We consider the problem of identifying intermediate variables (or mediators) that regulate the effect of a treatment on a response variable. While there has been significant research on this topic, little work has been done when the set of potential mediators is high-dimensional and when they are interrelated. In particular, we assume that the causal structure of the treatment, the potential mediators and the response is a directed acyclic graph (DAG). High-dimensional DAG models have previously been used for the estimation of causal effects from observational data and methods called IDA and joint-IDA have been developed for estimating the effects of single interventions and multiple simultaneous interventions respectively. In this paper, we propose an IDA-type method called MIDA for estimating mediation effects from high-dimensional observational data. Although IDA and joint-IDA estimators have been shown to be consistent in certain sparse high-dimensional settings, their asymptotic properties such as convergence in distribution and inferential tools in such settings remained unknown. We prove high-dimensional consistency of MIDA for linear structural equation models with sub-Gaussian errors. More importantly, we derive distributional convergence results for MIDA in similar high-dimensional settings, which are applicable to IDA and joint-IDA estimators as well. To the best of our knowledge, these are the first distributional convergence results facilitating inference for IDA-type estimators. These results have been built on our novel theoretical results regarding uniform bounds for linear regression estimators over varying subsets of high-dimensional covariates, which may be of independent interest. Finally, we empirically validate our asymptotic theory and demonstrate the usefulness of MIDA in identifying large mediation effects via simulations and application to real data in genomics.

Building deep reinforcement learning agents that can generalize and adapt to unseen environments remains a fundamental challenge for AI. This paper describes progresses on this challenge in the context of man-made environments, which are visually diverse but contain intrinsic semantic regularities. We propose a hybrid model-based and model-free approach, LEArning and Planning with Semantics (LEAPS), consisting of a multi-target sub-policy that acts on visual inputs, and a Bayesian model over semantic structures. When placed in an unseen environment, the agent plans with the semantic model to make high-level decisions, proposes the next sub-target for the sub-policy to execute, and updates the semantic model based on new observations. We perform experiments in visual navigation tasks using House3D, a 3D environment that contains diverse human-designed indoor scenes with real-world objects. LEAPS outperforms strong baselines that do not explicitly plan using the semantic content. Deep reinforcement learning (DRL) has undoubtedly witnessed strong achievements in recent years (Silver et al., 2016; Mnih et al., 2015; Levine et al., 2016).

This document is designed to be a first-year graduate-level introduction to probabilistic programming. It not only provides a thorough background for anyone wishing to use a probabilistic programming system, but also introduces the techniques needed to design and build these systems. It is aimed at people who have an undergraduate-level understanding of either or, ideally, both probabilistic machine learning and programming languages. We start with a discussion of model-based reasoning and explain why conditioning as a foundational computation is central to the fields of probabilistic machine learning and artificial intelligence. We then introduce a simple first-order probabilistic programming language (PPL) whose programs define static-computation-graph, finite-variable-cardinality models. In the context of this restricted PPL we introduce fundamental inference algorithms and describe how they can be implemented in the context of models denoted by probabilistic programs. In the second part of this document, we introduce a higher-order probabilistic programming language, with a functionality analogous to that of established programming languages. This affords the opportunity to define models with dynamic computation graphs, at the cost of requiring inference methods that generate samples by repeatedly executing the program. Foundational inference algorithms for this kind of probabilistic programming language are explained in the context of an interface between program executions and an inference controller. This document closes with a chapter on advanced topics which we believe to be, at the time of writing, interesting directions for probabilistic programming research; directions that point towards a tight integration with deep neural network research and the development of systems for next-generation artificial intelligence applications.

In this blog we will try to get the basic idea behind reinforcement learning and understand what is a multi arm bandit problem. We will also be trying to maximise CTR(click through rate) for advertisements for a advertising agency. Article includes: 1. Basics of reinforcement learning 2. Types of problems in reinforcement learning 3. Understamding multi-arm bandit problem 4. Basics of conditional probability and Thompson sampling 5. Optimizing ads CTR using Thompson sampling in R Reinforcement Learning Basics Reinforcement learning refers to goal-oriented algorithms, which learn how to attain a complex objective (goal) or maximise along a particular dimension over many steps; for example, maximise the points won in a game over many moves. They can start from a blank slate, and under the right conditions, they achieve superhuman performance. Like a child incentivized by spankings and candy, these algorithms are penalized when they make the wrong decisions and rewarded when they make the right ones -- this is reinforcement.

