We study a specific type of SCM, called a Dynamic Structural Causal Model (DSCM), whose endogenous variables represent functions of time, which is possibly cyclic and allows for latent confounding. As a motivating use-case, we show that certain systems of Stochastic Differential Equations (SDEs) can be appropriately represented with DSCMs. An immediate consequence of this construction is a graphical Markov property for systems of SDEs. We define a time-splitting operation, allowing us to analyse the concept of local independence (a notion of continuous-time Granger (non-)causality). We also define a subsampling operation, which returns a discrete-time DSCM, and which can be used for mathematical analysis of subsampled time-series. We give suggestions how DSCMs can be used for identification of the causal effect of time-dependent interventions, and how existing constraint-based causal discovery algorithms can be applied to time-series data.

There is an extensive literature about online controlled experiments, both on the statistical methods available to analyze experiment results as well as on the infrastructure built by several large scale Internet companies but also on the organizational challenges of embracing online experiments to inform product development. At Booking.com we have been conducting evidenced based product development using online experiments for more than ten years. Our methods and infrastructure were designed from their inception to reflect Booking.com culture, that is, with democratization and decentralization of experimentation and decision making in mind. In this paper we explain how building a central repository of successes and failures to allow for knowledge sharing, having a generic and extensible code library which enforces a loose coupling between experimentation and business logic, monitoring closely and transparently the quality and the reliability of the data gathering pipelines to build trust in the experimentation infrastructure, and putting in place safeguards to enable anyone to have end to end ownership of their experiments have allowed such a large organization as Booking.com to truly and successfully democratize experimentation.

A fundamental objective in intelligent robotics is to move towards lifelong learning robot that can learn and adapt to unseen scenarios over time. However, continually learning new tasks would introduce catastrophic forgetting problems due to data distribution shifts. To mitigate this, we store a subset of data from previous tasks and utilize it in two manners: leveraging experience replay to retain learned skills and applying a novel Retrieval-based Local Adaptation technique to restore relevant knowledge. Since a lifelong learning robot must operate in task-free scenarios, where task IDs and even boundaries are not available, our method performs effectively without relying on such information. We also incorporate a selective weighting mechanism to focus on the most "forgotten" skill segment, ensuring effective knowledge restoration. Experimental results across diverse manipulation tasks demonstrate that our framework provides a scalable paradigm for lifelong learning, enhancing robot performance in open-ended, task-free scenarios.

Causal Machine Learning has emerged as a powerful tool for flexibly estimating causal effects from observational data in both industry and academia. However, causal inference from observational data relies on untestable assumptions about the data-generating process, such as the absence of unobserved confounders. When these assumptions are violated, causal effect estimates may become biased, undermining the validity of research findings. In these contexts, sensitivity analysis plays a crucial role, by enabling data scientists to assess the robustness of their findings to plausible violations of unconfoundedness. This paper introduces sensitivity analysis and demonstrates its practical relevance through a (simulated) data example based on a use case at this http URL. We focus our presentation on a recently proposed method by Chernozhukov et al. (2023), which derives general non-parametric bounds on biases due to omitted variables, and is fully compatible with (though not limited to) modern inferential tools of Causal Machine Learning. By presenting this use case, we aim to raise awareness of sensitivity analysis and highlight its importance in real-world scenarios.

Unbiased learning-to-rank (ULTR) is a well-established framework for learning from user clicks, which are often biased by the ranker collecting the data. While theoretically justified and extensively tested in simulation, ULTR techniques lack empirical validation, especially on modern search engines. The Baidu-ULTR dataset released for the WSDM Cup 2023, collected from Baidu's search engine, offers a rare opportunity to assess the real-world performance of prominent ULTR techniques. Despite multiple submissions during the WSDM Cup 2023 and the subsequent NTCIR ULTRE-2 task, it remains unclear whether the observed improvements stem from applying ULTR or other learning techniques. In this work, we revisit and extend the available experiments on the Baidu-ULTR dataset. We find that standard unbiased learning-to-rank techniques robustly improve click predictions but struggle to consistently improve ranking performance, especially considering the stark differences obtained by choice of ranking loss and query-document features. Our experiments reveal that gains in click prediction do not necessarily translate to enhanced ranking performance on expert relevance annotations, implying that conclusions strongly depend on how success is measured in this benchmark.

