Research Papers

The lack of explainability of a decision from an Artificial Intelligence (AI) based "black box" system/model, despite its superiority in many real-world applications, is a key stumbling block for adopting AI in many high stakes applications of different domain or industry. While many popular Explainable Artificial Intelligence (XAI) methods or approaches are available to facilitate a human-friendly explanation of the decision, each has its own merits and demerits, with a plethora of open challenges. We demonstrate popular XAI methods with a mutual case study/task (i.e., credit default prediction), analyze for competitive advantages from multiple perspectives (e.g., local, global), provide meaningful insight on quantifying explainability, and recommend paths towards responsible or human-centered AI using XAI as a medium. Practitioners can use this work as a catalog to understand, compare, and correlate competitive advantages of popular XAI methods. In addition, this survey elicits future research directions towards responsible or human-centric AI systems, which is crucial to adopt AI in high stakes applications.

Unbiased Watermark for Large Language Models

Zhengmian Hu, Lichang Chen, Xidong Wu +3

The recent advancements in large language models (LLMs) have sparked a growing apprehension regarding the potential misuse. One approach to mitigating this risk is to incorporate watermarking techniques into LLMs, allowing for the tracking and attribution of model outputs. This study examines a crucial aspect of watermarking: how significantly watermarks impact the quality of model-generated outputs. Previous studies have suggested a trade-off between watermark strength and output quality. However, our research demonstrates that it is possible to integrate watermarks without affecting the output probability distribution with appropriate implementation. We refer to this type of watermark as an unbiased watermark. This has significant implications for the use of LLMs, as it becomes impossible for users to discern whether a service provider has incorporated watermarks or not. Furthermore, the presence of watermarks does not compromise the performance of the model in downstream tasks, ensuring that the overall utility of the language model is preserved. Our findings contribute to the ongoing discussion around responsible AI development, suggesting that unbiased watermarks can serve as an effective means of tracking and attributing model outputs without sacrificing output quality.

With the advances in artificial intelligence (AI), data-driven algorithms are becoming increasingly popular in the medical domain. However, due to the nonlinear and complex behavior of many of these algorithms, decision-making by such algorithms is not trustworthy for clinicians and is considered a black-box process. Hence, the scientific community has introduced explainable artificial intelligence (XAI) to remedy the problem. This systematic scoping review investigates the application of XAI in breast cancer detection and risk prediction. We conducted a comprehensive search on Scopus, IEEE Explore, PubMed, and Google Scholar (first 50 citations) using a systematic search strategy. The search spanned from January 2017 to July 2023, focusing on peer-reviewed studies implementing XAI methods in breast cancer datasets. Thirty studies met our inclusion criteria and were included in the analysis. The results revealed that SHapley Additive exPlanations (SHAP) is the top model-agnostic XAI technique in breast cancer research in terms of usage, explaining the model prediction results, diagnosis and classification of biomarkers, and prognosis and survival analysis. Additionally, the SHAP model primarily explained tree-based ensemble machine learning models. The most common reason is that SHAP is model agnostic, which makes it both popular and useful for explaining any model prediction. Additionally, it is relatively easy to implement effectively and completely suits performant models, such as tree-based models. Explainable AI improves the transparency, interpretability, fairness, and trustworthiness of AI-enabled health systems and medical devices and, ultimately, the quality of care and outcomes.

Seeing Like a Toolkit: How Toolkits Envision the Work of AI Ethics

Richmond Y. Wong, Michael A. Madaio, Nick Merrill

Numerous toolkits have been developed to support ethical AI development. However, toolkits, like all tools, encode assumptions in their design about what work should be done and how. In this paper, we conduct a qualitative analysis of 27 AI ethics toolkits to critically examine how the work of ethics is imagined and how it is supported by these toolkits. Specifically, we examine the discourses toolkits rely on when talking about ethical issues, who they imagine should do the work of ethics, and how they envision the work practices involved in addressing ethics. Among the toolkits, we identify a mismatch between the imagined work of ethics and the support the toolkits provide for doing that work. In particular, we identify a lack of guidance around how to navigate labor, organizational, and institutional power dynamics as they relate to performing ethical work. We use these omissions to chart future work for researchers and designers of AI ethics toolkits.

