A faithful and interpretable explanation of an AI model's behavior and internal structure is a high-level explanation that is human-intelligible but also consistent with the known, but often opaque low-level causal details of the model. We argue that the theory of causal abstraction provides the mathematical foundations for the desired kinds of model explanations. In causal abstraction analysis, we use interventions on model-internal states to rigorously assess whether an interpretable high-level causal model is a faithful description of an AI model. Our contributions in this area are: (1) We generalize causal abstraction to cyclic causal structures and typed high-level variables.

Over the past few years, the use of machine learning models has emerged as a generic and powerful means for prediction purposes. At the same time, there is a growing demand for interpretability of prediction models. To determine which features of a dataset are important to predict a target variable $Y$, a Feature Importance (FI) method can be used. By quantifying how important each feature is for predicting $Y$, irrelevant features can be identified and removed, which could increase the speed and accuracy of a model, and moreover, important features can be discovered, which could lead to valuable insights. A major problem with evaluating FI methods, is that the ground truth FI is often unknown. As a consequence, existing FI methods do not give the exact correct FI values. This is one of the many reasons why it can be hard to properly interpret the results of an FI method. Motivated by this, we introduce a new global approach named the Berkelmans-Pries FI method, which is based on a combination of Shapley values and the Berkelmans-Pries dependency function. We prove that our method has many useful properties, and accurately predicts the correct FI values for several cases where the ground truth FI can be derived in an exact manner. We experimentally show for a large collection of FI methods (468) that existing methods do not have the same useful properties. This shows that the Berkelmans-Pries FI method is a highly valuable tool for analyzing datasets with complex interdependencies.

Realizing when a model is right for a wrong reason is not trivial and requires a significant effort by model developers. In some cases an input salience method, which highlights the most important parts of the input, may reveal problematic reasoning. But scrutinizing highlights over many data instances is tedious and often infeasible. Furthermore, analyzing examples in isolation does not reveal general patterns in the data or in the model's behavior. In this paper we aim to address these issues and go from understanding single examples to understanding entire datasets and models. The methodology we propose is based on aggregated salience maps, to which we apply clustering, nearest neighbor search and visualizations. Using this methodology we address multiple distinct but common model developer needs by showing how problematic data and model behavior can be identified and explained -- a necessary first step for improving the model.

In the field of reinforcement learning, because of the high cost and risk of policy training in the real world, policies are trained in a simulation environment and transferred to the corresponding real-world environment. However, the simulation environment does not perfectly mimic the real-world environment, lead to model misspecification. Multiple studies report significant deterioration of policy performance in a real-world environment. In this study, we focus on scenarios involving a simulation environment with uncertainty parameters and the set of their possible values, called the uncertainty parameter set. The aim is to optimize the worst-case performance on the uncertainty parameter set to guarantee the performance in the corresponding real-world environment. To obtain a policy for the optimization, we propose an off-policy actor-critic approach called the Max-Min Twin Delayed Deep Deterministic Policy Gradient algorithm (M2TD3), which solves a max-min optimization problem using a simultaneous gradient ascent descent approach. Experiments in multi-joint dynamics with contact (MuJoCo) environments show that the proposed method exhibited a worst-case performance superior to several baseline approaches.

Central Banks Digital Currency (CBDC) is a digital currency or virtual asset backed by a central bank. According to the Atlantic Council CBDC Tracker, over 100 countries are currently exploring the CBDC opportunity and between 2020-2022 the central banks of many countries launched CBDC, while others started the pilot or research project. The Bahamas was the first country to launch CBDC which they named Sand Dollar. Nigeria launched e-Naira and Jamaica introduced JAM-DEX. China has been trialing Digital Yuan or e-CNY for a couple of years, Ghana is conducting eCedi pilot and India has started a pilot of both wholesale and retail CBDC Digital Rupee (e-Rupee) in the last quarter of 2022.