This paper studies grading algorithms for randomized exams. In a randomized exam, each student is asked a small number of random questions from a large question bank. The predominant grading rule is simple averaging, i.e., calculating grades by averaging scores on the questions each student is asked, which is fair ex-ante, over the randomized questions, but not fair ex-post, on the realized questions. The fair grading problem is to estimate the average grade of each student on the full question bank. The maximum-likelihood estimator for the Bradley-Terry-Luce model on the bipartite student-question graph is shown to be consistent with high probability when the number of questions asked to each student is at least the cubed-logarithm of the number of students. In an empirical study on exam data and in simulations, our algorithm based on the maximum-likelihood estimator significantly outperforms simple averaging in prediction accuracy and ex-post fairness even with a small class and exam size.

Uplift modeling is a key technique for promotion optimization in recommender systems, but standard methods typically fail to account for interference, where treating one item affects the outcomes of others. This violation of the Stable Unit Treatment Value Assumption (SUTVA) leads to suboptimal policies in real-world marketplaces. Recent developments in interference-aware estimators such as Additive Inverse Propensity Weighting (AddIPW) have not found their way into the uplift modeling literature yet, and optimising policies using these estimators is not well-established. This paper proposes a practical methodology to bridge this gap. We use the AddIPW estimator as a differentiable learning objective suitable for gradient-based optimization. We demonstrate how this framework can be integrated with proven response transformation techniques to directly optimize for economic outcomes like incremental profit. Through simulations, we show that our approach significantly outperforms interference-naive methods, especially as interference effects grow. Furthermore, we find that adapting profit-centric uplift strategies within our framework can yield superior performance in identifying the highest-impact interventions, offering a practical path toward more profitable incentive personalization.

Reasoning about fairness through correlation-based notions is rife with pitfalls. The 1973 University of California, Berkeley graduate school admissions case from Bickel et. al. (1975) is a classic example of one such pitfall, namely Simpson's paradox. The discrepancy in admission rates among males and female applicants, in the aggregate data over all departments, vanishes when admission rates per department are examined. We reason about the Berkeley graduate school admissions case through a causal lens. In the process, we introduce a statistical test for causal hypothesis testing based on Pearl's instrumental-variable inequalities (Pearl 1995). We compare different causal notions of fairness that are based on graphical, counterfactual and interventional queries on the causal model, and develop statistical tests for these notions that use only observational data. We study the logical relations between notions, and show that while notions may not be equivalent, their corresponding statistical tests coincide for the case at hand. We believe that a thorough case-based causal analysis helps develop a more principled understanding of both causal hypothesis testing and fairness.

Controlling the length of generated text can be crucial in various text-generation tasks, including summarization. Existing methods often require complex model alterations, limiting compatibility with pre-trained models. We address these limitations by developing a simple approach for controlling the length of automatic text summaries by increasing the importance of correctly predicting the EOS token in the cross-entropy loss computation. The proposed methodology is agnostic to architecture and decoding algorithms and orthogonal to other inference-time techniques to control the generation length. We tested it with encoder-decoder and modern GPT-style LLMs, and show that this method can control generation length, often without affecting the quality of the summary.

We study the generalization capabilities of Group Convolutional Neural Networks (GCNNs) with ReLU activation function by deriving upper and lower bounds for their Vapnik-Chervonenkis (VC) dimension. Specifically, we analyze how factors such as the number of layers, weights, and input dimension affect the VC dimension. We further compare the derived bounds to those known for other types of neural networks. Our findings extend previous results on the VC dimension of continuous GCNNs with two layers, thereby providing new insights into the generalization properties of GCNNs, particularly regarding the dependence on the input resolution of the data.

In this paper, we investigate properties and limitations of invariance learned by neural networks from the data compared to the genuine invariance achieved through invariant weight-tying. To do so, we adopt a group theoretical perspective and analyze invariance learning in neural networks without weight-tying constraints. We demonstrate that even when a network learns to correctly classify samples on a group orbit, the underlying decision-making in such a model does not attain genuine invariance. Instead, learned invariance is strongly conditioned on the input data, rendering it unreliable if the input distribution shifts. We next demonstrate how to guide invariance learning toward genuine invariance by regularizing the invariance of a model at the training. To this end, we propose several metrics to quantify learned invariance: (i) predictive distribution invariance, (ii) logit invariance, and (iii) saliency invariance similarity. We show that the invariance learned with the invariance error regularization closely reassembles the genuine invariance of weight-tying models and reliably holds even under a severe input distribution shift. Closer analysis of the learned invariance also reveals the spectral decay phenomenon, when a network chooses to achieve the invariance to a specific transformation group by reducing the sensitivity to any input perturbation.

