Automated feature discovery is a fundamental problem in machine learning. Although classical feature discovery methods do not guarantee optimal solutions in general, it has been recently noted that certain subspace learning and sparse coding problems can be solved efficiently, provided the number of features is not restricted a priori. We provide an extended characterization of this optimality result and describe the nature of the solutions under an expanded set of practical contexts. In particular, we apply the framework to a semi-supervised learning problem, and demonstrate that feature discovery can co-occur with input reconstruction and supervised training while still admitting globally optimal solutions. A comparison to existing semi-supervised feature discovery methods shows improved generalization and efficiency.

We address the problem of transferring information learned from experiments to a different environment, in which only passive observations can be collected. We introduce a formal representation called "selection diagrams" for expressing knowledge about differences and commonalities between environments and, using this representation, we derive procedures for deciding whether effects in the target environment can be inferred from experiments conducted elsewhere. When the answer is affirmative, the procedures identify the set of experiments and observations that need be conducted to license the transport. We further discuss how transportability analysis can guide the transfer of knowledge in non-experimental learning to minimize re-measurement cost and improve prediction power.

That drone and its basic capabilities are summarized in Figure 1. The flexibility and controllability of quadrotor helicopters have made them a recent focus of interest among robotics and AI research groups. At the same time, their popularity has led to a wide range of commercially available platforms, some at prices accessible for undergraduate educational use. This project evaluates the ARDrone quadrotor helicopter as a basis for use in undergraduate classes such as robotics, computer vision, or embodied AI. We have encountered both successes and frustrations in using the ARDrone to date. Looking forward, the quadrotor's capabilities do seem a promising basis for future curricular offerings.

We present two visual algorithms, called the affine and fractal methods, which each solve a considerable portion of the Raven’s Progressive Matrices (RPM) test. The RPM is considered to be one of the premier psychometric measures of general intelligence. Current computational accounts of the RPM assume that visual test inputs are translated into propositional representations before further reasoning takes place. We propose that visual strategies can also solve RPM problems, in line with behavioral evidence showing that humans do use visual strategies to some extent on the RPM. Our two visual methods currently solve RPM problems at the level of typical 9- to 10-year-olds.

The demand for computer-assisted language learning systems that can provide corrective feedback on language learners’ speaking has increased. However, it is not a trivial task to detect grammatical errors in oral conversations because of the unavoidable errors of automatic speech recognition systems. To provide corrective feedback, a novel method to detect grammatical errors in speaking performance is proposed. The proposed method consists of two sub-models: the grammaticality-checking model and the error-type classification model. We automatically generate grammatical errors that learners are likely to commit and construct error patterns based on the articulated errors. When a particular speech pattern is recognized, the grammaticality-checking model performs a binary classification based on the similarity between the error patterns and the recognition result using the confidence score. The error-type classification model chooses the error type based on the most similar error pattern and the error frequency extracted from a learner corpus. The grammaticality checking method largely outperformed the two comparative models by 56.36% and 42.61% in F-score while keeping the false positive rate very low. The error-type classification model exhibited very high performance with a 99.6% accuracy rate. Because high precision and a low false positive rate are important criteria for the language-tutoring setting, the proposed method will be helpful for intelligent computer-assisted language learning systems.

Inspired by “GoogleTM Sets” and Bayesian sets, we consider the problem of retrieving complex objects and relations among them, i.e., ground atoms from a logical concept, given a query consisting of a few atoms from that concept. We formulate this as a within-network relational learning problem using few labels only and describe an algorithm that ranks atoms using a score based on random walks with restart (RWR): the probability that a random surfer hits an atom starting from the query atoms. Specifically, we compute an initial ranking using personalized PageRank. Then, we find paths of atoms that are connected via their arguments, variablize the ground atoms in each path, in order to create features for the query. These features are used to re-personalize the original RWR and to finally compute the set completion, based on Label Propagation. Moreover, we exploit that RWR techniques can naturally be lifted and show that lifted inference for label propagation is possible. We evaluate our algorithm on a realworld relational dataset by finding completions of sets of objects describing the Roman city of Pompeii. We compare to Bayesian sets and show that our approach gives very reasonable set completions.

The initial idea of using Lego Mindstorms Robots for student courses had soon to be expanded to a simulation environment as the user base in students grew larger and the need for parallel development and testing arose. An easy to use and easy to set up means of providing positioning data led to the creation of an indoor positioning system so that new users can adapt quickly and successfully, as sensors on the actual robots are difficult to configure and hard to interpret in an environmental context. A global positioning system shared among robots can make local sensors obsolete and still deliver more precise information than currently available sensors, also providing the base necessary for the robots to effectively work on shared tasks as a group. Further more, a simulator for robots programmed with Fujaba and Java which was developed along the way can be used by many developers simultaneously and lets them evaluate their code in a simple way, while close to real-world results.

Integrating the expressive power of first-order logic with the ability of probabilistic reasoning of Bayesian networks has attracted the interest of many researchers for decades. We present an approach to integration that translates logical knowledge into Bayesian networks and uses Bayesian network composition to build a uniform representation that supports both logical and probabilistic reasoning. In particular, we propose a new way of translation of logical knowledge, relation search. Through the use of the proposed framework, without learning new languages or tools, modelers are allowed to 1) specify special knowledge using the most suitable languages, while reasoning in a uniform engine; 2) make use of pre-existing logical knowledge bases for probabilistic reasoning (to complete the model or minimize potential inconsistencies).

We propose a fast batch learning method for linear-chain Conditional Random Fields (CRFs) based on Newton-CG methods. Newton-CG methods are a variant of Newton method for high-dimensional problems. They only require the Hessian-vector products instead of the full Hessian matrices. To speed up Newton-CG methods for the CRF learning, we derive a novel dynamic programming procedure for the Hessian-vector products of the CRF objective function. The proposed procedure can reuse the byproducts of the time-consuming gradient computation for the Hessian-vector products to drastically reduce the total computation time of the Newton-CG methods. In experiments with tasks in natural language processing, the proposed method outperforms a conventional quasi-Newton method. Remarkably, the proposed method is competitive with online learning algorithms that are fast but unstable.

A key problem in statistical modeling is model selection, how to choose a model at an appropriate level of complexity. This problem appears in many settings, most prominently in choosing the number ofclusters in mixture models or the number of factors in factor analysis. In this tutorial we describe Bayesian nonparametric methods, a class of methods that side-steps this issue by allowing the data to determine the complexity of the model. This tutorial is a high-level introduction to Bayesian nonparametric methods and contains several examples of their application.