For a $d$-dimensional log-concave distribution $\pi(\theta)\propto e^{-f(\theta)}$ on a polytope $K$, we consider the problem of outputting samples from a distribution $\nu$ which is $O(\varepsilon)$-close in infinity-distance $\sup_{\theta\in K}|\log\frac{\nu(\theta)}{\pi(\theta)}|$ to $\pi$. Such samplers with infinity-distance guarantees are specifically desired for differentially private optimization as traditional sampling algorithms which come with total-variation distance or KL divergence bounds are insufficient to guarantee differential privacy. Our main result is an algorithm that outputs a point from a distribution $O(\varepsilon)$-close to $\pi$ in infinity-distance and requires $O((md+dL^2R^2)\times(LR+d\log(\frac{Rd+LRd}{\varepsilon r}))\times md^{\omega-1})$ arithmetic operations, where $f$ is $L$-Lipschitz, $K$ is defined by $m$ inequalities, is contained in a ball of radius $R$ and contains a ball of smaller radius $r$, and $\omega$ is the matrix-multiplication constant. In particular this runtime is logarithmic in $\frac{1}{\varepsilon}$ and significantly improves on prior works. Technically, we depart from the prior works that construct Markov chains on a $\frac{1}{\varepsilon^2}$-discretization of $K$ to achieve a sample with $O(\varepsilon)$ infinity-distance error, and present a method to convert continuous samples from $K$ with total-variation bounds to samples with infinity bounds. To achieve improved dependence on $d$, we present a "soft-threshold" version of the Dikin walk which may be of independent interest. Plugging our algorithm into the framework of the exponential mechanism yields similar improvements in the running time of $\varepsilon$-pure differentially private algorithms for optimization problems such as empirical risk minimization of Lipschitz-convex functions and low-rank approximation, while still achieving the tightest known utility bounds.

The fairness in machine learning is getting increasing attention, as its applications in different fields continue to expand and diversify. To mitigate the discriminated model behaviors between different demographic groups, we introduce a novel post-processing method to optimize over multiple fairness constraints through group-aware threshold adaptation. We propose to learn adaptive classification thresholds for each demographic group by optimizing the confusion matrix estimated from the probability distribution of a classification model output. As we only need an estimated probability distribution of model output instead of the classification model structure, our post-processing model can be applied to a wide range of classification models and improve fairness in a model-agnostic manner and ensure privacy. This even allows us to post-process existing fairness methods to further improve the trade-off between accuracy and fairness. Moreover, our model has low computational cost. We provide rigorous theoretical analysis on the convergence of our optimization algorithm and the trade-off between accuracy and fairness of our method. Our method theoretically enables a better upper bound in near optimality than existing method under same condition. Experimental results demonstrate that our method outperforms state-of-the-art methods and obtains the result that is closest to the theoretical accuracy-fairness trade-off boundary.

We present a sound and complete algorithm, called iterative causal discovery (ICD), for recovering causal graphs in the presence of latent confounders and selection bias. ICD relies on the causal Markov and faithfulness assumptions and recovers the equivalence class of the underlying causal graph. It starts with a complete graph, and consists of a single iterative stage that gradually refines this graph by identifying conditional independence (CI) between connected nodes. Independence and causal relations entailed after any iteration are correct, rendering ICD anytime. Essentially, we tie the size of the CI conditioning set to its distance on the graph from the tested nodes, and increase this value in the successive iteration. Thus, each iteration refines a graph that was recovered by previous iterations having smaller conditioning sets -- a higher statistical power -- which contributes to stability. We demonstrate empirically that ICD requires significantly fewer CI tests and learns more accurate causal graphs compared to FCI, FCI+, and RFCI algorithms.

University campuses are essentially a microcosm of a city. They comprise diverse facilities such as residences, sport centres, lecture theatres, parking spaces, and public transport stops. Universities are under constant pressure to improve efficiencies while offering a better experience to various stakeholders including students, staff, and visitors. Nonetheless, anecdotal evidence indicates that campus assets are not being utilised efficiently, often due to the lack of data collection and analysis, thereby limiting the ability to make informed decisions on the allocation and management of resources. Advances in the Internet of Things (IoT) technologies that can sense and communicate data from the physical world, coupled with data analytics and Artificial intelligence (AI) that can predict usage patterns, have opened up new opportunities for organisations to lower cost and improve user experience. This thesis explores this opportunity via theory and experimentation using UNSW Sydney as a living laboratory.

A Neyman-Scott process is a special case of a Cox process. The latent and observable stochastic processes are both Poisson processes. We consider a deep Neyman-Scott process in this paper, for which the building components of a network are all Poisson processes. We develop an efficient posterior sampling via Markov chain Monte Carlo and use it for likelihood-based inference. Our method opens up room for the inference in sophisticated hierarchical point processes. We show in the experiments that more hidden Poisson processes brings better performance for likelihood fitting and events types prediction. We also compare our method with state-of-the-art models for temporal real-world datasets and demonstrate competitive abilities for both data fitting and prediction, using far fewer parameters.

