Amir-Hossein Karimi

Ahmad-Reza Ehyaei, Amir-Hossein Karimi, Bernhard Schölkopf et al. · eth-zurich

Algorithmic recourse aims to disclose the inner workings of the black-box decision process in situations where decisions have significant consequences, by providing recommendations to empower beneficiaries to achieve a more favorable outcome. To ensure an effective remedy, suggested interventions must not only be low-cost but also robust and fair. This goal is accomplished by providing similar explanations to individuals who are alike. This study explores the concept of individual fairness and adversarial robustness in causal algorithmic recourse and addresses the challenge of achieving both. To resolve the challenges, we propose a new framework for defining adversarially robust recourse. The new setting views the protected feature as a pseudometric and demonstrates that individual fairness is a special case of adversarial robustness. Finally, we introduce the fair robust recourse problem to achieve both desirable properties and show how it can be satisfied both theoretically and empirically.

Amir-Hossein Karimi, Krikamol Muandet, Simon Kornblith et al. · eth-zurich

Being able to provide explanations for a model's decision has become a central requirement for the development, deployment, and adoption of machine learning models. However, we are yet to understand what explanation methods can and cannot do. How do upstream factors such as data, model prediction, hyperparameters, and random initialization influence downstream explanations? While previous work raised concerns that explanations (E) may have little relationship with the prediction (Y), there is a lack of conclusive study to quantify this relationship. Our work borrows tools from causal inference to systematically assay this relationship. More specifically, we study the relationship between E and Y by measuring the treatment effect when intervening on their causal ancestors, i.e., on hyperparameters and inputs used to generate saliency-based Es or Ys. Our results suggest that the relationships between E and Y is far from ideal. In fact, the gap between 'ideal' case only increase in higher-performing models -- models that are likely to be deployed. Our work is a promising first step towards providing a quantitative measure of the relationship between E and Y, which could also inform the future development of methods for E with a quantitative metric.

Saleh Afroogh, Seyd Ishtiaque Ahmed, Petra Ahrweiler et al. · cmu

This study provides a cross-disciplinary examination of Explainable Artificial Intelligence (XAI) approaches-focusing on deep neural networks (DNNs) and large language models (LLMs)-and identifies empirical and conceptual limitations in current XAI. We discuss critical symptoms that stem from deeper root causes (i.e., two paradoxes, two conceptual confusions, and five false assumptions). These fundamental problems within the current XAI research field reveal three insights: experimentally, XAI exhibits significant flaws; conceptually, it is paradoxical; and pragmatically, further attempts to reform the paradoxical XAI might exacerbate its confusion-demanding fundamental shifts and new research directions. To move beyond XAI's limitations, we propose a four-pronged synthesized paradigm shift toward reliable and certified AI development. These four components include: verification-focused Interactive AI (IAI) to establish scientific community protocols for certifying AI system performance rather than attempting post-hoc explanations, AI Epistemology for rigorous scientific foundations, User-Sensible AI to create context-aware systems tailored to specific user communities, and Model-Centered Interpretability for faithful technical analysis-together offering comprehensive post-XAI research directions.

Ahmad-Reza Ehyaei, Kiarash Mohammadi, Amir-Hossein Karimi et al. · eth-zurich

As responsible AI gains importance in machine learning algorithms, properties such as fairness, adversarial robustness, and causality have received considerable attention in recent years. However, despite their individual significance, there remains a critical gap in simultaneously exploring and integrating these properties. In this paper, we propose a novel approach that examines the relationship between individual fairness, adversarial robustness, and structural causal models in heterogeneous data spaces, particularly when dealing with discrete sensitive attributes. We use causal structural models and sensitive attributes to create a fair metric and apply it to measure semantic similarity among individuals. By introducing a novel causal adversarial perturbation and applying adversarial training, we create a new regularizer that combines individual fairness, causality, and robustness in the classifier. Our method is evaluated on both real-world and synthetic datasets, demonstrating its effectiveness in achieving an accurate classifier that simultaneously exhibits fairness, adversarial robustness, and causal awareness.

