Playing with Food: Learning Food Item Representations through Interactive Exploration

A key challenge in robotic food manipulation is modeling the material properties of diverse and deformable food items. We propose using a multimodal sensory approach to interact and play with food that facilitates the ability to distinguish these properties across food items. First, we use a robotic arm and an array of sensors, which are synchronized using ROS, to collect a diverse dataset consisting of 21 unique food items with varying slices and properties. Afterwards, we learn visual embedding networks that utilize a combination of proprioceptive, audio, and visual data to encode similarities among food items using a triplet loss formulation. Our evaluations show that embeddings learned through interactions can successfully increase performance in a wide range of material and shape classification tasks. We envision that these learned embeddings can be utilized as a basis for planning and selecting optimal parameters for more material-aware robotic food manipulation skills. Furthermore, we hope to stimulate further innovations in the field of food robotics by sharing this food playing dataset with the research community.

Learning Skills to Patch Plans Based on Inaccurate Models

Planners using accurate models can be effective for accomplishing manipulation tasks in the real world, but are typically highly specialized and require significant fine-tuning to be reliable. Meanwhile, learning is useful for adaptation, but can require a substantial amount of data collection. In this paper, we propose a method that improves the efficiency of sub-optimal planners with approximate but simple and fast models by switching to a model-free policy when unexpected transitions are observed. Unlike previous work, our method specifically addresses when the planner fails due to transition model error by patching with a local policy only where needed. First, we use a sub-optimal model-based planner to perform a task until model failure is detected. Next, we learn a local model-free policy from expert demonstrations to complete the task in regions where the model failed. To show the efficacy of our method, we perform experiments with a shape insertion puzzle and compare our results to both pure planning and imitation learning approaches. We then apply our method to a door opening task. Our experiments demonstrate that our patch-enhanced planner performs more reliably than pure planning and with lower overall sample complexity than pure imitation learning.

Anomaly Detection in Bitcoin Network Using Unsupervised Learning Methods

The problem of anomaly detection has been studied for a long time. In short, anomalies are abnormal or unlikely things. In financial networks, thieves and illegal activities are often anomalous in nature. Members of a network want to detect anomalies as soon as possible to prevent them from harming the network's community and integrity. Many Machine Learning techniques have been proposed to deal with this problem; some results appear to be quite promising but there is no obvious superior method. In this paper, we consider anomaly detection particular to the Bitcoin transaction network. Our goal is to detect which users and transactions are the most suspicious; in this case, anomalous behavior is a proxy for suspicious behavior. To this end, we use three unsupervised learning methods including k-means clustering, Mahalanobis distance, and Unsupervised Support Vector Machine (SVM) on two graphs generated by the Bitcoin transaction network: one graph has users as nodes, and the other has transactions as nodes.