As ChatGPT et al. conquer the world, the optimal liability framework for AI systems remains an unsolved problem across the globe. In a much-anticipated move, the European Commission advanced two proposals outlining the European approach to AI liability in September 2022: a novel AI Liability Directive and a revision of the Product Liability Directive. They constitute the final cornerstone of EU AI regulation. Crucially, the liability proposals and the EU AI Act are inherently intertwined: the latter does not contain any individual rights of affected persons, and the former lack specific, substantive rules on AI development and deployment. Taken together, these acts may well trigger a Brussels Effect in AI regulation, with significant consequences for the US and beyond. This paper makes three novel contributions. First, it examines in detail the Commission proposals and shows that, while making steps in the right direction, they ultimately represent a half-hearted approach: if enacted as foreseen, AI liability in the EU will primarily rest on disclosure of evidence mechanisms and a set of narrowly defined presumptions concerning fault, defectiveness and causality. Hence, second, the article suggests amendments, which are collected in an Annex at the end of the paper. Third, based on an analysis of the key risks AI poses, the final part of the paper maps out a road for the future of AI liability and regulation, in the EU and beyond. This includes: a comprehensive framework for AI liability; provisions to support innovation; an extension to non-discrimination/algorithmic fairness, as well as explainable AI; and sustainability. I propose to jump-start sustainable AI regulation via sustainability impact assessments in the AI Act and sustainable design defects in the liability regime. In this way, the law may help spur not only fair AI and XAI, but potentially also sustainable AI (SAI).

Artificial Intelligence (AI) has created the single biggest technology revolution the world has ever seen. For the finance sector, it provides great opportunities to enhance customer experience, democratize financial services, ensure consumer protection and significantly improve risk management. While it is easier than ever to run state-of-the-art machine learning models, designing and implementing systems that support real-world finance applications have been challenging. In large part because they lack transparency and explainability which are important factors in establishing reliable technology and the research on this topic with a specific focus on applications in credit risk management. In this paper, we implement two advanced post-hoc model agnostic explainability techniques called Local Interpretable Model Agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP) to machine learning (ML)-based credit scoring models applied to the open-access data set offered by the US-based P2P Lending Platform, Lending Club. Specifically, we use LIME to explain instances locally and SHAP to get both local and global explanations. We discuss the results in detail and present multiple comparison scenarios by using various kernels available for explaining graphs generated using SHAP values. We also discuss the practical challenges associated with the implementation of these state-of-art eXplainabale AI (XAI) methods and document them for future reference. We have made an effort to document every technical aspect of this research, while at the same time providing a general summary of the conclusions.

Relating the Partial Dependence Plot and Permutation Feature Importance to the Data Generating Process

Christoph Molnar, Timo Freiesleben, Gunnar König +3

Scientists and practitioners increasingly rely on machine learning to model data and draw conclusions. Compared to statistical modeling approaches, machine learning makes fewer explicit assumptions about data structures, such as linearity. However, their model parameters usually cannot be easily related to the data generating process. To learn about the modeled relationships, partial dependence (PD) plots and permutation feature importance (PFI) are often used as interpretation methods. However, PD and PFI lack a theory that relates them to the data generating process. We formalize PD and PFI as statistical estimators of ground truth estimands rooted in the data generating process. We show that PD and PFI estimates deviate from this ground truth due to statistical biases, model variance and Monte Carlo approximation errors. To account for model variance in PD and PFI estimation, we propose the learner-PD and the learner-PFI based on model refits, and propose corrected variance and confidence interval estimators.