Hosna Oyarhoseini, Jimmy Lin, Amir-Hossein Karimi

Evaluation leaderboards such as LMArena play a central role in benchmarking large language models by aggregating pairwise human preferences into model rankings, yet the robustness of these rankings remains poorly understood. We present a unified perturbation framework for analyzing Bradley-Terry leaderboards under structured data modifications using influence-based approximations. Our framework studies three match-level perturbations -- Drop, Add, and Flip -- together with player removal, and evaluates their effects on top-k membership, global ranking consistency via Kendall's tau, and confidence-interval-based uncertainty. Across Chatbot Arena and six additional pairwise-comparison datasets, we show that modern leaderboards are non-robust across all three objectives: sub-1% targeted perturbations can change the top-ranked model, degrade Kendall's tau, and alter confidence intervals. Beyond robustness auditing, we show that the same influence scores enable efficient targeted perturbations, promoting or demoting specific models and reducing target-model uncertainty with fewer actions than previous manipulation and active-sampling baselines. By summarizing these effects with normalized dataset-level robustness scores, our framework provides a practical and helpful tool for auditing leaderboard stability and motivating more robust evaluation protocols.

Eugene Yu Ji, Igor Grossmann, Amir-Hossein Karimi

Generative AI research increasingly confronts a shared problem: systems must sustain yet govern their own generative activity when uncertainty is high, evidence is missing, or context is insufficient. This position paper argues that metacognition should become the scientific framework for bounded and effective self governance in generative AI, where output generation is properly evaluated together with the capacities through which generative systems navigate and regulate their own activity. We advance this position by showing that bounded and effective AI self-governance requires metacognitive alignment across computational, algorithmic, and ecological levels. At the computational level, metacognition specifies the meta-level functions a system is meant to serve, such as monitoring, evaluation, control, and adaptation. At the algorithmic level, these functions are realized through procedures such as elicitation, iteration, and modularization. At the ecological level, metacognitive signals become meaningful, actionable, and accountable within the interface, workflow, and accountability arrangements. Metacognition thus makes it possible to conceive generative AI as both capable and well-governed, rather than treating capability and governance as competing aims.

Amir-Hossein Karimi

The demand for Explainable AI (XAI) has triggered an explosion of methods, producing a landscape so fragmented that we now rely on surveys of surveys. Yet, fundamental challenges persist: conflicting metrics, failed sanity checks, and unresolved debates over robustness and fairness. The only consensus on how to achieve explainability is a lack of one. This has led many to point to the absence of a ground truth for defining ``the'' correct explanation as the main culprit. This position paper posits that the persistent discord in XAI arises not from an absent ground truth but from a ground truth that exists, albeit as an elusive and challenging target: the causal model that governs the relevant system. By reframing XAI queries about data, models, or decisions as causal inquiries, we prove the necessity and sufficiency of causal models for XAI. We contend that without this causal grounding, XAI remains unmoored. Ultimately, we encourage the community to converge around advanced concept and causal discovery to escape this entrenched uncertainty.

Samuel G. B. Johnson, Amir-Hossein Karimi, Yoshua Bengio et al.

Although AI has become increasingly smart, its wisdom has not kept pace. In this article, we examine what is known about human wisdom and sketch a vision of its AI counterpart. We analyze human wisdom as a set of strategies for solving intractable problems-those outside the scope of analytic techniques-including both object-level strategies like heuristics [for managing problems] and metacognitive strategies like intellectual humility, perspective-taking, or context-adaptability [for managing object-level strategies]. We argue that AI systems particularly struggle with metacognition; improved metacognition would lead to AI more robust to novel environments, explainable to users, cooperative with others, and safer in risking fewer misaligned goals with human users. We discuss how wise AI might be benchmarked, trained, and implemented.

Hamdi Altaheri, Fakhri Karray, Md. Milon Islam et al.

Foundation models (FMs), powered by self-supervised learning (SSL), have redefined the capabilities of artificial intelligence, demonstrating exceptional performance in domains like natural language processing and computer vision. These advances present a transformative opportunity for brain signal analysis. Unlike traditional supervised learning, which is limited by the scarcity of labeled neural data, SSL offers a promising solution by enabling models to learn meaningful representations from unlabeled data. This is particularly valuable in addressing the unique challenges of brain signals, including high noise levels, inter-subject variability, and low signal-to-noise ratios. This survey systematically reviews the emerging field of bridging brain signals with foundation models through the innovative application of SSL. It explores key SSL techniques, the development of brain-specific foundation models, their adaptation to downstream tasks, and the integration of brain signals with other modalities in multimodal SSL frameworks. The review also covers commonly used evaluation metrics and benchmark datasets that support comparative analysis. Finally, it highlights key challenges and outlines future research directions. This work aims to provide researchers with a structured understanding of this rapidly evolving field and a roadmap for developing generalizable brain foundation models powered by self-supervision.

