Decision trees are regarded for high interpretability arising from their hierarchical partitioning structure built on simple decision rules. However, in practice, this is not realized because axis-aligned partitioning of realistic data results in deep trees, and because ensemble methods are used to mitigate overfitting. Even then, model complexity and performance remain sensitive to transformation of the input, and extensive expert crafting of features from the raw data is common. We propose the first system to alternate sparse feature learning with differentiable decision tree construction to produce small, interpretable trees with good performance. It benchmarks favorably against conventional tree-based models and demonstrates several notions of interpretability of a model and its predictions.

"I believe the technology was deployed too quickly in too vast amounts, with hundreds of vehicles, when it wasn't really ready," one police official told federal regulators last month. Emergency first-responder leaders told federal regulators in a private meeting last month that they were frustrated with the performance of autonomous vehicles on their streets--that city firefighters, police officers, EMTs, and paramedics are forced to spend time during emergencies resolving issues with frozen or stuck cars. One fire official called them "a safety issue for our crews as well as the victims." WIRED obtained an audio recording of the meeting. Officials from San Francisco and Austin, where Waymo has been ferrying passengers without drivers for more than a year, said the vehicles' performance is getting worse.

Researchers show scammers are using AI-manipulated footage of celebrity interviews to trick users into sharing their personal data. Last week, Taylor Swift filed a trio of trademark applications to protect her image and voice. One is meant to cover a well-known photograph of the pop singer holding a pink guitar during a concert on her record-breaking Eras tour, while the two sound trademarks are for simple identifying phrases: "Hey, it's Taylor Swift" and "Hey, it's Taylor." The move comes as AI deepfakes continue to proliferate across social media. Any individual stands to have their likeness exploited in the creation of nonconsensual AI-generated material; earlier this month, an Ohio man was the first person convicted under a new federal law criminalizing "intimate" visual deceptions of this sort.

Incorporating causality into reinforcement learning methods increases the interpretability of artificial636 intelligence, which helps humans understand the underlying mechanism of algorithms and check637 the source of failures. However, the learned causal transition model may contain human-readable638 private information about the environment, which could raise privacy issues. To mitigate this potential639 negative societal impact, the causal transition model needs to be encrypted and only accessible to640 algorithms and trustworthy users.641 In this section, besides the most related formulation, robust RL introduced in Sec 3.3, we also643 introduce some other related RL problem formulations partially shown in Figure 3. Then, we limit644 our discussion to mainly two lines of work that are related to ours: (1) promoting robustness in RL;645 (2) concerning the spurious correlation issues in RL.646 B.1 Related RL formulations647 Robustness to noisy state: POMDPs and SA-MDPs.

Robustness has been extensively studied in reinforcement learning (RL) to handle various forms of uncertainty such as random perturbations, rare events, and malicious attacks. In this work, we consider one critical type of robustness against spurious correlation, where different portions of the state do not have correlations induced by unobserved confounders. These spurious correlations are ubiquitous in real-world tasks, for instance, a self-driving car usually observes heavy traffic in the daytime and light traffic at night due to unobservable human activity. A model that learns such useless or even harmful correlation could catastrophically fail when the confounder in the test case deviates from the training one. Although motivated, enabling robustness against spurious correlation poses significant challenges since the uncertainty set, shaped by the unobserved confounder and causal structure, is difficult to characterize and identify. Existing robust algorithms that assume simple and unstructured uncertainty sets are therefore inadequate to address this challenge. To solve this issue, we propose Robust State-Confounded Markov Decision Processes (RSC-MDPs) and theoretically demonstrate its superiority in avoiding learning spurious correlations compared with other robust RL counterparts. We also design an empirical algorithm to learn the robust optimal policy for RSC-MDPs, which outperforms all baselines in eight realistic self-driving and manipulation tasks. Please refer to the website for more details.