A Bayesian approach termed BAyesian Least Squares Optimization with Nonnegative L1-norm constraint (BALSON) is proposed. The error distribution of data fitting is described by Gaussian likelihood. The parameter distribution is assumed to be a Dirichlet distribution. With the Bayes rule, searching for the optimal parameters is equivalent to finding the mode of the posterior distribution. In order to explicitly characterize the nonnegative L1-norm constraint of the parameters, we further approximate the true posterior distribution by a Dirichlet distribution. We estimate the statistics of the approximating Dirichlet posterior distribution by sampling methods. Four sampling methods have been introduced. With the estimated posterior distributions, the original parameters can be effectively reconstructed in polynomial fitting problems, and the BALSON framework is found to perform better than conventional methods.

Deep learning and deep architectures are emerging as the best machine learning methods so far in many practical applications such as reducing the dimensionality of data, image classification, speech recognition or object segmentation.... In fact, many leading technology companies such as Google, Microsoft or IBM are researching and using deep architectures in their systems to replace other traditional models. Therefore, improving the performance of these models could make a very strong impact in the area of machine learning. However, deep learning is a very fast-growing research domain with many core methodologies and paradigms just discovered over the last few years. This thesis will first serve as a short summary of deep learning, which tries to include all of the most important ideas in this research area. Based on this knowledge, we suggested, and conducted some experiments to investigate the possibility of improving the deep learning based on automatic programming (ADATE). Although our experiments did produce good results, there are still many more possibilities that we could not try due to limited time as well as some limitations of the current ADATE version. I hope that this thesis can promote future work on this topic, especially when the next version of ADATE comes out. This thesis also includes a short analysis of the power of ADATE system, which could be very useful for other researchers who want to know what it is capable of.

Combining logic and probability has been a long stand- ing goal of AI research. Markov Logic Networks (MLNs) achieve this by attaching weights to formulas in first-order logic, and can be seen as templates for constructing features for ground Markov networks. Most techniques for learning weights of MLNs are domain-size agnostic, i.e., the size of the domain is not explicitly taken into account while learn- ing the parameters of the model. This often results in ex- treme probabilities when testing on domain sizes different from those seen during training. In this paper, we propose Domain Aware Markov logic Networks (DA-MLNs) which present a principled solution to this problem. While defin- ing the ground network distribution, DA-MLNs divide the ground feature weight by a scaling factor which is a function of the number of connections the ground atoms appearing in the feature are involved in. We show that standard MLNs fall out as a special case of our formalism when this func- tion evaluates to a constant equal to 1. Experiments on the benchmark Friends & Smokers domain show that our ap- proach results in significantly higher accuracies compared to existing methods when testing on domains whose sizes different from those seen during training.

While designing the state space of an MDP, it is common to include states that are transient or not reachable by any policy (e.g., in mountain car, the product space of speed and position contains configurations that are not physically reachable). This leads to defining weakly-communicating or multi-chain MDPs. In this paper, we introduce \tucrl, the first algorithm able to perform efficient exploration-exploitation in any finite Markov Decision Process (MDP) without requiring any form of prior knowledge. In particular, for any MDP with $S^{\texttt{C}}$ communicating states, $A$ actions and $\Gamma^{\texttt{C}} \leq S^{\texttt{C}}$ possible communicating next states, we derive a $\widetilde{O}(D^{\texttt{C}} \sqrt{\Gamma^{\texttt{C}} S^{\texttt{C}} AT})$ regret bound, where $D^{\texttt{C}}$ is the diameter (i.e., the longest shortest path) of the communicating part of the MDP. This is in contrast with optimistic algorithms (e.g., UCRL, Optimistic PSRL) that suffer linear regret in weakly-communicating MDPs, as well as posterior sampling or regularised algorithms (e.g., REGAL), which require prior knowledge on the bias span of the optimal policy to bias the exploration to achieve sub-linear regret. We also prove that in weakly-communicating MDPs, no algorithm can ever achieve a logarithmic growth of the regret without first suffering a linear regret for a number of steps that is exponential in the parameters of the MDP. Finally, we report numerical simulations supporting our theoretical findings and showing how TUCRL overcomes the limitations of the state-of-the-art.

