Meta-learning is a line of research that develops the ability to leverage past experiences to efficiently solve new learning problems. Meta-Reinforcement Learning (meta-RL) methods demonstrate a capability to learn behaviors that efficiently acquire and exploit information in several meta-RL problems. In this context, the Alchemy benchmark has been proposed by Wang et al. [2021]. Alchemy features a rich structured latent space that is challenging for state-of-the-art model-free RL methods. These methods fail to learn to properly explore then exploit. We develop a model-based algorithm. We train a model whose principal block is a Transformer Encoder to fit the symbolic Alchemy environment dynamics. Then we define an online planner with the learned model using a tree search method. This algorithm significantly outperforms previously applied model-free RL methods on the symbolic Alchemy problem. Our results reveal the relevance of model-based approaches with online planning to perform exploration and exploitation successfully in meta-RL. Moreover, we show the efficiency of the Transformer architecture to learn complex dynamics that arise from latent spaces present in meta-RL problems.

The evaluation of the free energy of a stochastic model is considered a significant issue in various fields of physics and machine learning. However, the exact free energy evaluation is computationally infeasible because the free energy expression includes an intractable partition function. Annealed importance sampling (AIS) is a type of importance sampling based on the Markov chain Monte Carlo method that is similar to a simulated annealing and can effectively approximate the free energy. This study proposes an AIS-based approach, which is referred to as marginalized AIS (mAIS). The statistical efficiency of mAIS is investigated in detail based on theoretical and numerical perspectives. Based on the investigation, it is proved that mAIS is more effective than AIS under a certain condition.

We study offline reinforcement learning under a novel model called strategic MDP, which characterizes the strategic interactions between a principal and a sequence of myopic agents with private types. Due to the bilevel structure and private types, strategic MDP involves information asymmetry between the principal and the agents. We focus on the offline RL problem, where the goal is to learn the optimal policy of the principal concerning a target population of agents based on a pre-collected dataset that consists of historical interactions. The unobserved private types confound such a dataset as they affect both the rewards and observations received by the principal. We propose a novel algorithm, Pessimistic policy Learning with Algorithmic iNstruments (PLAN), which leverages the ideas of instrumental variable regression and the pessimism principle to learn a near-optimal principal's policy in the context of general function approximation. Our algorithm is based on the critical observation that the principal's actions serve as valid instrumental variables. In particular, under a partial coverage assumption on the offline dataset, we prove that PLAN outputs a $1 / \sqrt{K}$-optimal policy with $K$ being the number of collected trajectories. We further apply our framework to some special cases of strategic MDP, including strategic regression, strategic bandit, and noncompliance in recommendation systems.

User simulators (USs) are commonly used to train task-oriented dialogue systems (DSs) via reinforcement learning. The interactions often take place on semantic level for efficiency, but there is still a gap from semantic actions to natural language, which causes a mismatch between training and deployment environment. Incorporating a natural language generation (NLG) module with USs during training can partly deal with this problem. However, since the policy and NLG of USs are optimised separately, these simulated user utterances may not be natural enough in a given context. In this work, we propose a generative transformer-based user simulator (GenTUS). GenTUS consists of an encoder-decoder structure, which means it can optimise both the user policy and natural language generation jointly. GenTUS generates both semantic actions and natural language utterances, preserving interpretability and enhancing language variation. In addition, by representing the inputs and outputs as word sequences and by using a large pre-trained language model we can achieve generalisability in feature representation. We evaluate GenTUS with automatic metrics and human evaluation. Our results show that GenTUS generates more natural language and is able to transfer to an unseen ontology in a zero-shot fashion. In addition, its behaviour can be further shaped with reinforcement learning opening the door to training specialised user simulators.

Accurate, detailed, and timely crop type mapping is a very valuable information for the institutions in order to create more accurate policies according to the needs of the citizens. In the last decade, the amount of available data dramatically increased, whether it can come from Remote Sensing (using Copernicus Sentinel-2 data) or directly from the farmers (providing in-situ crop information throughout the years and information on crop rotation). Nevertheless, the majority of the studies are restricted to the use of one modality (Remote Sensing data or crop rotation) and never fuse the Earth Observation data with domain knowledge like crop rotations. Moreover, when they use Earth Observation data they are mainly restrained to one year of data, not taking into account the past years. In this context, we propose to tackle a land use and crop type classification task using three data types, by using a Hierarchical Deep Learning algorithm modeling the crop rotations like a language model, the satellite signals like a speech signal and using the crop distribution as additional context vector. We obtained very promising results compared to classical approaches with significant performances, increasing the Accuracy by 5.1 points in a 28-class setting (.948), and the micro-F1 by 9.6 points in a 10-class setting (.887) using only a set of crop of interests selected by an expert. We finally proposed a data-augmentation technique to allow the model to classify the crop before the end of the season, which works surprisingly well in a multimodal setting.