Gautam Machiraju, Alexander Derry, Arjun Desai et al.

Feature attribution, the ability to localize regions of the input data that are relevant for classification, is an important capability for ML models in scientific and biomedical domains. Current methods for feature attribution, which rely on "explaining" the predictions of end-to-end classifiers, suffer from imprecise feature localization and are inadequate for use with small sample sizes and high-dimensional datasets due to computational challenges. We introduce prospector heads, an efficient and interpretable alternative to explanation-based attribution methods that can be applied to any encoder and any data modality. Prospector heads generalize across modalities through experiments on sequences (text), images (pathology), and graphs (protein structures), outperforming baseline attribution methods by up to 26.3 points in mean localization AUPRC. We also demonstrate how prospector heads enable improved interpretation and discovery of class-specific patterns in input data. Through their high performance, flexibility, and generalizability, prospectors provide a framework for improving trust and transparency for ML models in complex domains.

Zahra Khotanlou, Kate Larson, Amir-Hossein Karimi

Decision makers are increasingly relying on machine learning in sensitive situations. In such settings, algorithmic recourse aims to provide individuals with actionable and minimally costly steps to reverse unfavorable AI-driven decisions. While existing research predominantly focuses on single-individual (i.e., seeker) and single-model (i.e., provider) scenarios, real-world applications often involve multiple interacting stakeholders. Optimizing outcomes for seekers under an individual welfare approach overlooks the inherently multi-agent nature of real-world systems, where individuals interact and compete for limited resources. To address this, we introduce a novel framework for multi-agent algorithmic recourse that accounts for multiple recourse seekers and recourse providers. We model this many-to-many interaction as a capacitated weighted bipartite matching problem, where matches are guided by both recourse cost and provider capacity. Edge weights, reflecting recourse costs, are optimized for social welfare while quantifying the welfare gap between individual welfare and this collectively feasible outcome. We propose a three-layer optimization framework: (1) basic capacitated matching, (2) optimal capacity redistribution to minimize the welfare gap, and (3) cost-aware optimization balancing welfare maximization with capacity adjustment costs. Experimental validation on synthetic and real-world datasets demonstrates that our framework enables the many-to-many algorithmic recourse to achieve near-optimal welfare with minimum modification in system settings. This work extends algorithmic recourse from individual recommendations to system-level design, providing a tractable path toward higher social welfare while maintaining individual actionability.

Ricardo Dominguez-Olmedo, Amir-Hossein Karimi, Bernhard Schölkopf

Algorithmic recourse seeks to provide actionable recommendations for individuals to overcome unfavorable classification outcomes from automated decision-making systems. Recourse recommendations should ideally be robust to reasonably small uncertainty in the features of the individual seeking recourse. In this work, we formulate the adversarially robust recourse problem and show that recourse methods that offer minimally costly recourse fail to be robust. We then present methods for generating adversarially robust recourse for linear and for differentiable classifiers. Finally, we show that regularizing the decision-making classifier to behave locally linearly and to rely more strongly on actionable features facilitates the existence of adversarially robust recourse.

Julius von Kügelgen, Amir-Hossein Karimi, Umang Bhatt et al.

Algorithmic fairness is typically studied from the perspective of predictions. Instead, here we investigate fairness from the perspective of recourse actions suggested to individuals to remedy an unfavourable classification. We propose two new fairness criteria at the group and individual level, which -- unlike prior work on equalising the average group-wise distance from the decision boundary -- explicitly account for causal relationships between features, thereby capturing downstream effects of recourse actions performed in the physical world. We explore how our criteria relate to others, such as counterfactual fairness, and show that fairness of recourse is complementary to fairness of prediction. We study theoretically and empirically how to enforce fair causal recourse by altering the classifier and perform a case study on the Adult dataset. Finally, we discuss whether fairness violations in the data generating process revealed by our criteria may be better addressed by societal interventions as opposed to constraints on the classifier.

Kiarash Mohammadi, Amir-Hossein Karimi, Gilles Barthe et al.

