KNN can easily be mapped to our real lives. If you want to learn about a person, of whom you have no information, you might like to find out about his close friends and the circles he moves in and gain access to his/her information!

In my recent blog, Marrying Kalman Filtering & Machine Learning, we saw the merger of Bayesian exact recursive estimation (algorithm for which is Kalman Filter/Smoother in the linear, Gaussian case) and Machine Learning. We developed a solution called Kernel Projection Kalman Filter for business applications that require static or dynamical, dynamical or time-varying dynamical, linear or non-linear Machine Learning, i.e., pretty much all applications - therefore, Kernel Projection Kalman Filter is a "universal" solution . . . But who needs anything more than STATIC Machine Learning (ML)? Indeed, university courses in ML largely teach static ML. Given a set of inputs and outputs, find a static map between the two during supervised "Training" and use this static map for business purposes during "Operation" (which is called "Testing" during pre-operation evaluation).

Modeling causal dependencies often demands cycles at a coarse-grained temporal scale. If Bayesian networks are to be used for modeling uncertainties, cycles are eliminated with dynamic Bayesian networks, spreading indirect dependencies over time and enforcing an infinitesimal resolution of time. Without a ``causal design,'' i.e., without anticipating indirect influences appropriately in time, we argue that such networks return spurious results. By identifying activator random variables, we propose activator dynamic Bayesian networks (ADBNs) which are able to rapidly adapt to contexts under a causal use of time, anticipating indirect influences on a solid mathematical basis using familiar Bayesian network semantics. ADBNs are well-defined dynamic probabilistic graphical models allowing one to model cyclic dependencies from local and causal perspectives while preserving a classical, familiar calculus and classically known algorithms, without introducing any overhead in modeling or inference.

In this paper, we consider an infinite dimensional exponential family, $\mathcal{P}$ of probability densities, which are parametrized by functions in a reproducing kernel Hilbert space, $H$ and show it to be quite rich in the sense that a broad class of densities on $\mathbb{R}^d$ can be approximated arbitrarily well in Kullback-Leibler (KL) divergence by elements in $\mathcal{P}$. The main goal of the paper is to estimate an unknown density, $p_0$ through an element in $\mathcal{P}$. Standard techniques like maximum likelihood estimation (MLE) or pseudo MLE (based on the method of sieves), which are based on minimizing the KL divergence between $p_0$ and $\mathcal{P}$, do not yield practically useful estimators because of their inability to efficiently handle the log-partition function. Instead, we propose an estimator, $\hat{p}_n$ based on minimizing the \emph{Fisher divergence}, $J(p_0\Vert p)$ between $p_0$ and $p\in \mathcal{P}$, which involves solving a simple finite-dimensional linear system. When $p_0\in\mathcal{P}$, we show that the proposed estimator is consistent, and provide a convergence rate of $n^{-\min\left\{\frac{2}{3},\frac{2\beta+1}{2\beta+2}\right\}}$ in Fisher divergence under the smoothness assumption that $\log p_0\in\mathcal{R}(C^\beta)$ for some $\beta\ge 0$, where $C$ is a certain Hilbert-Schmidt operator on $H$ and $\mathcal{R}(C^\beta)$ denotes the image of $C^\beta$. We also investigate the misspecified case of $p_0\notin\mathcal{P}$ and show that $J(p_0\Vert\hat{p}_n)\rightarrow \inf_{p\in\mathcal{P}}J(p_0\Vert p)$ as $n\rightarrow\infty$, and provide a rate for this convergence under a similar smoothness condition as above. Through numerical simulations we demonstrate that the proposed estimator outperforms the non-parametric kernel density estimator, and that the advantage with the proposed estimator grows as $d$ increases.

The Bayesian community should really start going to ICLR. They really should have started going years ago. For too long we Bayesians have, quite arrogantly, dismissed deep neural networks as unprincipled, dumb black boxes that lack elegance. We said that highly over-parametrised models fitted via maximum likelihood can't possibly work, they will overfit, won't generalise, etc. We touted our Bayesian nonparametric models instead: Chinese restaurants, Indian buffets, Gaussian processes. And, when things started looking really dire for us Bayesians, we even formed an alliance with kernel people, who used to be our mortal enemies just years before because they like convex optimisation.

