Deep learning-based grasp prediction models have become an industry standard for robotic bin-picking systems. To maximize pick success, production environments are often equipped with several end-effector tools that can be swapped on-the-fly, based on the target object. Tool-change, however, takes time. Choosing the order of grasps to perform, and corresponding tool-change actions, can improve system throughput; this is the topic of our work. The main challenge in planning tool change is uncertainty - we typically cannot see objects in the bin that are currently occluded. Inspired by queuing and admission control problems, we model the problem as a Markov Decision Process (MDP), where the goal is to maximize expected throughput, and we pursue an approximate solution based on model predictive control, where at each time step we plan based only on the currently visible objects. Special to our method is the idea of void zones, which are geometrical boundaries in which an unknown object will be present, and therefore cannot be accounted for during planning. Our planning problem can be solved using integer linear programming (ILP). However, we find that an approximate solution based on sparse tree search yields near optimal performance at a fraction of the time. Another question that we explore is how to measure the performance of tool-change planning: we find that throughput alone can fail to capture delicate and smooth behavior, and propose a principled alternative. Finally, we demonstrate our algorithms on both synthetic and real world bin picking tasks.

User experiences can be made more engaging by incorporating surprise. For example, online shoppers may like to view unique products. In this paper we propose an approach for detecting surprising documents, such as product titles. As the concept of surprise is subjective, there is currently no principled method for measuring the surprisingness score of a document. We present such a method; an unsupervised approach for automatically discovering surprising documents in an unlabeled corpus. Our approach is based on a probabilistic model of surprise, and a construction of effective distributional word embeddings, which can be adapted to the semantic context in which the word appears. As the performance of our model does not degrade with the length of the document, it is particularly well suited for very short documents (even a single sentence). We evaluate our model both in supervised and unsupervised settings, demonstrating its state-of-the-art performance on two real-world data sets: a collection of e-commerce products from eBay, and a corpus of NSF proposals. These experiments show that our surprisingness score exhibits high correlation with human annotated labels.

Large, relational factor graphs with structure defined by first-order logic or other languages give rise to notoriously difficult inference problems. Because unrolling the structure necessary to represent distributions over all hypotheses has exponential blow-up, solutions are often derived from MCMC. However, because of limitations in the design and parameterization of the jump function, these sampling-based methods suffer from local minima|the system must transition through lower-scoring configurations before arriving at a better MAP solution. This paper presents a new method of explicitly selecting fruitful downward jumps by leveraging reinforcement learning (RL). Rather than setting parameters to maximize the likelihood of the training data, parameters of the factor graph are treated as a log-linear function approximator and learned with temporal difference (TD); MAP inference is performed by executing the resulting policy on held out test data. Our method allows efficient gradient updates since only factors in the neighborhood of variables affected by an action need to be computed|we bypass the need to compute marginals entirely. Our method provides dramatic empirical success, producing new state-of-the-art results on a complex joint model of ontology alignment, with a 48\% reduction in error over state-of-the-art in that domain.

Recently, many advanced machine learning approaches have been proposed for coreference resolution; however, all of the discriminatively-trained models reason over mentions rather than entities. That is, they do not explicitly contain variables indicating the “canonical” values for each attribute of an entity (e.g., name, venue, title, etc.). This canonicalization step is typically implemented as a post-processing routine to coreference resolution prior to adding the extracted entity to a database. In this paper, we propose a discriminatively-trained model that jointly performs coreference resolution and canonicalization, enabling features over hypothesized entities. We validate our approach on two different coreference problems: newswire anaphora resolution and research paper citation matching, demonstrating improvements in both tasks and achieving an error reduction of up to 62% when compared to a method that reasons about mentions only.