Counterfactual explanations (CFE) are being widely used to explain algorithmic decisions, especially in consequential decision-making contexts (e.g., loan approval or pretrial bail). In this context, CFEs aim to provide individuals affected by an algorithmic decision with the most similar individual (i.e., nearest individual) with a different outcome. However, while an increasing number of works propose algorithms to compute CFEs, such approaches either lack in optimality of distance (i.e., they do not return the nearest individual) and perfect coverage (i.e., they do not provide a CFE for all individuals); or they cannot handle complex models, such as neural networks. In this work, we provide a framework based on Mixed-Integer Programming (MIP) to compute nearest counterfactual explanations with provable guarantees and with runtimes comparable to gradient-based approaches. Our experiments on the Adult, COMPAS, and Credit datasets show that, in contrast with previous methods, our approach allows for efficiently computing diverse CFEs with both distance guarantees and perfect coverage.

Amir-Hossein Karimi, Gilles Barthe, Bernhard Schölkopf et al.

Machine learning is increasingly used to inform decision-making in sensitive situations where decisions have consequential effects on individuals' lives. In these settings, in addition to requiring models to be accurate and robust, socially relevant values such as fairness, privacy, accountability, and explainability play an important role for the adoption and impact of said technologies. In this work, we focus on algorithmic recourse, which is concerned with providing explanations and recommendations to individuals who are unfavourably treated by automated decision-making systems. We first perform an extensive literature review, and align the efforts of many authors by presenting unified definitions, formulations, and solutions to recourse. Then, we provide an overview of the prospective research directions towards which the community may engage, challenging existing assumptions and making explicit connections to other ethical challenges such as security, privacy, and fairness.

Amir-Hossein Karimi, Julius von Kügelgen, Bernhard Schölkopf et al.

Recent work has discussed the limitations of counterfactual explanations to recommend actions for algorithmic recourse, and argued for the need of taking causal relationships between features into consideration. Unfortunately, in practice, the true underlying structural causal model is generally unknown. In this work, we first show that it is impossible to guarantee recourse without access to the true structural equations. To address this limitation, we propose two probabilistic approaches to select optimal actions that achieve recourse with high probability given limited causal knowledge (e.g., only the causal graph). The first captures uncertainty over structural equations under additive Gaussian noise, and uses Bayesian model averaging to estimate the counterfactual distribution. The second removes any assumptions on the structural equations by instead computing the average effect of recourse actions on individuals similar to the person who seeks recourse, leading to a novel subpopulation-based interventional notion of recourse. We then derive a gradient-based procedure for selecting optimal recourse actions, and empirically show that the proposed approaches lead to more reliable recommendations under imperfect causal knowledge than non-probabilistic baselines.

Amir-Hossein Karimi, Bernhard Schölkopf, Isabel Valera

As machine learning is increasingly used to inform consequential decision-making (e.g., pre-trial bail and loan approval), it becomes important to explain how the system arrived at its decision, and also suggest actions to achieve a favorable decision. Counterfactual explanations -- "how the world would have (had) to be different for a desirable outcome to occur" -- aim to satisfy these criteria. Existing works have primarily focused on designing algorithms to obtain counterfactual explanations for a wide range of settings. However, one of the main objectives of "explanations as a means to help a data-subject act rather than merely understand" has been overlooked. In layman's terms, counterfactual explanations inform an individual where they need to get to, but not how to get there. In this work, we rely on causal reasoning to caution against the use of counterfactual explanations as a recommendable set of actions for recourse. Instead, we propose a shift of paradigm from recourse via nearest counterfactual explanations to recourse through minimal interventions, moving the focus from explanations to recommendations. Finally, we provide the reader with an extensive discussion on how to realistically achieve recourse beyond structural interventions.

Amir-Hossein Karimi, Gilles Barthe, Borja Balle et al.

Predictive models are being increasingly used to support consequential decision making at the individual level in contexts such as pretrial bail and loan approval. As a result, there is increasing social and legal pressure to provide explanations that help the affected individuals not only to understand why a prediction was output, but also how to act to obtain a desired outcome. To this end, several works have proposed optimization-based methods to generate nearest counterfactual explanations. However, these methods are often restricted to a particular subset of models (e.g., decision trees or linear models) and differentiable distance functions. In contrast, we build on standard theory and tools from formal verification and propose a novel algorithm that solves a sequence of satisfiability problems, where both the distance function (objective) and predictive model (constraints) are represented as logic formulae. As shown by our experiments on real-world data, our algorithm is: i) model-agnostic ({non-}linear, {non-}differentiable, {non-}convex); ii) data-type-agnostic (heterogeneous features); iii) distance-agnostic ($\ell_0, \ell_1, \ell_\infty$, and combinations thereof); iv) able to generate plausible and diverse counterfactuals for any sample (i.e., 100% coverage); and v) at provably optimal distances.

