Abstract Machine learning predictive models rely on data to make predictions for new input data. However, accurate predictions are not always the end goal; practitioners often aim to make informed decisions through optimization problems (OPs) based on these predictions. While the idea that better predictions lead to better decisions was widely accepted, the latest literature highlights that even small inaccuracies in predictions can lead to poor decisions depending on the structure of the OP. Therefore, recent research has been focused on end-to-end learning approaches that directly improve decision quality without considering prediction accuracy when solving data-driven OPs. Some of these end-to-end learning approaches are mainly called “predict-and-optimize” (PaO), and they aim to learn a predictor based on the quality of the downstream task decisions by incorporating mathematical programming into the learning process. This literature review discusses the variations of and approaches to PaO problems by proposing a unified notation and a taxonomy for them. Throughout the paper, we aim to provide a valuable roadmap for researchers and practitioners in the field, guiding them to choose data-driven methods to solve their decision problems effectively.

Dynamic multi-relational graphs are an expressive relational representation for data enclosing entities and relations of different types, and where relationships are allowed to vary in time. Addressing predictive tasks over such data requires the ability to find structure embeddings that capture the diversity of the relationships involved, as well as their dynamic evolution. In this work, we establish a novel class of challenging tasks for dynamic multi-relational graphs involving out-of-domain link prediction, where the relationship being predicted is not available in the input graph. We then introduce a novel Graph Neural Network model, named GOOD, designed specifically to tackle the out-of-domain generalization problem. GOOD introduces a novel design concept for multi-relation embedding aggregation, based on the idea that good representations are such when it is possible to disentangle the mixing proportions of the different relational embeddings that have produced it. We also propose five benchmarks based on two retail domains, where we show that GOOD can effectively generalize predictions out of known relationship types and achieve state-of-the-art results. Most importantly, we provide insights into problems where out-of-domain prediction might be preferred to an in-domain formulation, that is, where the relationship to be predicted has very few positive examples.

Global retailers have assortments that contain hundreds of thousands of products that can be linked by several types of relationships like style compatibility, "bought together", "watched together", etc. Graphs are a natural representation for assortments, where products are nodes and relations are edges.Relations like style compatibility are often produced by a manual process and therefore do not cover uniformly the whole graph.We propose to use inductive learning to enhance a graph encoding style compatibility of a fashion assortment, leveraging rich node information comprising textual descriptions and visual data.Then, we show how the proposed graph enhancement improves substantially the performance on transductive tasks with a minor impact on graph sparsity.

In this paper, we propose a controller for smooth autonomous path following. The controller is formulated as an economic model predictive controller. The economic cost introduced in the objective function leads to a smooth driving, since we minimize the first and second derivatives of the curvature function (i.e., we encourage linear curvature profiles). Since the curvature in clothoids varies linearly with the path arc-length, we use the smoothness and comfort characteristics of clothoid-driving to obtain a compact and intuitive controller formulation. We enforce convergence of the controller to the reference path with soft constraints that avoid deviations from the reference path. Finally, we present real life experiments where the controller is deployed on a Scania construction truck that show that the proposed controller outperforms a pure-pursuit controller. Moreover, we detail how the few tuning parameters can affect the obtained solution in practice.

The problem of gas detection is relevant to many real-world applications, such as leak detection in industrial settings and surveillance. In this paper we address the problem of gas detection in large areas with a mobile robotic platform equipped with a remote gas sensor. We propose a novel method based on convex relaxation for quickly finding an exploration plan that guarantees a complete coverage of the environment. Our method proves to be highly efficient in terms of computational requirements and to provide nearly-optimal solutions. We validate our approach both in simulation and in real environments, thus demonstrating its applicability to real-world problems.

In this paper an extension to the sampling based motion planning framework CL-RRT is presented. The framework uses a system model and a stabilizing controller to sample the perceived environment and build a tree of possible trajectories that are evaluated for execution. Complex system models and constraints are easily handled by a forward simulation making the framework widely applicable. To increase operational safety we propose a sampling recovery scheme that performs a deterministic brake profile regeneration using collision information from the forward simulation. This greatly increases the number of safe trajectories and also reduces the number of samples that produce infeasible results. We apply the framework to a Scania G480 mining truck and evaluate the algorithm in a simple yet challenging obstacle course and show that our approach greatly increases the number of feasible paths available for execution.