We propose an adaptation of the curriculum training framework, applicable to state-of-the-art meta learning techniques for few-shot classification. Curriculum-based training popularly attempts to mimic human learning by progressively increasing the training complexity to enable incremental concept learning. As the meta-learner's goal is learning how to learn from as few samples as possible, the exact number of those samples (i.e. the size of the support set) arises as a natural proxy of a given task's difficulty. We define a simple yet novel curriculum schedule that begins with a larger support size and progressively reduces it throughout training to eventually match the desired shot-size of the test setup. This proposed method boosts the learning efficiency as well as the generalization capability. Our experiments with the MAML algorithm on two few-shot image classification tasks show significant gains with the curriculum training framework. Ablation studies corroborate the independence of our proposed method from the model architecture as well as the meta-learning hyperparameters

Uplift modeling is a collection of machine learning techniques for estimating causal effects of a treatment at the individual or subgroup levels. Over the last years, causality and uplift modeling have become key trends in personalization at online e-commerce platforms, enabling the selection of the best treatment for each user in order to maximize the target business metric. Uplift modeling can be particularly useful for personalized promotional campaigns, where the potential benefit caused by a promotion needs to be weighed against the potential costs. In this tutorial we will cover basic concepts of causality and introduce the audience to state-of-the-art techniques in uplift modeling. We will discuss the advantages and the limitations of different approaches and dive into the unique setup of constrained uplift modeling. Finally, we will present real-life applications and discuss challenges in implementing these models in production.

Promotions play a crucial role in e-commerce platforms, and various cost structures are employed to drive user engagement. This paper focuses on promotions with response-dependent costs, where expenses are incurred only when a purchase is made. Such promotions include discounts and coupons. While existing uplift model approaches aim to address this challenge, these approaches often necessitate training multiple models, like meta-learners, or encounter complications when estimating profit due to zero-inflated values stemming from non-converted individuals with zero cost and profit. To address these challenges, we introduce Incremental Profit per Conversion (IPC), a novel uplift measure of promotional campaigns' efficiency in unit economics. Through a proposed response transformation, we demonstrate that IPC requires only converted data, its propensity, and a single model to be estimated. As a result, IPC resolves the issues mentioned above while mitigating the noise typically associated with the class imbalance in conversion datasets and biases arising from the many-to-one mapping between search and purchase data. Lastly, we validate the efficacy of our approach by presenting results obtained from a synthetic simulation of a discount coupon campaign.

Fraud detection and prevention play an important part in ensuring the sustained operation of any e-commerce business. Machine learning (ML) often plays an important role in these anti-fraud operations, but the organizational context in which these ML models operate cannot be ignored. In this paper, we take an organization-centric view on the topic of fraud detection by formulating an operational model of the anti-fraud departments in e-commerce organizations. We derive 6 research topics and 12 practical challenges for fraud detection from this operational model. We summarize the state of the literature for each research topic, discuss potential solutions to the practical challenges, and identify 22 open research challenges.

Privacy in Social Network Sites (SNSs) like Facebook or Instagram is closely related to people's self-disclosure decisions and their ability to foresee the consequences of sharing personal information with large and diverse audiences. Nonetheless, online privacy decisions are often based on spurious risk judgements that make people liable to reveal sensitive data to untrusted recipients and become victims of social engineering attacks. Artificial Intelligence (AI) in combination with persuasive mechanisms like nudging is a promising approach for promoting preventative privacy behaviour among the users of SNSs. Nevertheless, combining behavioural interventions with high levels of personalization can be a potential threat to people's agency and autonomy even when applied to the design of social engineering countermeasures. This paper elaborates on the ethical challenges that nudging mechanisms can introduce to the development of AI-based countermeasures, particularly to those addressing unsafe self-disclosure practices in SNSs. Overall, it endorses the elaboration of personalized risk awareness solutions as i) an ethical approach to counteract social engineering, and ii) as an effective means for promoting reflective privacy decisions.

A well-known problem when learning from user clicks are inherent biases prevalent in the data, such as position or trust bias. Click models are a common method for extracting information from user clicks, such as document relevance in web search, or to estimate click biases for downstream applications such as counterfactual learning-to-rank, ad placement, or fair ranking. Recent work shows that the current evaluation practices in the community fail to guarantee that a well-performing click model generalizes well to downstream tasks in which the ranking distribution differs from the training distribution, i.e., under covariate shift. In this work, we propose an evaluation metric based on conditional independence testing to detect a lack of robustness to covariate shift in click models. We introduce the concept of debiasedness in click modeling and derive a metric for measuring it. In extensive semi-synthetic experiments, we show that our proposed metric helps to predict the downstream performance of click models under covariate shift and is useful in an off-policy model selection setting.