Ershad Banijamali, Amir-Hossein Karimi, Ali Ghodsi

We consider the problem of sufficient dimensionality reduction (SDR), where the high-dimensional observation is transformed to a low-dimensional sub-space in which the information of the observations regarding the label variable is preserved. We propose DVSDR, a deep variational approach for sufficient dimensionality reduction. The deep structure in our model has a bottleneck that represent the low-dimensional embedding of the data. We explain the SDR problem using graphical models and use the framework of variational autoencoders to maximize the lower bound of the log-likelihood of the joint distribution of the observation and label. We show that such a maximization problem can be interpreted as solving the SDR problem. DVSDR can be easily adopted to semi-supervised learning setting. In our experiment we show that DVSDR performs competitively on classification tasks while being able to generate novel data samples.

Amir-Hossein Karimi, Alexander Wong, Ali Ghodsi

Supervised dimensionality reduction strategies have been of great interest. However, current supervised dimensionality reduction approaches are difficult to scale for situations characterized by large datasets given the high computational complexities associated with such methods. While stochastic approximation strategies have been explored for unsupervised dimensionality reduction to tackle this challenge, such approaches are not well-suited for accelerating computational speed for supervised dimensionality reduction. Motivated to tackle this challenge, in this study we explore a novel direction of directly learning optimal class-aware embeddings in a supervised manner via the notion of supervised random projections (SRP). The key idea behind SRP is that, rather than performing spectral decomposition (or approximations thereof) which are computationally prohibitive for large-scale data, we instead perform a direct decomposition by leveraging kernel approximation theory and the symmetry of the Hilbert-Schmidt Independence Criterion (HSIC) measure of dependence between the embedded data and the labels. Experimental results on five different synthetic and real-world datasets demonstrate that the proposed SRP strategy for class-aware embedding learning can be very promising in producing embeddings that are highly competitive with existing supervised dimensionality reduction methods (e.g., SPCA and KSPCA) while achieving 1-2 orders of magnitude better computational performance. As such, such an efficient approach to learning embeddings for dimensionality reduction can be a powerful tool for large-scale data analysis and visualization.

Ershad Banijamali, Amir-Hossein Karimi, Alexander Wong et al.

The problem of feature disentanglement has been explored in the literature, for the purpose of image and video processing and text analysis. State-of-the-art methods for disentangling feature representations rely on the presence of many labeled samples. In this work, we present a novel method for disentangling factors of variation in data-scarce regimes. Specifically, we explore the application of feature disentangling for the problem of supervised classification in a setting where few labeled samples exist, and there are no unlabeled samples for use in unsupervised training. Instead, a similar datasets exists which shares at least one direction of variation with the sample-constrained datasets. We train our model end-to-end using the framework of variational autoencoders and are able to experimentally demonstrate that using an auxiliary dataset with similar variation factors contribute positively to classification performance, yielding competitive results with the state-of-the-art in unsupervised learning.

Amir-Hossein Karimi

Kernel-based learning algorithms are widely used in machine learning for problems that make use of the similarity between object pairs. Such algorithms first embed all data points into an alternative space, where the inner product between object pairs specifies their distance in the embedding space. Applying kernel methods to partially labeled datasets is a classical challenge in this regard, requiring that the distances between unlabeled pairs must somehow be learnt using the labeled data. In this independent study, I will summarize the work of G. Lanckriet et al.'s work on "Learning the Kernel Matrix with Semidefinite Programming" used in support vector machines (SVM) algorithms for the transduction problem. Throughout the report, I have provide alternative explanations / derivations / analysis related to this work which is designed to ease the understanding of the original article.

Alexander Miller, Adam Fisch, Jesse Dodge et al.

Directly reading documents and being able to answer questions from them is an unsolved challenge. To avoid its inherent difficulty, question answering (QA) has been directed towards using Knowledge Bases (KBs) instead, which has proven effective. Unfortunately KBs often suffer from being too restrictive, as the schema cannot support certain types of answers, and too sparse, e.g. Wikipedia contains much more information than Freebase. In this work we introduce a new method, Key-Value Memory Networks, that makes reading documents more viable by utilizing different encodings in the addressing and output stages of the memory read operation. To compare using KBs, information extraction or Wikipedia documents directly in a single framework we construct an analysis tool, WikiMovies, a QA dataset that contains raw text alongside a preprocessed KB, in the domain of movies. Our method reduces the gap between all three settings. It also achieves state-of-the-art results on the existing WikiQA benchmark.