This lecture note is intended to prepare early-year master's and PhD students in data science or a related discipline with foundational ideas in machine learning. It starts with basic ideas in modern machine learning with classification as a main target task. These basic ideas include loss formulation, backpropagation, stochastic gradient descent, generalization, model selection as well as fundamental blocks of artificial neural networks. Based on these basic ideas, the lecture note explores in depth the probablistic approach to unsupervised learning, covering directed latent variable models, product of experts, generative adversarial networks and autoregressive models. Finally, the note ends by covering a diverse set of further topics, such as reinforcement learning, ensemble methods and meta-learning. After reading this lecture note, a student should be ready to embark on studying and researching more advanced topics in machine learning and more broadly artificial intelligence.

When the likelihood is analytically unavailable and computationally intractable, approximate Bayesian computation (ABC) has emerged as a widely used methodology for approximate posterior inference; however, it suffers from severe computational inefficiency in high-dimensional settings or under diffuse priors. To overcome these limitations, we propose Adaptive Bayesian Inference (ABI), a framework that bypasses traditional data-space discrepancies and instead compares distributions directly in posterior space through nonparametric distribution matching. By leveraging a novel Marginally-augmented Sliced Wasserstein (MSW) distance on posterior measures and exploiting its quantile representation, ABI transforms the challenging problem of measuring divergence between posterior distributions into a tractable sequence of one-dimensional conditional quantile regression tasks. Moreover, we introduce a new adaptive rejection sampling scheme that iteratively refines the posterior approximation by updating the proposal distribution via generative density estimation. Theoretically, we establish parametric convergence rates for the trimmed MSW distance and prove that the ABI posterior converges to the true posterior as the tolerance threshold vanishes. Through extensive empirical evaluation, we demonstrate that ABI significantly outperforms data-based Wasserstein ABC, summary-based ABC, and state-of-the-art likelihood-free simulators, especially in high-dimensional or dependent observation regimes.

Deep neural networks (DNNs) have emerged as a powerful methodology with significant practical successes in fields such as computer vision and natural language processing. Recent works have demonstrated that sparsely connected DNNs with carefully designed architectures can achieve minimax estimation rates under classical smoothness assumptions. However, subsequent studies revealed that simple fully connected DNNs can achieve comparable convergence rates, challenging the necessity of sparsity. Theoretical advances in Bayesian neural networks (BNNs) have been more fragmented. Much of those work has concentrated on sparse networks, leaving the theoretical properties of fully connected BNNs underexplored. In this paper, we address this gap by investigating fully connected Bayesian DNNs with Gaussian prior using PAC-Bayes bounds. We establish upper bounds on the prediction risk for a probabilistic deep neural network method, showing that these bounds match (up to logarithmic factors) the minimax-optimal rates in Besov space, for both nonparametric regression and binary classification with logistic loss. Importantly, our results hold for a broad class of practical activation functions that are Lipschitz continuous.

To cluster data is to separate samples into distinctive groups that should ideally have some cohesive properties. Today, numerous clustering algorithms exist, and their differences lie essentially in what can be perceived as ``cohesive properties''. Therefore, hypotheses on the nature of clusters must be set: they can be either generative or discriminative. As the last decade witnessed the impressive growth of deep clustering methods that involve neural networks to handle high-dimensional data often in a discriminative manner; we concentrate mainly on the discriminative hypotheses. In this paper, our aim is to provide an accessible historical perspective on the evolution of discriminative clustering methods and notably how the nature of assumptions of the discriminative models changed over time: from decision boundaries to invariance critics. We notably highlight how mutual information has been a historical cornerstone of the progress of (deep) discriminative clustering methods. We also show some known limitations of mutual information and how discriminative clustering methods tried to circumvent those. We then discuss the challenges that discriminative clustering faces with respect to the selection of the number of clusters. Finally, we showcase these techniques using the dedicated Python package, GemClus, that we have developed for discriminative clustering.

