Increased defense spending is a boon for Germany's ailing industrial sector. Numerous companies, even those with no previous military experience, are now hoping to get in on the act. Visiting the works of Ilsenburger Grobblech GmbH is like taking a trip back in time. Way back in the 16th century, copper used to be produced at this site in the northern Harz Mountains, not far from eastern Germany' tallest peak, the Brocken. Today, slabs of steel up to 35 centimeters thick are piled up in front of the factory halls, delivered from the blast furnaces and converters of parent company Salzgitter, less than an hour's drive away. What is happening behind the factory walls, though, is part of a new hype that has gripped Germany's crisis-ridden industrial sector. A hype which many are hoping will be enough to revive it.

We consider the problem of strategic classification, where a learner must build a model to classify agents based on features that have been strategically modified. Previous work in this area has concentrated on the case when the learner is restricted to deterministic classifiers. In contrast, we perform a theoretical analysis of an extension to this setting that allows the learner to produce a randomised classifier. We show that, under certain conditions, the optimal randomised classifier can achieve better accuracy than the optimal deterministic classifier, but under no conditions can it be worse. When a finite set of training data is available, we show that the excess risk of Strategic Empirical Risk Minimisation over the class of randomised classifiers is bounded in a similar manner as the deterministic case. In both the deterministic and randomised cases, the risk of the classifier produced by the learner converges to that of the corresponding optimal classifier as the volume of available training data grows. Moreover, this convergence happens at the same rate as in the i.i.d. case. Our findings are compared with previous theoretical work analysing the problem of strategic classification. We conclude that randomisation has the potential to alleviate some issues that could be faced in practice without introducing any substantial downsides.

Often in prediction tasks, the predictive model itself can influence the distribution of the target variable, a phenomenon termed performative prediction. Generally, this influence stems from strategic actions taken by stakeholders with a vested interest in predictive models. A key challenge that hinders the widespread adaptation of performative prediction in machine learning is that practitioners are generally unaware of the social impacts of their predictions. To address this gap, we propose a methodology for learning the distribution map that encapsulates the long-term impacts of predictive models on the population. Specifically, we model agents' responses as a cost-adjusted utility maximization problem and propose estimates for said cost. Our approach leverages optimal transport to align pre-model exposure (ex ante) and post-model exposure (ex post) distributions. We provide a rate of convergence for this proposed estimate and assess its quality through empirical demonstrations on a credit-scoring dataset.

Scaling laws have driven remarkable progress across machine learning domains like language modeling and computer vision. However, the exploration of scaling laws in embodied AI and robotics has been limited, despite the rapidly increasing usage of machine learning in this field. This paper presents the first study to quantify scaling laws for Robot Foundation Models (RFMs) and the use of LLMs in robotics tasks. Through a meta-analysis spanning 198 research papers, we analyze how key factors like compute, model size, and training data quantity impact model performance across various robotic tasks. Our findings confirm that scaling laws apply to both RFMs and LLMs in robotics, with performance consistently improving as resources increase. The power law coefficients for RFMs closely match those of LLMs in robotics, resembling those found in computer vision and outperforming those for LLMs in the language domain. We also note that these coefficients vary with task complexity, with familiar tasks scaling more efficiently than unfamiliar ones, emphasizing the need for large and diverse datasets. Furthermore, we highlight the absence of standardized benchmarks in embodied AI. Most studies indicate diminishing returns, suggesting that significant resources are necessary to achieve high performance, posing challenges due to data and computational limitations. Finally, as models scale, we observe the emergence of new capabilities, particularly related to data and model size.

A class-action lawsuit filed in a US federal court last Valentine's Day accuses Match Group – the owners of Tinder, Hinge and OkCupid dating apps, among others – of using a "predatory business model" and of doing everything in its power to keep users hooked, in flagrant opposition to Hinge's claim that it is "designed to be deleted". The lawsuit crystallised an ocean of dissatisfaction with the apps, and stimulated a new round of debate over their potential to harm mental health, but for scientists who study romantic relationships it sidestepped the central issue: do they work? Does using the apps increase your chances of finding your soulmate, or not? The answer is, nobody knows. "The science isn't there," says sociologist Elizabeth Bruch of the University of Michigan, who has studied online dating for a decade.

