Search for: All records

Trust is predictive of civic cooperation and economic growth. Recently, the U.S. public has demonstrated increased partisan division and a surveyed decline in trust in institutions. There is a need to quantify individual and community levels of trust unobtrusively and at scale. Using observations of language across more than 16,000 Facebook users, along with their self-reported generalized trust score, we develop and evaluate a language-based assessment of generalized trust. We then apply the assessment to more than 1.6 billion geotagged tweets collected between 2009 and 2015 and derive estimates of trust across 2,041 U.S. counties. We find generalized trust was associated with more affiliative words (love, we, andfriends) and less angry words (hateandstupid) but only had a weak association with social words primarily driven by strong negative associations with general othering terms (“they” and “people”). At the county level, associations with the Centers for Disease Control and Prevention (CDC) and Gallup surveys suggest that people in high-trust counties were physically healthier and more satisfied with their community and their lives. Our study demonstrates that generalized trust levels can be estimated from language as a low-cost, unobtrusive method to monitor variations in trust in large populations. 

Replicating human-like dexterity in robot hands represents one of the largest open problems in robotics. Reinforcement learning is a promising approach that has achieved impressive progress in the last few years; however, the class of problems it has typically addressed corresponds to a rather narrow definition of dexterity as compared to human capabilities. To address this gap, we investigate piano-playing, a skill that challenges even the human limits of dexterity, as a means to test high-dimensional control, and which requires high spatial and temporal precision, and complex finger coordination and planning. We introduce RoboPianist, a system that enables simulated anthropomorphic hands to learn an extensive repertoire of 150 piano pieces where traditional model-based optimization struggles. We additionally introduce an open-sourced environment, benchmark of tasks, interpretable evaluation metrics, and open challenges for future study. 

Conveying complex objectives to reinforcement learning (RL) agents can often be difficult, involving meticulous design of reward functions that are sufficiently informative yet easy enough to provide. Human-in-the-loop RL methods allow practitioners to instead interactively teach agents through tailored feedback; however, such approaches have been challenging to scale since human feedback is very expensive. In this work, we aim to make this process more sample- and feedback-efficient. We present an off-policy, interactive RL algorithm that capitalizes on the strengths of both feedback and off-policy learning. Specifically, we learn a reward model by actively querying a teacher’s preferences between two clips of behavior and use it to train an agent. To enable off-policy learning, we relabel all the agent’s past experience when its reward model changes. We additionally show that pre-training our agents with unsupervised exploration substantially increases the mileage of its queries. We demonstrate that our approach is capable of learning tasks of higher complexity than previously considered by human-in-the-loop methods, including a variety of locomotion and robotic manipulation skills. We also show that our method is able to utilize real-time human feedback to effectively prevent reward exploitation and learn new behaviors that are difficult to specify with standard reward functions.