We present a novel approach for determining robot movements that efficiently accomplish the robot's tasks while not hindering of people within environment. Our models goal-directed trajectories pedestrians using maximum entropy inverse optimal control. The advantage this modeling is generality its learned cost function to changes in environment and entirely different environments. employ predictions model pedestrian incremental planner quantitatively show improvement hindrance-sensitive...

We present PROCAB, an efficient method for Probabilistically Reasoning from Observed Context-Aware Behavior. It models the context-dependent utilities and underlying reasons that people take different actions. The model generalizes to unseen situations scales incorporate rich contextual information. train our using route preferences of 25 taxi drivers demonstrated in over 100,000 miles collected data, demonstrate performance by inferring: (1) decision at next intersection, (2) known...

The principle of maximum entropy provides a powerful framework for statistical models joint, conditional, and marginal distributions. However, there are many important distributions with elements interaction feedback where its applicability has not been established. This work presents the causal entropy—an approach based on causally conditioned probabilities that can appropriately model availability influence sequentially revealed side information. Using this principle, we derive sequential...

Numerous interaction techniques have been developed that make "virtual" pointing at targets in graphical user interfaces easier than analogous physical tasks by invoking target-based interface modifications. These facilitation crucially depend on methods for estimating the relevance of potential targets. Unfortunately, many simple employed to date are inaccurate common settings with selectable close proximity. In this paper, we bring recent advances statistical machine learning bear...

Abstract In many animal societies, groups of individuals form stable social units that are shaped by well-delineated dominance hierarchies and a range affiliative relationships. How do socially complex maintain cohesion achieve collective movement? Using high-resolution GPS tracking members wild baboon troop, we test whether movement in is governed interactions among local neighbours (commonly found with largely anonymous memberships), affiliates, and/or paying attention to global group...

Reducing the large energy consumption of temperature regulation systems is a challenge for researchers and practitioners alike. In this paper, we explore compare two common types solutions: A manual that encourages reduced use, an intelligent automatic control system. We deployed eco-feedback system with ability to remotely one's thermostat ten participants three months. Participants appreciated thermostat, controlled their heating 78.8% accuracy, 6.3% improvement over not having However,...

The principle of maximum entropy provides a powerful framework for estimating joint, conditional, and marginal probability distributions. However, there are many important distributions with elements interaction feedback where its applicability has not been established. This paper presents the causal entropy-an approach based on directed information theory an unknown process interactions known process. We demonstrate breadth using two applications: predictive solution inverse optimal control...

Making predictions that are fair with regard to protected attributes (race, gender, age, etc.) has become an important requirement for classification algorithms. Existing techniques derive a model from sampled labeled data relying on the assumption training and testing identically independently drawn (iid) same distribution. In practice, distribution shift can does occur between datasets as characteristics of individuals interacting machine learning system change. We investigate fairness...

To facilitate interaction with people, robots must not only recognize current actions, but also infer a person's intentions and future behavior. Recent advances in depth camera technology have significantly improved human motion tracking. However, the inherent high dimensionality of interacting physical world makes efficiently forecasting intention behavior challenging task. Predictive methods that estimate uncertainty are therefore critical for supporting appropriate robotic responses to...

Advertisements simultaneously provide both economic support for most free web content and one of the largest annoyances to end users. Furthermore, modern advertisement ecosystem is rife with tracking methods which violate user privacy. A natural reaction users install ad blockers prevent advertisers from or displaying ads. Traditional blocking software relies upon hand-crafted filter expressions generate large, unwieldy regular matched against resources being included within pages. This...

Part of being a parent is taking responsibility for arranging and supplying transportation children between various events. Dual-income parents frequently develop routines to help manage with minimal amount attention. On days when families deviate from their routines, effective logistics can often depend on knowledge the routine location, availability intentions other family members. Since most rarely document activities, making that needed information unavailable, coordination breakdowns...

Word sense induction (WSI) seeks to automatically discover the senses of a word in corpus via unsupervised methods. We propose sense-topic model for WSI, which treats and topic as two separate latent variables be inferred jointly. Topics are informed by entire document, while local context surrounding ambiguous word. also discuss ways enriching original order improve performance, including using neural embeddings external corpora expand each data instance. demonstrate significant...

Pervasive computing allows the coupling of physical world to information world, and provides a wealth ubiquitous services applications that allow users, machines, data, applications, spaces interact seamlessly with one another. In this paper, we propose benchmark for evaluating pervasive environments. These proposed metrics facilitate assessment evaluation different aspects its support wide variety tasks.

Developing classification methods with high accuracy that also avoid unfair treatment of different groups has become increasingly important for data-driven decision making in social applications. Many existing enforce fairness constraints on a selected classifier (e.g., logistic regression) by directly forming constrained optimizations. We instead re-derive new from the first principles distributional robustness incorporates criteria into worst-case logarithmic loss minimization. This...