When facing uncertainty, decision-makers want predictions they can trust. A machine learning provider can convey confidence to decision-makers by guaranteeing their predictions are distribution calibrated--- amongst the inputs that receive a predicted vector of class probabilities q, the actual distribution over classes is given by q. For multi-class prediction problems, however, directly optimizing predictions under distribution calibration tends to be infeasible, requiring sample complexity that grows exponentially in the number of classes C. In this work, we introduce a new notion---decision calibration---that requires the predicted distribution and true distribution over classes to be ``indistinguishable'' to downstream decision-makers. This perspective gives a new characterization of distribution calibration: a predictor is distribution calibrated if and only if it is decision calibrated with respect to all decision-makers. Our main result shows that under a mild restriction, unlike distribution calibration, decision calibration is actually feasible. We design a recalibration algorithm that provably achieves decision calibration efficiently, provided that the decision-makers have a bounded number of actions (e.g., polynomial in C). We validate our recalibration algorithm empirically: compared to existing methods, decision calibration improves decision-making on skin lesion and ImageNet classification with modern neural network predictors.

As data-driven predictive models are increasingly used to inform decisions, it has been argued that decision makers should provide explanations that help individuals understand what would have to change for these decisions to be beneficial ones. However, there has been little discussion on the possibility that individuals may use the above counterfactual explanations to invest effort strategically and maximize their chances of receiving a beneficial decision. In this paper, our goal is to find policies and counterfactual explanations that are optimal in terms of utility in such a strategic setting. We first show that, given a pre-defined policy, the problem of finding the optimal set of counterfactual explanations is NP-hard. Then, we show that the corresponding objective is nondecreasing and satisfies submodularity and this allows a standard greedy algorithm to enjoy approximation guarantees. In addition, we further show that the problem of jointly finding both the optimal policy and set of counterfactual explanations reduces to maximizing a non-monotone submodular function. As a result, we can use a recent randomized algorithm to solve the problem, which also offers approximation guarantees. Finally, we demonstrate that, by incorporating a matroid constraint into the problem formulation, we can increase the diversity of the optimal set of counterfactual explanations and incentivize individuals across the whole spectrum of the population to self improve. Experiments on synthetic and real lending and credit card data illustrate our theoretical findings and show that the counterfactual explanations and decision policies found by our algorithms achieve higher utility than several competitive baselines.

Machine learning is a powerful tool for predicting human-related outcomes, from creditworthiness to heart attack risks. But when deployed transparently, learned models also affect how users act in order to improve outcomes. The standard approach to learning predictive models is agnostic to induced user actions and provides no guarantees as to the effect of actions. We provide a framework for learning predictors that are accurate, while also considering interactions between the learned model and user decisions. For this, we introduce look-ahead regularization which, by anticipating user actions, encourages predictive models to also induce actions that improve outcomes. This regularization carefully tailors the uncertainty estimates that govern confidence in this improvement to the distribution of model-induced actions. We report the results of experiments on real and synthetic data that show the effectiveness of this approach.

Learning controllers with offline data in decision-making systems is an essential area of research due to its potential to reduce the risk of applications in real-world systems. However, in responsibility-sensitive settings such as healthcare, decision accountability is of paramount importance, yet has not been adequately addressed by the literature.This paper introduces the Accountable Offline Controller (AOC) that employs the offline dataset as the Decision Corpus and performs accountable control based on a tailored selection of examples, referred to as the Corpus Subset. AOC operates effectively in low-data scenarios, can be extended to the strictly offline imitation setting, and displays qualities of both conservation and adaptability.We assess AOC's performance in both simulated and real-world healthcare scenarios, emphasizing its capability to manage offline control tasks with high levels of performance while maintaining accountability.

The authors motivate the proposed model with the setting in which items have "true" but unobserved labels/ratings and the observed labels/ratings given by evaluators are potentially incorrect. This differs from the very common problem in recommendation systems or collaborative filtering where evaluators provide their subjective ratings but there is not assumed to be any "true" rating (e.g., users of Netflix giving 1-5 star ratings to movies). This seems like a common but underexplored setting that is worthy of further study within machine learning. The authors are also right to highlight interpretability as a desired aspect of any machine learning solution that may yield post-hoc insights into common human biases and thus suggest corrective measures. This paper does a good job of motivating the proposed model and situating it within the crowdsourcing and human annotation literature.

The dynamics of simple decisions are well understood and modeled as a class of random walk models [e.g.

The news: Self-driving car startup Wayve can now interrogate its vehicles, asking them questions about their driving decisions--and getting answers back thanks to a chatbot. How it works: The idea is to use the same tech behind ChatGPT to help train driverless cars. The company combined its existing self-driving software with a large language model, creating a hybrid model that syncs up video data and driving data with natural-language descriptions that capture what the car sees and what it does. Why it matters: Wayve is treating the news as a breakthrough in AI safety. By quizzing its self-driving software every step of the way, Wayve hopes to understand exactly why and how its cars make certain decisions--and to uncover mistakes more quickly.

UN High Commissioner for Human Rights Michelle Bachelet called Wednesday for a moratorium on artificial intelligence (AI) systems that threaten human rights until enough safeguards are in place. Bachelet said in a press release: "Given the rapid and continuous growth of AI, filling the immense accountability gap in how data is collected, stored, shared and used is one of the most urgent human rights questions we face." As a part of its work on technology and human rights, the UN Human Rights Office published a report analyzing how AI affects people's right to privacy and other human rights such as the right to health, education, freedom of movement and freedom of expression. The report highlights the "undeniable and steadily growing impacts of AI technologies on the exercise of the right to privacy and other human rights." Bachelet noted that "the risk of discrimination linked to AI-driven decisions--decisions that can change, define or damage human lives--is all too real."