh +2

DDoS attacks involve overwhelming a target system with a large number of requests or traffic from multiple sources, disrupting the normal traffic of a targeted server, service, or network. Distinguishing between legitimate traffic and malicious traffic is a challenging task. It is possible to classify legitimate traffic and malicious traffic and analysis the network traffic by using machine learning and deep learning techniques. However, an inter-model explanation implemented to classify a traffic flow whether is benign or malicious is an important investigation of the inner working theory of the model to increase the trustworthiness of the model. Explainable Artificial Intelligence (XAI) can explain the decision-making of the machine learning models that can be classified and identify DDoS traffic. In this context, we proposed a framework that can not only classify legitimate traffic and malicious traffic of DDoS attacks but also use SHAP to explain the decision-making of the classifier model. To address this concern, we first adopt feature selection techniques to select the top 20 important features based on feature importance techniques (e.g., XGB-based SHAP feature importance). Following that, the Multi-layer Perceptron Network (MLP) part of our proposed model uses the optimized features of the DDoS attack dataset as inputs to classify legitimate and malicious traffic. We perform extensive experiments with all features and selected features. The evaluation results show that the model performance with selected features achieves above 99\% accuracy. Finally, to provide interpretability, XAI can be adopted to explain the model performance between the prediction results and features based on global and local explanations by SHAP, which can better explain the results achieved by our proposed framework.

Artificial Intelligence (AI) has made impressive progress in recent years and represents a key technology that has a crucial impact on the economy and society. However, it is clear that AI and business models based on it can only reach their full potential if AI applications are developed according to high quality standards and are effectively protected against new AI risks. For instance, AI bears the risk of unfair treatment of individuals when processing personal data e.g., to support credit lending or staff recruitment decisions. The emergence of these new risks is closely linked to the fact that the behavior of AI applications, particularly those based on Machine Learning (ML), is essentially learned from large volumes of data and is not predetermined by fixed programmed rules. Thus, the issue of the trustworthiness of AI applications is crucial and is the subject of numerous major publications by stakeholders in politics, business and society. In addition, there is mutual agreement that the requirements for trustworthy AI, which are often described in an abstract way, must now be made clear and tangible. One challenge to overcome here relates to the fact that the specific quality criteria for an AI application depend heavily on the application context and possible measures to fulfill them in turn depend heavily on the AI technology used. Lastly, practical assessment procedures are needed to evaluate whether specific AI applications have been developed according to adequate quality standards. This AI assessment catalog addresses exactly this point and is intended for two target groups: Firstly, it provides developers with a guideline for systematically making their AI applications trustworthy. Secondly, it guides assessors and auditors on how to examine AI applications for trustworthiness in a structured way.

Complying with the EU AI Act (AIA) guidelines while developing and implementing AI systems will soon be mandatory within the EU. However, practitioners lack actionable instructions to operationalise ethics during AI systems development. A literature review of different ethical guidelines revealed inconsistencies in the principles addressed and the terminology used to describe them. Furthermore, requirements engineering (RE), which is identified to foster trustworthiness in the AI development process from the early stages was observed to be absent in a lot of frameworks that support the development of ethical and trustworthy AI. This incongruous phrasing combined with a lack of concrete development practices makes trustworthy AI development harder. To address this concern, we formulated a comparison table for the terminology used and the coverage of the ethical AI principles in major ethical AI guidelines. We then examined the applicability of ethical AI development frameworks for performing effective RE during the development of trustworthy AI systems. A tertiary review and meta-analysis of literature discussing ethical AI frameworks revealed their limitations when developing trustworthy AI. Based on our findings, we propose recommendations to address such limitations during the development of trustworthy AI.

A Survey of Trojans in Neural Models of Source Code: Taxonomy and Techniques

Aftab Hussain, Md Rafiqul Islam Rabin, Toufique Ahmed +4

In this work, we study literature in Explainable AI and Safe AI to understand poisoning of neural models of code. In order to do so, we first establish a novel taxonomy for Trojan AI for code, and present a new aspect-based classification of triggers in neural models of code. Next, we highlight recent works that help us deepen our conception of how these models understand software code. Then we pick some of the recent, state-of-art poisoning strategies that can be used to manipulate such models. The insights we draw can potentially help to foster future research in the area of Trojan AI for code.

Deep-BIAS: Detecting Structural Bias using Explainable AI

Bas van Stein, Diederick Vermetten, Fabio Caraffini +1

Evaluating the performance of heuristic optimisation algorithms is essential to determine how well they perform under various conditions. Recently, the BIAS toolbox was introduced as a behaviour benchmark to detect structural bias (SB) in search algorithms. The toolbox can be used to identify biases in existing algorithms, as well as to test for bias in newly developed algorithms. In this article, we introduce a novel and explainable deep-learning expansion of the BIAS toolbox, called Deep-BIAS. Where the original toolbox uses 39 statistical tests and a Random Forest model to predict the existence and type of SB, the Deep-BIAS method uses a trained deep-learning model to immediately detect the strength and type of SB based on the raw performance distributions. Through a series of experiments with a variety of structurally biased scenarios, we demonstrate the effectiveness of Deep-BIAS. We also present the results of using the toolbox on 336 state-of-the-art optimisation algorithms, which showed the presence of various types of structural bias, particularly towards the centre of the objective space or exhibiting discretisation behaviour. The Deep-BIAS method outperforms the BIAS toolbox both in detecting bias and for classifying the type of SB. Furthermore, explanations can be derived using XAI techniques.