At its WWDC 2017 keynote on Monday, Apple showed off the fruits of its AI research labors. We saw a Siri assistant that's smart enough to interpret your intentions, an updated Metal 2 graphics suite designed for machine learning and a Photos app that can do everything its Google rival does without an internet connection. Being at the front of the AI pack is a new position for Apple to find itself in. Despite setting off the AI arms race when it introduced Siri in 2010, Apple has long lagged behind its competitors in this field. It's amazing what a year of intense R&D can do. Well, technically, it's been three years of R&D, but Apple had a bit of trouble getting out of its own way for the first two.

This is part two of a blog post on probabilistic programming. The first part of the blog can be found here. Markov chains are mathematical constructs with a wide range of applications in physics, mathematical biology, speech recognition, statistics and many others. The simplest way to think about them is considering the above animation. A person (the circle) is trying to find out where their friend lives in a neighbourhood block.

In several realistic situations, an interactive learning agent can practice and refine its strategy before going on to be evaluated. For instance, consider a student preparing for a series of tests. She would typically take a few practice tests to know which areas she needs to improve upon. Based of the scores she obtains in these practice tests, she would formulate a strategy for maximizing her scores in the actual tests. We treat this scenario in the context of an agent exploring a fixed-horizon episodic Markov Decision Process (MDP), where the agent can practice on the MDP for some number of episodes (not necessarily known in advance) before starting to incur regret for its actions. During practice, the agent's goal must be to maximize the probability of following an optimal policy. This is akin to the problem of Pure Exploration (PE). We extend the PE problem of Multi Armed Bandits (MAB) to MDPs and propose a Bayesian algorithm called Posterior Sampling for Pure Exploration (PSPE), which is similar to its bandit counterpart. We show that the Bayesian simple regret converges at an optimal exponential rate when using PSPE. When the agent starts being evaluated, its goal would be to minimize the cumulative regret incurred. This is akin to the problem of Reinforcement Learning (RL). The agent uses the Posterior Sampling for Reinforcement Learning algorithm (PSRL) initialized with the posteriors of the practice phase. We hypothesize that this PSPE + PSRL combination is an optimal strategy for minimizing regret in RL problems with an initial practice phase. We show empirical results which prove that having a lower simple regret at the end of the practice phase results in having lower cumulative regret during evaluation.

A Monte Carlo algorithm typically simulates some prescribed number of samples, taking some random real time to complete the computations necessary. This work considers the converse: to impose a real-time budget on the computation, so that the number of samples simulated is random. To complicate matters, the real time taken for each simulation may depend on the sample produced, so that the samples themselves are not independent of their number, and a length bias with respect to compute time is apparent. This is especially problematic when a Markov chain Monte Carlo (MCMC) algorithm is used and the final state of the Markov chain---rather than an average over all states---is required. The length bias does not diminish with the compute budget in this case. It occurs, for example, in sequential Monte Carlo (SMC) algorithms. We propose an anytime framework to address the concern, using a continuous-time Markov jump process to study the progress of the computation in real time. We show that the length bias can be eliminated for any MCMC algorithm by using a multiple chain construction. The utility of this construction is demonstrated on a large-scale SMC-squared implementation, using four billion particles distributed across a cluster of 128 graphics processing units on the Amazon EC2 service. The anytime framework imposes a real-time budget on the MCMC move steps within SMC-squared, ensuring that all processors are simultaneously ready for the resampling step, demonstrably reducing wait times and providing substantial control over the total compute budget.

The official title of this free book available in PDF format is Machine Learning Cheat Sheet. See table of content screenshot below. The chapters 17 to 28 (the most interesting ones in my opinion) seem like a work in progress - I'm sure the authors intend to make them a bit bigger. For a more modern and applied book, get Dr Granville's book on data science. And here's the detailed table of content:

We study the problem of online learning in a class of Markov decision processes known as linearly solvable MDPs. In the stationary version of this problem, a learner interacts with its environment by directly controlling the state transitions, attempting to balance a fixed state-dependent cost and a certain smooth cost penalizing extreme control inputs. In the current paper, we consider an online setting where the state costs may change arbitrarily between consecutive rounds, and the learner only observes the costs at the end of each respective round. We are interested in constructing algorithms for the learner that guarantee small regret against the best stationary control policy chosen in full knowledge of the cost sequence. Our main result is showing that the smoothness of the control cost enables the simple algorithm of following the leader to achieve a regret of order $\log^2 T$ after $T$ rounds, vastly improving on the best known regret bound of order $T^{3/4}$ for this setting.