With rise of digital age, there is an explosion of information in the form of news, articles, social media, and so on. Much of this data lies in unstructured form and manually managing and effectively making use of it is tedious, boring and labor intensive. This explosion of information and need for more sophisticated and efficient information handling tools gives rise to Information Extraction(IE) and Information Retrieval(IR) technology. Information Extraction systems takes natural language text as input and produces structured information specified by certain criteria, that is relevant to a particular application. Various sub-tasks of IE such as Named Entity Recognition, Coreference Resolution, Named Entity Linking, Relation Extraction, Knowledge Base reasoning forms the building blocks of various high end Natural Language Processing (NLP) tasks such as Machine Translation, Question-Answering System, Natural Language Understanding, Text Summarization and Digital Assistants like Siri, Cortana and Google Now. This paper introduces Information Extraction technology, its various sub-tasks, highlights state-of-the-art research in various IE subtasks, current challenges and future research directions.

The design of personalized incentives or recommendations to improve user engagement is gaining prominence as digital platform providers continually emerge. We propose a multi-armed bandit framework for matching incentives to users, whose preferences are unknown a priori and evolving dynamically in time, in a resource constrained environment. We design an algorithm that combines ideas from three distinct domains: (i) a greedy matching paradigm, (ii) the upper confidence bound algorithm (UCB) for bandits, and (iii) mixing times from the theory of Markov chains. For this algorithm, we provide theoretical bounds on the regret and demonstrate its performance via both synthetic and realistic (matching supply and demand in a bike-sharing platform) examples.

--We consider higher-order linear-chain conditional random fields (HO-LC-CRFs) for sequence modelling, and use sum-product networks (SPNs) for representing higher-order input-and output-dependent factors. SPNs are a recently introduced class of deep models for which exact and efficient inference can be performed. By combining HO-LC-CRFs with SPNs, expressive models over both the output labels and the hidden variables are instantiated while still enabling efficient exact inference. Furthermore, the use of higher-order factors allows us to capture relations of multiple input segments and multiple output labels as often present in real-world data. These relations can not be modeled by the commonly used first-order models and higher-order models with local factors including only a single output label. We demonstrate the effectiveness of our proposed models for sequence labeling. In extensive experiments, we outperform other state-of-the-art methods in optical character recognition and achieve competitive results in phone classification. For instance, they have been successfully used for speech recognition [3], optical character recognition and natural language processing [4]. Due to several advantages, LC-CRFs achieve better performance compared to their generative counterparts, i.e. hidden Markov models (HMMs) [3]. While LC-CRFs are normalized over the whole sequence, thereby counteracting the label bias problem, MEMMs are normalized locally .

The problem of recognizing actions or activities has been widely addressed in the computer vision research community: it consists in the classification of a fully or partially observed action, typically observed through cameras or external motion capture [2]. In robotics, recognizing the human activity is paramount for enabling a proper interaction and providing assistance to the human: an assistive device or prosthetics could switch control modes depending on the current human activity (e.g., walking or sitting) [3], [4]; a mobile robot may adapt its navigation depending on the prediction of the human motion [5]. More generally, prediction is important to provide the robot with anticipation capabilities [6]. In collaborative robotics applications in manufacturing, such as in assembly lines, recognizing the current activity of the operator is necessary for ergonomics evaluations [7] and for the optimization of the robot actions. However, there are two critical issues that prevent the direct application of existing techniques in such scenarios. The first issue is the availability of external sensing devices (cameras or motion captures) that poses constraints on the application for many tasks and application scenarios.

We consider reinforcement learning in input-driven environments, where an exogenous, stochastic input process affects the dynamics of the system. Input processes arise in many applications, including queuing systems, robotics control with disturbances, and object tracking. Since the state dynamics and rewards depend on the input process, the state alone provides limited information for the expected future returns. Therefore, policy gradient methods with standard state-dependent baselines suffer high variance during training. We derive a bias-free, input-dependent baseline to reduce this variance, and analytically show its benefits over state-dependent baselines. We then propose a meta-learning approach to overcome the complexity of learning a baseline that depends on a long sequence of inputs. Our experimental results show that across environments from queuing systems, computer networks, and MuJoCo robotic locomotion, input-dependent baselines consistently improve training stability and result in better eventual policies.

This paper is an attempt to bridge the conceptual gaps between researchers working on the two widely used approaches based on positive definite kernels: Bayesian learning or inference using Gaussian processes on the one side, and frequentist kernel methods based on reproducing kernel Hilbert spaces on the other. It is widely known in machine learning that these two formalisms are closely related; for instance, the estimator of kernel ridge regression is identical to the posterior mean of Gaussian process regression. However, they have been studied and developed almost independently by two essentially separate communities, and this makes it difficult to seamlessly transfer results between them. Our aim is to overcome this potential difficulty. To this end, we review several old and new results and concepts from either side, and juxtapose algorithmic quantities from each framework to highlight close similarities. We also provide discussions on subtle philosophical and theoretical differences between the two approaches.