SSUP-HRI: Social Signaling in Urban Public Human-Robot Interaction dataset

This paper introduces our dataset featuring human-robot interactions (HRI) in urban public environments. This dataset is rich with social signals that we believe can be modeled to help understand naturalistic human-robot interaction. Our dataset currently comprises approximately 15 hours of video footage recorded from the robots' perspectives, within which we annotated a total of 274 observable interactions featuring a wide range of naturalistic human-robot interactions. The data was collected by two mobile trash barrel robots deployed in Astor Place, New York City, over the course of a week. We invite the HRI community to access and utilize our dataset. To the best of our knowledge, this is the first dataset showcasing robot deployments in a complete public, non-controlled setting involving urban residents.

Human-Robot Commensality: Bite Timing Prediction for Robot-Assisted Feeding in Groups

We develop data-driven models to predict when a robot should feed during social dining scenarios. Being able to eat independently with friends and family is considered one of the most memorable and important activities for people with mobility limitations. Robots can potentially help with this activity but robot-assisted feeding is a multi-faceted problem with challenges in bite acquisition, bite timing, and bite transfer. Bite timing in particular becomes uniquely challenging in social dining scenarios due to the possibility of interrupting a social human-robot group interaction during commensality. Our key insight is that bite timing strategies that take into account the delicate balance of social cues can lead to seamless interactions during robot-assisted feeding in a social dining scenario. We approach this problem by collecting a multimodal Human-Human Commensality Dataset (HHCD) containing 30 groups of three people eating together. We use this dataset to analyze human-human commensality behaviors and develop bite timing prediction models in social dining scenarios. We also transfer these models to human-robot commensality scenarios. Our user studies show that prediction improves when our algorithm uses multimodal social signaling cues between diners to model bite timing. The HHCD dataset, videos of user studies, and code will be publicly released after acceptance.

Learning Representations from Temporally Smooth Data

Events in the real world are correlated across nearby points in time, and we must learn from this temporally smooth data. However, when neural networks are trained to categorize or reconstruct single items, the common practice is to randomize the order of training items. What are the effects of temporally smooth training data on the efficiency of learning? We first tested the effects of smoothness in training data on incremental learning in feedforward nets and found that smoother data slowed learning. Moreover, sampling so as to minimize temporal smoothness produced more efficient learning than sampling randomly. If smoothness generally impairs incremental learning, then how can networks be modified to benefit from smoothness in the training data? We hypothesized that two simple brain-inspired mechanisms, leaky memory in activation units and memory-gating, could enable networks to rapidly extract useful representations from smooth data. Across all levels of data smoothness, these brain-inspired architectures achieved more efficient category learning than feedforward networks. This advantage persisted, even when leaky memory networks with gating were trained on smooth data and tested on randomly-ordered data. Finally, we investigated how these brain-inspired mechanisms altered the internal representations learned by the networks. We found that networks with multi-scale leaky memory and memory-gating could learn internal representations that un-mixed data sources which vary on fast and slow timescales across training samples. Altogether, we identified simple mechanisms enabling neural networks to learn more quickly from temporally smooth data, and to generate internal representations that separate timescales in the training signal.

Object Permanence Through Audio-Visual Representations

As robots perform manipulation tasks and interact with objects, it is probable that they accidentally drop objects that subsequently bounce out of their visual fields (e.g., due to an inadequate grasp of an unfamiliar object). To enable robots to recover from such errors, we draw upon the concept of object permanence---objects remain in existence even when they are not being sensed (e.g., seen) directly. In particular, we developed a multimodal neural network model---using a partial, observed bounce trajectory and the audio resulting from drop impact as its inputs---to predict the full bounce trajectory and the end location of a dropped object. We empirically show that: (1) our multimodal method predicted end locations close in proximity (i.e., within the visual field of the robot's wrist camera) to the actual locations and (2) the robot was able to retrieve dropped objects by applying minimal vision-based pick-up adjustments. Additionally, we show that our multimodal method outperformed the vision-only and audio-only baselines in retrieving dropped objects. Our results provide insights in enabling object permanence for robots and offer foundations for ensuring robust robot autonomy in task execution.