Words then point to shared ideas in our minds [Gärdenfors, 2014]. 5. Language as shared convention A B Wittgenstein and his language games, Philosophical Investigations, 1958 block pillar slab beam Two people building something. Eventually turns into a shared convention for a community 6. Our brains map community conventions to personal sensations and actions sensation representation action • When someone says "beam," we map that to our experience with beams. See: Benjamin Bergen, Steven Pinker, Mark Johnson, Jerome Feldman, and Murray Shanahan "beam" This is what is means for language to be grounded. We negotiate language and meaning as we go Modified from Gärdenfors (2014), which was based on Winter (1998) A fishing pole is a stick, string and hook You can catch fish with a fishing pole Get me some fish Acknowledgement Acknowledgement Break Break • Levels of discourse • Complicated to go up and down the pyramid 9. Conversation has its own rules (pragmatics) • Conversational maxims: Grice (1975, 1978) • Breaking these rules is a way to communicate more than the meaning of the words. Maxim of Quantity: Say only what is not implied. What did she mean by that?

Stochastic Gradient Langevin Dynamics (SGLD) is a popular variant of Stochastic Gradient Descent, where properly scaled isotropic Gaussian noise is added to an unbiased estimate of the gradient at each iteration. This modest change allows SGLD to escape local minima and suffices to guarantee asymptotic convergence to global minimizers for sufficiently regular non-convex objectives (Gelfand and Mitter, 1991). The present work provides a nonasymptotic analysis in the context of non-convex learning problems, giving finite-time guarantees for SGLD to find approximate minimizers of both empirical and population risks. As in the asymptotic setting, our analysis relates the discrete-time SGLD Markov chain to a continuous-time diffusion process. A new tool that drives the results is the use of weighted transportation cost inequalities to quantify the rate of convergence of SGLD to a stationary distribution in the Euclidean $2$-Wasserstein distance.

Manifold Markov chain Monte Carlo algorithms have been introduced to sample more effectively from challenging target densities exhibiting multiple modes or strong correlations. Such algorithms exploit the local geometry of the parameter space, thus enabling chains to achieve a faster convergence rate when measured in number of steps. However, often acquiring local geometric information increases computational complexity per step to the extent that sampling from high-dimensional targets becomes inefficient in terms of total computational time. This paper analyzes the computational complexity of manifold Langevin Monte Carlo and proposes a manifold adaptive Monte Carlo sampler aimed at balancing the benefits of exploiting local geometry with computational requirements to achieve a high effective sample size for a given computational cost. The suggested strategy randomly switches between a local geometric and an adaptive proposal kernel via a schedule to regulate the frequency of manifold-based updates. An exponentially decaying schedule is put forward that enables more frequent updates of geometric information in early transient phases of the chain, while saving computational time in late stationary phases. The average complexity can be manually set depending on the need for geometric exploitation posed by the underlying model.

Web crawling, snowball sampling, and respondent-driven sampling (RDS) are three types of network sampling techniques used to contact individuals in hard-to-reach populations. This paper studies these procedures as a Markov process on the social network that is indexed by a tree. Each node in this tree corresponds to an observation and each edge in the tree corresponds to a referral. Indexing with a tree (instead of a chain) allows for the sampled units to refer multiple future units into the sample. In survey sampling, the design effect characterizes the additional variance induced by a novel sampling strategy. If the design effect is some value $DE$, then constructing an estimator from the novel design makes the variance of the estimator $DE$ times greater than it would be under a simple random sample with the same sample size $n$. Under certain assumptions on the referral tree, the design effect of network sampling has a critical threshold that is a function of the referral rate $m$ and the clustering structure in the social network, represented by the second eigenvalue of the Markov transition matrix, $\lambda_2$. If $m < 1/\lambda_2^2$, then the design effect is finite (i.e. the standard estimator is $\sqrt{n}$-consistent). However, if $m > 1/\lambda_2^2$, then the design effect grows with $n$ (i.e. the standard estimator is no longer $\sqrt{n}$-consistent). Past this critical threshold, the standard error of the estimator converges at the slower rate of $n^{\log_m \lambda_2}$. The Markov model allows for nodes to be resampled; computational results show that the findings hold in without-replacement sampling. To estimate confidence intervals that adapt to the correct level of uncertainty, a novel resampling procedure is proposed. Computational experiments compare this procedure to previous techniques.