AI systems empowered by reinforcement learning (RL) algorithms harbor the immense potential to catalyze societal advancement, yet their deployment is often impeded by significant safety concerns. Particularly in safety-critical applications, researchers have raised concerns about unintended harms or unsafe behaviors of unaligned RL agents. The philosophy of safe reinforcement learning (SafeRL) is to align RL agents with harmless intentions and safe behavioral patterns. In SafeRL, agents learn to develop optimal policies by receiving feedback from the environment, while also fulfilling the requirement of minimizing the risk of unintended harm or unsafe behavior. However, due to the intricate nature of SafeRL algorithm implementation, combining methodologies across various domains presents a formidable challenge. This had led to an absence of a cohesive and efficacious learning framework within the contemporary SafeRL research milieu. In this work, we introduce a foundational framework designed to expedite SafeRL research endeavors. Our comprehensive framework encompasses an array of algorithms spanning different RL domains and places heavy emphasis on safety elements. Our efforts are to make the SafeRL-related research process more streamlined and efficient, therefore facilitating further research in AI safety. Our project is released at: https://github.com/PKU-Alignment/omnisafe.

Companies have considered adoption of various high-level artificial intelligence (AI) principles for responsible AI, but there is less clarity on how to implement these principles as organizational practices. This paper reviews the principles-to-practices gap. We outline five explanations for this gap ranging from a disciplinary divide to an overabundance of tools. In turn, we argue that an impact assessment framework which is broad, operationalizable, flexible, iterative, guided, and participatory is a promising approach to close the principles-to-practices gap. Finally, to help practitioners with applying these recommendations, we review a case study of AI's use in forest ecosystem restoration, demonstrating how an impact assessment framework can translate into effective and responsible AI practices.

Ethics and Responsible AI Deployment

Petar Radanliev, Omar Santos

As Artificial Intelligence (AI) becomes more prevalent, protecting personal privacy is a critical ethical issue that must be addressed. This article explores the need for ethical AI systems that safeguard individual privacy while complying with ethical standards. By taking a multidisciplinary approach, the research examines innovative algorithmic techniques such as differential privacy, homomorphic encryption, federated learning, international regulatory frameworks, and ethical guidelines. The study concludes that these algorithms effectively enhance privacy protection while balancing the utility of AI with the need to protect personal data. The article emphasises the importance of a comprehensive approach that combines technological innovation with ethical and regulatory strategies to harness the power of AI in a way that respects and protects individual privacy.

Explaining Black-Box Algorithms Using Probabilistic Contrastive Counterfactuals

Sainyam Galhotra, Romila Pradhan, Babak Salimi

There has been a recent resurgence of interest in explainable artificial intelligence (XAI) that aims to reduce the opaqueness of AI-based decision-making systems, allowing humans to scrutinize and trust them. Prior work in this context has focused on the attribution of responsibility for an algorithm's decisions to its inputs wherein responsibility is typically approached as a purely associational concept. In this paper, we propose a principled causality-based approach for explaining black-box decision-making systems that addresses limitations of existing methods in XAI. At the core of our framework lies probabilistic contrastive counterfactuals, a concept that can be traced back to philosophical, cognitive, and social foundations of theories on how humans generate and select explanations. We show how such counterfactuals can quantify the direct and indirect influences of a variable on decisions made by an algorithm, and provide actionable recourse for individuals negatively affected by the algorithm's decision. Unlike prior work, our system, LEWIS: (1)can compute provably effective explanations and recourse at local, global and contextual levels (2)is designed to work with users with varying levels of background knowledge of the underlying causal model and (3)makes no assumptions about the internals of an algorithmic system except for the availability of its input-output data. We empirically evaluate LEWIS on three real-world datasets and show that it generates human-understandable explanations that improve upon state-of-the-art approaches in XAI, including the popular LIME and SHAP. Experiments on synthetic data further demonstrate the correctness of LEWIS's explanations and the scalability of its recourse algorithm.

One of the most concrete measures to take towards meaningful AI accountability is to consequentially assess and report the systems' performance and impact. However, the practical nature of the "AI audit" ecosystem is muddled and imprecise, making it difficult to work through various concepts and map out the stakeholders involved in the practice. First, we taxonomize current AI audit practices as completed by regulators, law firms, civil society, journalism, academia, consulting agencies. Next, we assess the impact of audits done by stakeholders within each domain. We find that only a subset of AI audit studies translate to desired accountability outcomes. We thus assess and isolate practices necessary for effective AI audit results, articulating the observed connections between AI audit design, methodology and institutional context on its effectiveness as a meaningful mechanism for accountability.