Have you ever gone to a sporting event and spent what seems like an eternity just trying to get in? Security technologies can slow lines down to a grinding halt, and they're not always effective – your necklace, keys, and belt may set the metal detector off, while weapons can still get through. At the FirstEnergy Stadium in Cleveland, it turns out a lot of football fans wear steel-toed boots. "Everyone who wore these boots were setting off when they were coming in," says Brandon Covert, vice president of information technology for the Cleveland Browns. The team has managed to overcome this problem with artificial intelligence, after adopting security screening technology from the company Evolv.

Researchers have proposed a wide variety of model explanation approaches, but it remains unclear how most methods are related or when one method is preferable to another. We establish a new class of methods, removal-based explanations, that are based on the principle of simulating feature removal to quantify each feature's influence. These methods vary in several respects, so we develop a framework that characterizes each method along three dimensions: 1) how the method removes features, 2) what model behavior the method explains, and 3) how the method summarizes each feature's influence. Our framework unifies 25 existing methods, including several of the most widely used approaches (SHAP, LIME, Meaningful Perturbations, permutation tests). This new class of explanation methods has rich connections that we examine using tools that have been largely overlooked by the explainability literature. To anchor removal-based explanations in cognitive psychology, we show that feature removal is a simple application of subtractive counterfactual reasoning. Ideas from cooperative game theory shed light on the relationships and trade-offs among different methods, and we derive conditions under which all removal-based explanations have information-theoretic interpretations. Through this analysis, we develop a unified framework that helps practitioners better understand model explanation tools, and that offers a strong theoretical foundation upon which future explainability research can build.

Billionaire philanthropist Paul Allen has announced a 100 million commitment over 10 years to fund scientific endeavors at the "frontiers of bioscience" that he describes as having major implications for humankind. An initial set of grants, announced Wednesday, will go to Stanford and Tufts universities for the creation of new research centers and to individual scientists with unconventional approaches to projects in tissue regeneration, antibiotic resistance, gene editing and the development of brain circuitry. Allen said his commitment grew out of a realization that the biological sciences are at a critical point in history, with technology now able to take the field in a more quantitative direction than ever before. New tools can manipulate DNA, next-generation microscopes measure and create images of the tiniest parts of living systems, and super-powerful computers are able to make sense of massive amounts of data. "What I believe is that this is potentially a game-changer for our understanding of complex biological systems," Allen said.