Previous research has shown the benefit Bayesian methods can bring to certain problems within deep learning Gal et al. (2017). However, computing the exact posterior distributions of BNNs is a difficult task as traditional methods such as Markov chain Monte Carlo (MCMC) Hastings (1970) are computationally poorly suited to exploring high dimensional spaces and dealing with large amounts of data. Parametric methods such as variational inference are better suited to these difficulties, but only give an approximation to the posterior distribution. These spaces have been found to be highly complex Izmailov et al. (2021a) and therefore variational methods often give a poor approximation of the posterior. Sequential Monte Carlo (SMC) samplers Doucet et al. (2001) are an alternative to MCMC methods which also provide an empirical estimate of the posterior distribution. SMC samplers are instantly parallelisable Varsi et al. (2021b) and therefore can take advantage of the GPU resources commonly used in machine learning to speed up the training process. MCMC methods often require a warm-up period to adapt the hyperparameters, after which the chains can be parallelised. However, the hyperparameters must remain fixed after this warm-up period to obey stationarity. This means that SMC samplers can be more flexible than 1 arXiv:2505.03797v1

Multi-step manipulation tasks where robots interact with their environment and must apply process forces based on the perceived situation remain challenging to learn and prone to execution errors. Accurately simulating these tasks is also difficult. Hence, it is crucial for robust task performance to learn how to coordinate end-effector pose and applied force, monitor execution, and react to deviations. To address these challenges, we propose a learning approach that directly infers both low- and high-level task representations from user demonstrations on the real system. We developed an unsupervised task segmentation algorithm that combines intention recognition and feature clustering to infer the skills of a task. We leverage the inferred characteristic features of each skill in a novel unsupervised anomaly detection approach to identify deviations from the intended task execution. Together, these components form a comprehensive framework capable of incrementally learning task decisions and new behaviors as new situations arise. Compared to state-of-the-art learning techniques, our approach significantly reduces the required amount of training data and computational complexity while efficiently learning complex in-contact behaviors and recovery strategies. Our proposed task segmentation and anomaly detection approaches outperform state-of-the-art methods on force-based tasks evaluated on two different robotic systems.

Robust Markov Decision Processes (RMDPs) generalize classical MDPs that consider uncertainties in transition probabilities by defining a set of possible transition functions. An objective is a set of runs (or infinite trajectories) of the RMDP, and the value for an objective is the maximal probability that the agent can guarantee against the adversarial environment. We consider (a) reachability objectives, where given a target set of states, the goal is to eventually arrive at one of them; and (b) parity objectives, which are a canonical representation for $ω$-regular objectives. The qualitative analysis problem asks whether the objective can be ensured with probability 1. In this work, we study the qualitative problem for reachability and parity objectives on RMDPs without making any assumption over the structures of the RMDPs, e.g., unichain or aperiodic. Our contributions are twofold. We first present efficient algorithms with oracle access to uncertainty sets that solve qualitative problems of reachability and parity objectives. We then report experimental results demonstrating the effectiveness of our oracle-based approach on classical RMDP examples from the literature scaling up to thousands of states.

We study reinforcement learning by combining recent advances in regularized linear programming formulations with the classical theory of stochastic approximation. Motivated by the challenge of designing algorithms that leverage off-policy data while maintaining on-policy exploration, we propose PGDA-RL, a novel primal-dual Projected Gradient Descent-Ascent algorithm for solving regularized Markov Decision Processes (MDPs). PGDA-RL integrates experience replay-based gradient estimation with a two-timescale decomposition of the underlying nested optimization problem. The algorithm operates asynchronously, interacts with the environment through a single trajectory of correlated data, and updates its policy online in response to the dual variable associated with the occupation measure of the underlying MDP. We prove that PGDA-RL converges almost surely to the optimal value function and policy of the regularized MDP. Our convergence analysis relies on tools from stochastic approximation theory and holds under weaker assumptions than those required by existing primal-dual RL approaches, notably removing the need for a simulator or a fixed behavioral policy.

Model-based planning and execution systems offer a principled approach to building flexible autonomous robots that can perform diverse tasks by automatically combining a host of basic skills. This idea is almost as old as modern robotics. Yet, while diverse general-purpose reasoning architectures have been proposed since, general-purpose systems that are integrated with modern robotic platforms have emerged only recently, starting with the influential ROSPlan system. Since then, a growing number of model-based systems for robot task-level control have emerged. In this paper, we consider the diverse design choices and issues existing systems attempt to address, the different solutions proposed so far, and suggest avenues for future development.

Automatic music transcription (AMT) is the problem of analyzing an audio recording of a musical piece and detecting notes that are being played. AMT is a challenging problem, particularly when it comes to polyphonic music. The goal of AMT is to produce a score representation of a music piece, by analyzing a sound signal containing multiple notes played simultaneously. In this work, we design a processing pipeline that can transform classical piano audio files in .wav format into a music score representation. The features from the audio signals are extracted using the constant-Q transform, and the resulting coefficients are used as an input to the convolutional neural network (CNN) model.