Innovation diffusion has been studied extensively in a variety of disciplines, including sociology, economics, marketing, ecology, and computer science. Traditional literature on innovation diffusion has been dominated by models of aggregate behavior and trends. However, the agent-based modeling (ABM) paradigm is gaining popularity as it captures agent heterogeneity and enables fine-grained modeling of interactions mediated by social and geographic networks. While most ABM work on innovation diffusion is theoretical, empirically grounded models are increasingly important, particularly in guiding policy decisions. We present a critical review of empirically grounded agent-based models of innovation diffusion, developing a categorization of this research based on types of agent models as well as applications. By connecting the modeling methodologies in the fields of information and innovation diffusion, we suggest that the maximum likelihood estimation framework widely used in the former is a promising paradigm for calibration of agent-based models for innovation diffusion. Although many advances have been made to standardize ABM methodology, we identify four major issues in model calibration and validation, and suggest potential solutions.

Process mining techniques focus on extracting insight in processes from event logs. Process mining has the potential to provide valuable insights in (un)healthy habits and to contribute to ambient assisted living solutions when applied on data from smart home environments. However, events recorded in smart home environments are on the level of sensor triggers, at which process discovery algorithms produce overgeneralizing process models that allow for too much behavior and that are difficult to interpret for human experts. We show that abstracting the events to a higher-level interpretation can enable discovery of more precise and more comprehensible models. We present a framework for the extraction of features that can be used for abstraction with supervised learning methods that is based on the XES IEEE standard for event logs. This framework can automatically abstract sensor-level events to their interpretation at the human activity level, after training it on training data for which both the sensor and human activity events are known. We demonstrate our abstraction framework on three real-life smart home event logs and show that the process models that can be discovered after abstraction are more precise indeed.

We empirically characterize the performance of discriminative and generative LSTM models for text classification. We find that although RNN-based generative models are more powerful than their bag-of-words ancestors (e.g., they account for conditional dependencies across words in a document), they have higher asymptotic error rates than discriminatively trained RNN models. However we also find that generative models approach their asymptotic error rate more rapidly than their discriminative counterparts---the same pattern that Ng & Jordan (2001) proved holds for linear classification models that make more naive conditional independence assumptions. Building on this finding, we hypothesize that RNN-based generative classification models will be more robust to shifts in the data distribution. This hypothesis is confirmed in a series of experiments in zero-shot and continual learning settings that show that generative models substantially outperform discriminative models.

We introduce dropout compaction, a novel method for training feed-forward neural networks which realizes the performance gains of training a large model with dropout regularization, yet extracts a compact neural network for run-time efficiency. In the proposed method, we introduce a sparsity-inducing prior on the per unit dropout retention probability so that the optimizer can effectively prune hidden units during training. By changing the prior hyperparameters, we can control the size of the resulting network. We performed a systematic comparison of dropout compaction and competing methods on several real-world speech recognition tasks and found that dropout compaction achieved comparable accuracy with fewer than 50% of the hidden units, translating to a 2.5x speedup in run-time.

The volume of data generated by internet and social networks is increasing every day, and there is a clear need for efficient ways of extracting useful information from them. As those data can take different forms, it is important to use all the available data representations for prediction. In this paper, we focus our attention on supervised classification using both regular plain, tabular, data and structural information coming from a network structure. 14 techniques are investigated and compared in this study and can be divided in three classes: the first one uses only the plain data to build a classification model, the second uses only the graph structure and the last uses both information sources. The relative performances in these three cases are investigated. Furthermore, the effect of using a graph embedding and well-known indicators in spatial statistics is also studied. Possible applications are automatic classification of web pages or other linked documents, of people in a social network or of proteins in a biological complex system, to name a few. Based on our comparison, we draw some general conclusions and advices to tackle this particular classification task: some datasets can be better explained by their graph structure (graph-driven), or by their feature set (features-driven). The most efficient methods are discussed in both cases.