The automatic consolidation of database records from many heterogeneous sources into a single repository requires solving several information integration tasks. Although tasks such as coreference, schema matching, and canonicalization are closely related, they are most commonly studied in isolation. Systems that do tackle multiple integration problems traditionally solve each independently, allowing errors to propagate from one task to another. In this paper, we describe a discriminatively-trained model that reasons about schema matching, coreference, and canonicalization jointly. We evaluate our model on a real-world data set of people and demonstrate that simultaneously solving these tasks reduces errors over a cascaded or isolated approach. Our experiments show that a joint model is able to improve substantially over systems that either solve each task in isolation or with the conventional cascade. We demonstrate nearly a 50% error reduction for coreference and a 40% error reduction for schema matching.

We propose a fast approach for detecting collision- free swing-foot trajectories for legged locomotion over extreme terrains. Instead of simulating the swing trajectories and checking for collisions along them, our approach uses machine learning techniques to predict whether a swing trajectory is collision-free. Using a set of local terrain features, we apply supervised learning to train a classifier to predict collisions. Both in simulation and on a real quadruped platform, our results show that our classifiers can improve the accuracy of collision detection compared to a real-time geometric approach without significantly increasing the computation time.

We study an approach for performing concurrent activities in Markov decision processes (MDPs) based on the coarticulation framework. We assume that the agent has multiple degrees of freedom (DOF) in the action space which enables it to perform activities simultaneously. We demonstrate that one natural way for generating concurrency in the system is by coarticulating among the set of learned activities available to the agent. In general due to the multiple DOF in the system, often there exists a redundant set of admissible sub-optimal policies associated with each learned activity. Such flexibility enables the agent to concurrently commit to several subgoals according to their priority levels, given a new task defined in terms of a set of prioritized subgoals. We present efficient approximate algorithms for computing such policies and for generating concurrent plans. We also evaluate our approach in a simulated domain.

In sequence modeling, we often wish to represent complex interaction between labels, such as when performing multiple, cascaded labeling tasks on the same sequence, or when long-range dependencies exist. We present dynamic conditional random fields (DCRFs), a generalization of linear-chain conditional random fields (CRFs) in which each time slice contains a set of state variables and edges---a distributed state representation as in dynamic Bayesian networks (DBNs)---and parameters are tied across slices. Since exact inference can be intractable in such models, we perform approximate inference using several schedules for belief propagation, including tree-based reparameterization (TRP). On a natural-language chunking task, we show that a DCRF performs better than a series of linear-chain CRFs, achieving comparable performance using only half the training data.

We investigate an approach for simultaneously committing to mul- tiple activities, each modeled as a temporally extended action in a semi-Markov decision process (SMDP). For each activity we de- fine a set of admissible solutions consisting of the redundant set of optimal policies, and those policies that ascend the optimal state- value function associated with them. A plan is then generated by merging them in such a way that the solutions to the subordinate activities are realized in the set of admissible solutions satisfying the superior activities. We present our theoretical results and em- pirically evaluate our approach in a simulated domain.

We investigate a general semi-Markov Decision Process (SMDP) framework for modeling concurrent decision making, where agents learn optimal plans over concurrent temporally extended actions. We introduce three types of parallel termination schemes { all, any and continue { and theoretically and experimentally compare them.

We propose and investigate a general framework for hierarchical modeling of partially observable environments, such as office buildings, using hierarchical hidden Markov models (HHMMs). Our main goal is to explore hierarchical modeling as a basis for designing more efficient methods for model construction and usage. As a case study we focus on indoor robot navigation and show how this framework can be used to learn a hierarchy of models of the environment at different levels of spatial abstraction. We introduce the idea of model reuse that can be used to combine already learned models into a larger model. We describe an extension of the HHMM model to includes actions, which we call hierarchical POMDPs, and describe a modified hierarchical Baum-Welch algorithm to learn these models. We train different families of hierarchical models for a simulated and a real world corridor environment and compare them with the standard "flat" representation of the same environment. We show that the hierarchical POMDP approach, combined with model reuse, allows learning hierarchical models that fit the data better and train faster than flat models.