Modern dating can be severed into two eras: before the swipe, and after. When Tinder and other dating apps took off in the early 2010s, they unleashed a way to more easily access potential love interests than ever before. By 2017, about five years after Tinder introduced the swipe, more than a quarter of different-sex couples were meeting on apps and dating websites, according to a study led by the Stanford sociologist Michael Rosenfeld. Suddenly, saying "We met on Hinge" was as normal as saying "We met in college" or "We met through a friend." The share of couples meeting on apps has remained pretty consistent in the years since his 2017 study, Rosenfeld told me.

A primary goal in strategic classification is to learn decision rules which are robust to strategic input manipulation. Earlier works assume that strategic responses are known; while some recent works address the important challenge of unknown responses, they exclusively study sequential settings which allow multiple model deployments over time. But there are many domains$\unicode{x2014}$particularly in public policy, a common motivating use-case$\unicode{x2014}$where multiple deployments are unrealistic, or where even a single bad round is undesirable. To address this gap, we initiate the study of strategic classification under unknown responses in the one-shot setting, which requires committing to a single classifier once. Focusing on the users' cost function as the source of uncertainty, we begin by proving that for a broad class of costs, even a small mis-estimation of the true cost can entail arbitrarily low accuracy in the worst case. In light of this, we frame the one-shot task as a minimax problem, with the goal of identifying the classifier with the smallest worst-case risk over an uncertainty set of possible costs. Our main contribution is efficient algorithms for both the full-batch and stochastic settings, which we prove converge (offline) to the minimax optimal solution at the dimension-independent rate of $\tilde{\mathcal{O}}(T^{-\frac{1}{2}})$. Our analysis reveals important structure stemming from the strategic nature of user responses, particularly the importance of dual norm regularization with respect to the cost function.

When users can benefit from certain predictive outcomes, they may be prone to act to achieve those outcome, e.g., by strategically modifying their features. The goal in strategic classification is therefore to train predictive models that are robust to such behavior. However, the conventional framework assumes that changing features does not change actual outcomes, which depicts users as "gaming" the system. Here we remove this assumption, and study learning in a causal strategic setting where true outcomes do change. Focusing on accuracy as our primary objective, we show how strategic behavior and causal effects underlie two complementing forms of distribution shift. We characterize these shifts, and propose a learning algorithm that balances between these two forces and over time, and permits end-to-end training. Experiments on synthetic and semi-synthetic data demonstrate the utility of our approach.

Any classifier can be "smoothed out" under Gaussian noise to build a new classifier that is provably robust to $\ell_2$-adversarial perturbations, viz., by averaging its predictions over the noise via randomized smoothing. Under the smoothed classifiers, the fundamental trade-off between accuracy and (adversarial) robustness has been well evidenced in the literature: i.e., increasing the robustness of a classifier for an input can be at the expense of decreased accuracy for some other inputs. In this paper, we propose a simple training method leveraging this trade-off to obtain robust smoothed classifiers, in particular, through a sample-wise control of robustness over the training samples. We make this control feasible by using "accuracy under Gaussian noise" as an easy-to-compute proxy of adversarial robustness for an input. Specifically, we differentiate the training objective depending on this proxy to filter out samples that are unlikely to benefit from the worst-case (adversarial) objective. Our experiments show that the proposed method, despite its simplicity, consistently exhibits improved certified robustness upon state-of-the-art training methods. Somewhat surprisingly, we find these improvements persist even for other notions of robustness, e.g., to various types of common corruptions.

The number of multi-robot systems deployed in field applications has risen dramatically over the years. Nevertheless, supervising and operating multiple robots at once is a difficult task for a single operator to execute. In this paper we propose a novel approach for utilizing advising automated agents when assisting an operator to better manage a team of multiple robots in complex environments. We introduce the Myopic Advice Optimization (MYAO) Problem and exemplify its implementation using an agent for the Search And Rescue (SAR) task. Our intelligent advising agent was evaluated through extensive field trials, with 44 non-expert human operators and 10 low-cost mobile robots, in simulation and physical deployment, and showed a significant improvement in both team performance and the operator's satisfaction.