We study risk-sensitive reinforcement learning (RL) based on the entropic risk measure. Although existing works have established non-asymptotic regret guarantees for this problem, they leave open an exponential gap between the upper and lower bounds. We identify the deficiencies in existing algorithms and their analysis that result in such a gap. To remedy these deficiencies, we investigate a simple transformation of the risk-sensitive Bellman equations, which we call the exponential Bellman equation. The exponential Bellman equation inspires us to develop a novel analysis of Bellman backup procedures in risk-sensitive RL algorithms, and further motivates the design of a novel exploration mechanism. We show that these analytic and algorithmic innovations together lead to improved regret upper bounds over existing ones.

Episodic control enables sample efficiency in reinforcement learning by recalling past experiences from an episodic memory. We propose a new model-based episodic memory of trajectories addressing current limitations of episodic control. Our memory estimates trajectory values, guiding the agent towards good policies. Built upon the memory, we construct a complementary learning model via a dynamic hybrid control unifying model-based, episodic and habitual learning into a single architecture. Experiments demonstrate that our model allows significantly faster and better learning than other strong reinforcement learning agents across a variety of environments including stochastic and non-Markovian settings.

The use of automated decision support and decision-making systems (ADM) (Hardt, Price, and Srebro 2016) in applications with direct impact on people's lives has increasingly become a fact of life, e,g. in criminal justice (Kleinberg, Contributions. We propose a technique that uses Bias Mullainathan, and Raghavan 2016; Jain et al. 2020b; Dressel Parity Score (BPS) measures to characterize fairness and develop and Farid 2018), medical diagnosis (Kleinberg, Mullainathan, a family of corresponding loss functions that are used and Raghavan 2016; Ahsen, Ayvaci, and Raghunathan as regularizers during training of Neural Networks to enhance 2019), insurance (Baudry and Robert 2019), credit fairness of the trained models. The goal here is to permit card fraud detection (Dal Pozzolo et al. 2014), electronic the system to actively pursue fair solutions during training health record data (Gianfrancesco et al. 2018), credit scoring while maintaining as high a performance on the task as (Huang, Chen, and Wang 2007) and many more diverse possible. We apply the approach in the context of several domains. This, in turn, has lead to an urgent need fairness measures and investigate multiple loss function formulations for study and scrutiny of the bias-magnifying effects of machine and regularization weights in order to study the learning and Artificial Intelligence algorithms and thus performance as well as potential drawbacks and deployment their potential to introduce and emphasize social inequalities considerations. In these experiments we show that, if used and systematic discrimination in our society. Appropriately, with appropriate settings, the technique measurably reduces much research is being done currently to mitigate bias race-based bias in recidivism prediction, and demonstrate in AI-based decision support systems (Ahsen, Ayvaci, and on the gender-based Adult Income dataset that the proposed Raghunathan 2019; Kleinberg, Mullainathan, and Raghavan method can outperform state-of-the art techniques aimed at 2016; Noriega-Campero et al. 2019; Feldman 2015; more targeted aspects of bias and fairness.

Bayesian optimization (BO) is an efficient method to optimize expensive black-box functions. It has been generalized to scenarios where objective function evaluations return stochastic binary feedback, such as success/failure in a given test, or preference between different parameter settings. In many real-world situations, the objective function can be evaluated in controlled 'contexts' or 'environments' that directly influence the observations. For example, one could directly alter the 'difficulty' of the test that is used to evaluate a system's performance. With binary feedback, the context determines the information obtained from each observation. For example, if the test is too easy/hard, the system will always succeed/fail, yielding uninformative binary outputs. Here we combine ideas from Bayesian active learning and optimization to efficiently choose the best context and optimization parameter on each iteration. We demonstrate the performance of our algorithm and illustrate how it can be used to tackle a concrete application in visual psychophysics: efficiently improving patients' vision via corrective lenses, using psychophysics measurements.

Ad hoc teamwork problem describes situations where an agent has to cooperate with previously unseen agents to achieve a common goal. For an agent to be successful in these scenarios, it has to have a suitable cooperative skill. One could implement cooperative skills into an agent by using domain knowledge to design the agent's behavior. However, in complex domains, domain knowledge might not be available. Therefore, it is worthwhile to explore how to directly learn cooperative skills from data. In this work, we apply meta-reinforcement learning (meta-RL) formulation in the context of the ad hoc teamwork problem. Our empirical results show that such a method could produce robust cooperative agents in two cooperative environments with different cooperative circumstances: social compliance and language interpretation. (This is a full paper of the extended abstract version.)