Abstract--Principal component analysis (PCA) is a popular method for projecting data onto uncorrelated components in lower dimension, although the optimal number of components is not specified. Likewise, multiple signal classification (MUSIC) algorithm is a popular PCA-based method for estimating directions of arrival (DOAs) of sinusoidal sources, yet it requires the number of sources to be known a priori. The accurate estimation of the number of sources is hence a crucial issue for performance of these algorithms. In this paper, we will show that both PCA and MUSIC actually return the exact joint maximum-a-posteriori (MAP) estimate for uncorrelated steering vectors, although they can only compute this MAP estimate approximately in correlated case. We then use Bayesian method to, for the first time, compute the MAP estimate for the number of sources in PCA and MUSIC algorithms. Intuitively, this MAP estimate corresponds to the highest probability that signal-plus- noise's variance still dominates projected noise's variance on signal subspace. In simulations of overlapping multi-tone sources for linear sensor array, our exact MAP estimate is far superior to the asymptotic Akaike information criterion (AIC), which is a popular method for estimating the number of components in PCA and MUSIC algorithms. In many systems of array signal processing, e.g. in radar, sonar and antenna systems, linear sensor array is the most basic and universal mathematical model. Because far distant sources with different directions of arrival (DOAs) will oscillate the steering sensor array with different angular frequencies, the array's output data is then a superposition of sinusoidal signals [1]. Hence, a common problem of array systems is to detect the number of sources, as well as their tone frequencies and DOAs, from noisy sinusoidal signals. In literature, most papers only consider the case of single-tone or narrowband sources (i.e. When the number of sources is small, the DOA's line spectra are sparse and can be estimated effectively via sparse techniques like atomic norm (also known as total variation norm) [1], [2], LASSO [4], [5] and Bayesian compressed sensing [6], [7].

We propose a new way of thinking about deep neural networks, in which the linear and non-linear components of the network are naturally derived and justified in terms of principles in probability theory. In particular, the models constructed in our framework assign probabilities to uncertain realizations, leading to Kullback-Leibler Divergence (KLD) as the linear layer. In our model construction, we also arrive at a structure similar to ReLU activation supported with Bayes' theorem. The non-linearities in our framework are normalization layers with ReLU and Sigmoid as element-wise approximations. Additionally, the pooling function is derived as a marginalization of spatial random variables according to the mechanics of the framework. As such, Max Pooling is an approximation to the aforementioned marginalization process. Since our models are comprised of finite state distributions (FSD) as variables and parameters, exact computation of information-theoretic quantities such as entropy and KLD is possible, thereby providing more objective measures to analyze networks. Unlike existing designs that rely on heuristics, the proposed framework restricts subjective interpretations of CNNs and sheds light on the functionality of neural networks from a completely new perspective.

Question: Why do you square the error in a regression machine learning task? Ans: "Why, of course, it turns out all the errors (residuals) into positive quantities!" Question: "OK, why not use a simpler absolute value function x to make all the errors positive?" Ans: "Aha, you are trying to trick me. Absolute value function is not differentiable everywhere!" Question: "That should not matter much for numerical algorithms. LASSO regression uses a term with absolute value and it can be handled.

This article follows my previous one on Bayesian probability & probabilistic programming that I published few months ago on LinkedIn. And for the purpose of this article, I am going to assume that most this article readers have some idea what a Neural Network or Artificial Neural Network is. Neural Network is a non-linear function approximator. We can think of it as a parameterized function where the parameters are the weights & biases of Neural Network through which we will be typically passing our data (inputs), that will be converted to a probability between 0 and 1, to some kind of non-linearity such as a sigmoid function and help make our predictions or estimations. These non-linear functions can be composed together hence Deep Learning Neural Network with multiple layers of this function compositions.

With the availability of vast amounts of user visitation history on location-based social networks (LBSN), the problem of Point-of-Interest (POI) prediction has been extensively studied. However, much of the research has been conducted solely on voluntary checkin datasets collected from social apps such as Foursquare or Yelp. While these data contain rich information about recreational activities (e.g., restaurants, nightlife, and entertainment), information about more prosaic aspects of people's lives is sparse. This not only limits our understanding of users' daily routines, but more importantly the modeling assumptions developed based on characteristics of recreation-based data may not be suitable for richer check-in data. In this work, we present an analysis of education "check-in" data using WiFi access logs collected at Purdue University. We propose a heterogeneous graph-based method to encode the correlations between users, POIs, and activities, and then jointly learn embeddings for the vertices. We evaluate our method compared to previous state-of-the-art POI prediction methods, and show that the assumptions made by previous methods significantly degrade performance on our data with dense(r) activity signals. We also show how our learned embeddings could be used to identify similar students (e.g., for friend suggestions).

Abstract--When selecting a classification algorithm to be applied to a particular problem, one has to simultaneously select the best algorithm for that dataset and the best set of hyperparameters for the chosen model. The usual approach is to apply a nested cross-validation procedure; hyperparameter selection is performed in the inner crossvalidation, while the outer cross-validation computes an unbiased estimate of the expected accuracy of the algorithm with cross-validation based hyperparameter tuning. The alternative approach, which we shall call "flat cross-validation", uses a single cross-validation step both to select the optimal hyperparameter values and to provide an estimate of the expected accuracy of the algorithm, that while biased may nevertheless still be used to select the best learning algorithm. We tested both procedures using 12 different algorithms on 115 real life binary datasets and conclude that using the less computationally expensive flat crossvalidation procedure will generally result in the selection of an algorithm that is, for all practical purposes, of similar quality to that selected via nested cross-validation, provided the learning algorithms have relatively few hyperparameters to be optimised. A practitioner who builds a classification model has to select the best algorithm for that particular problem. There are hundreds of classification algorithms described in the literature, such as k-nearest neighbour [1], SVM [2], neural networks [3], naïve Bayes [4], gradient boosting machines [5], and so on.