We study the problem of globally optimizing the causal effect on a target variable of an unknown causal graph in which interventions can be performed. This problem arises in many areas of science including biology, operations research and healthcare. We propose Causal Entropy Optimization (CEO), a framework that generalizes Causal Bayesian Optimization (CBO) to account for all sources of uncertainty, including the one arising from the causal graph structure. CEO incorporates the causal structure uncertainty both in the surrogate models for the causal effects and in the mechanism used to select interventions via an information-theoretic acquisition function. The resulting algorithm automatically trades-off structure learning and causal effect optimization, while naturally accounting for observation noise. For various synthetic and real-world structural causal models, CEO achieves faster convergence to the global optimum compared with CBO while also learning the graph. Furthermore, our joint approach to structure learning and causal optimization improves upon sequential, structure-learning-first approaches.

Causal graph, as an effective and powerful tool for causal modeling, is usually assumed as a Directed Acyclic Graph (DAG). However, recommender systems usually involve feedback loops, defined as the cyclic process of recommending items, incorporating user feedback in model updates, and repeating the procedure. As a result, it is important to incorporate loops into the causal graphs to accurately model the dynamic and iterative data generation process for recommender systems. However, feedback loops are not always beneficial since over time they may encourage more and more narrowed content exposure, which if left unattended, may results in echo chambers. As a result, it is important to understand when the recommendations will lead to echo chambers and how to mitigate echo chambers without hurting the recommendation performance. In this paper, we design a causal graph with loops to describe the dynamic process of recommendation. We then take Markov process to analyze the mathematical properties of echo chamber such as the conditions that lead to echo chambers. Inspired by the theoretical analysis, we propose a Dynamic Causal Collaborative Filtering ($\partial$CCF) model, which estimates users' post-intervention preference on items based on back-door adjustment and mitigates echo chamber with counterfactual reasoning. Multiple experiments are conducted on real-world datasets and results show that our framework can mitigate echo chambers better than other state-of-the-art frameworks while achieving comparable recommendation performance with the base recommendation models.

Machine learning algorithms are based on correlation – they do not specify cause and effect relations. A Bayesian network is a probabilistic graphical model representing a set of variables and their conditional dependencies via a directed acyclic graph (DAG). Bayesian networks are used for causal inference because they can consider an event and predict the likelihood of several causes that could contribute to the event's occurrence. For example, you could model a disease and its symptoms in a Bayesian network such that you could predict the disease given a set of symptoms. Bayesian networks are directed acyclic graphs (DAGs) whose nodes represent Bayesian variables (as observable quantities, latent variables, unknown parameters or hypotheses).

Event-triggered model predictive control (eMPC) is a popular optimal control method with an aim to alleviate the computation and/or communication burden of MPC. However, it generally requires priori knowledge of the closed-loop system behavior along with the communication characteristics for designing the event-trigger policy. This paper attempts to solve this challenge by proposing an efficient eMPC framework and demonstrate successful implementation of this framework on the autonomous vehicle path following. First of all, a model-free reinforcement learning (RL) agent is used to learn the optimal event-trigger policy without the need for a complete dynamical system and communication knowledge in this framework. Furthermore, techniques including prioritized experience replay (PER) buffer and long-short term memory (LSTM) are employed to foster exploration and improve training efficiency. In this paper, we use the proposed framework with three deep RL algorithms, i.e., Double Q-learning (DDQN), Proximal Policy Optimization (PPO), and Soft Actor-Critic (SAC), to solve this problem. Experimental results show that all three deep RL-based eMPC (deep-RL-eMPC) can achieve better evaluation performance than the conventional threshold-based and previous linear Q-based approach in the autonomous path following. In particular, PPO-eMPC with LSTM and DDQN-eMPC with PER and LSTM obtains a superior balance between the closed-loop control performance and event-trigger frequency. The associated code is open-sourced and available at: https://github.com/DangFengying/RL-based-event-triggered-MPC.

Safe exploration is a common problem in reinforcement learning (RL) that aims to prevent agents from making disastrous decisions while exploring their environment. A family of approaches to this problem assume domain knowledge in the form of a (partial) model of this environment to decide upon the safety of an action. A so-called shield forces the RL agent to select only safe actions. However, for adoption in various applications, one must look beyond enforcing safety and also ensure the applicability of RL with good performance. We extend the applicability of shields via tight integration with state-of-the-art deep RL, and provide an extensive, empirical study in challenging, sparse-reward environments under partial observability. We show that a carefully integrated shield ensures safety and can improve the convergence rate and final performance of RL agents. We furthermore show that a shield can be used to bootstrap state-of-the-art RL agents: they remain safe after initial learning in a shielded setting, allowing us to disable a potentially too conservative shield eventually.