This dissertation proposes a framework of user-centered security in Natural Language Processing (NLP), and demonstrates how it can improve the accessibility of related research. Accordingly, it focuses on two security domains within NLP with great public interest. First, that of author profiling, which can be employed to compromise online privacy through invasive inferences. Without access and detailed insight into these models' predictions, there is no reasonable heuristic by which Internet users might defend themselves from such inferences. Secondly, that of cyberbullying detection, which by default presupposes a centralized implementation; i.e., content moderation across social platforms. As access to appropriate data is restricted, and the nature of the task rapidly evolves (both through lexical variation, and cultural shifts), the effectiveness of its classifiers is greatly diminished and thereby often misrepresented. Under the proposed framework, we predominantly investigate the use of adversarial attacks on language; i.e., changing a given input (generating adversarial samples) such that a given model does not function as intended. These attacks form a common thread between our user-centered security problems; they are highly relevant for privacy-preserving obfuscation methods against author profiling, and adversarial samples might also prove useful to assess the influence of lexical variation and augmentation on cyberbullying detection.

Despite strong performance in many sequence-to-sequence tasks, autoregressive models trained with maximum likelihood estimation suffer from exposure bias, i.e. the discrepancy between the ground-truth prefixes used during training and the model-generated prefixes used at inference time. Scheduled sampling is a simple and empirically successful approach which addresses this issue by incorporating model-generated prefixes into training. However, it has been argued that it is an inconsistent training objective leading to models ignoring the prefixes altogether. In this paper, we conduct systematic experiments and find that scheduled sampling, while it ameliorates exposure bias by increasing model reliance on the input sequence, worsens performance when the prefix at inference time is correct, a form of catastrophic forgetting. We propose to use Elastic Weight Consolidation to better balance mitigating exposure bias with retaining performance. Experiments on four IWSLT'14 and WMT'14 translation datasets demonstrate that our approach alleviates catastrophic forgetting and significantly outperforms maximum likelihood estimation and scheduled sampling baselines.

Identifying whether a given sample is an outlier or not is an important issue in various real-world domains. This study aims to solve the unsupervised outlier detection problem where training data contain outliers, but any label information about inliers and outliers is not given. We propose a powerful and efficient learning framework to identify outliers in a training data set using deep neural networks. We start with a new observation called the inlier-memorization (IM) effect. When we train a deep generative model with data contaminated with outliers, the model first memorizes inliers before outliers. Exploiting this finding, we develop a new method called the outlier detection via the IM effect (ODIM). The ODIM only requires a few updates; thus, it is computationally efficient, tens of times faster than other deep-learning-based algorithms. Also, the ODIM filters out outliers successfully, regardless of the types of data, such as tabular, image, and sequential. We empirically demonstrate the superiority and efficiency of the ODIM by analyzing 20 data sets.

Recent years have seen the development of many open-source ML fairness toolkits aimed at helping ML practitioners assess and address unfairness in their systems. However, there has been little research investigating how ML practitioners actually use these toolkits in practice. In this paper, we conducted the first in-depth empirical exploration of how industry practitioners (try to) work with existing fairness toolkits. In particular, we conducted think-aloud interviews to understand how participants learn about and use fairness toolkits, and explored the generality of our findings through an anonymous online survey. We identified several opportunities for fairness toolkits to better address practitioner needs and scaffold them in using toolkits effectively and responsibly. Based on these findings, we highlight implications for the design of future open-source fairness toolkits that can support practitioners in better contextualizing, communicating, and collaborating around ML fairness efforts.

In December 2019, a novel coronavirus (COVID-19) spread so quickly around the world that many countries had to set mandatory face mask rules in public areas to reduce the transmission of the virus. To monitor public adherence, researchers aimed to rapidly develop efficient systems that can detect faces with masks automatically. However, the lack of representative and novel datasets proved to be the biggest challenge. Early attempts to collect face mask datasets did not account for potential race, gender, and age biases. Therefore, the resulting models show inherent biases toward specific race groups, such as Asian or Caucasian. In this work, we present a novel face mask detection dataset that contains images posted on Twitter during the pandemic from around the world. Unlike previous datasets, the proposed Bias-Aware Face Mask Detection (BAFMD) dataset contains more images from underrepresented race and age groups to mitigate the problem for the face mask detection task. We perform experiments to investigate potential biases in widely used face mask detection datasets and illustrate that the BAFMD dataset yields models with better performance and generalization ability. The dataset is publicly available at https://github.com/Alpkant/BAFMD.