In this paper we present an inspection robot to produce gas distribution maps and localize gas sources in large outdoor environments. The robot is equipped with a 3D laser range finder and a remote gas sensor that returns integral concentration measurements. We apply principles of tomography to create a spatial gas distribution model from integral gas concentration measurements. The gas distribution algorithm is framed as a convex optimization problem and it models the mean distribution and the fluctuations of gases. This is important since gas dispersion is not an static phenomenon and furthermore, areas of high fluctuation can be correlated with the location of an emitting source. We use a compact surface representation created from the measurements of the 3D laser range finder with a state of the art mapping algorithm to get a very accurate localization and estimation of the path of the laser beams. In addition, a conic model for the beam of the remote gas sensor is introduced. We observe a substantial improvement in the gas source localization capabilities over previous state-of-the-art in our evaluation carried out in an open field environment.

Due to its environmental, economical and safety implications, methane leak detection is a crucial task to address in the biogas production industry. In this paper, we introduce Gasbot, a robotic platform that aims to automatize methane emission monitoring in landfills and biogas production sites. The distinctive characteristic of the Gasbot platform is the use of a Tunable Laser Absorption Spectroscopy (TDLAS) sensor. This sensor provides integral concentration measurements over the path of the laser beam. Existing gas distribution mapping algorithms can only handle local measurements obtained from traditional in-situ chemical sensors. In this paper we also describe an algorithm to generate 3D methane concentration maps from integral concentration and depth measurements. The Gasbot platform has been tested in two different scenarios: an underground corridor, where a pipeline leak was simulated and in a decommissioned landfill site, where an artificial methane emission source was introduced.

In this paper we analyze the problem of discrimination of gases with mobile robots. Previously, it has been shown that the conditions in which data is collected heavily influence the characteristics of the signal to be identified. As a result, the already difficult task of selecting features which characterize a gas is made more challenging by the absence of a steady state response. This is often due to the movement of the robot, and/or the physical properties of the environment, e.g., turbulent airflow creating patches and eddies in the plume. In this work we compare two approaches for feature selection which are able to consider explicitly the information on the experimental setup and optimize the subset of features used in the recognition process. The approaches are tested on a large data set collected with a mobile robot moving in different environments (outdoors and indoors). The results show that the classification performance is improved resulting in a higher average accuracy and lower variance in the accuracy across the different experimental setups.

Gas distribution modelling constitutes an ideal application area for mobile robots, which - as intelligent mobile gas sensors - offer several advantages compared to stationary sensor networks. In this paper we propose the Kernel DM+V algorithm to learn a statistical 2-d gas distribution model from a sequence of localized gas sensor measurements. The algorithm does not make strong assumptions about the sensing locations and can thus be applied on a mobile robot that is not primarily used for gas distribution monitoring, and also in the case of stationary measurements. Kernel DM+V treats distribution modelling as a density estimation problem. In contrast to most previous approaches, it models the variance in addition to the distribution mean. Estimating the predictive variance entails a significant improvement for gas distribution modelling since it allows to evaluate the model quality in terms of the data likelihood. This offers a solution to the problem of ground truth evaluation, which has always been a critical issue for gas distribution modelling. Estimating the predictive variance also provides the means to learn meta parameters and to suggest new measurement locations based on the current model. We derive the Kernel DM+V algorithm and present a method for learning the hyper-parameters. Based on real world data collected with a mobile robot we demonstrate the consistency of the obtained maps and present a quantitative comparison, in terms of the data likelihood of unseen samples, with an alternative approach that estimates the predictive variance.

In this paper we investigate how a mobile robot equipped with tin dioxide gas sensors and an anemometer can use an online classification algorithm in order to improve the exploration strategy. The purpose of the platform is to establish the character of a gas source with accuracy while minimizing the time required for exploration. For this to be possible, the output of the classification algorithm is probabilistic, feeding in a sequence of posterior probabilities to a path planner. To further assist path planning, a 3d-ultrasonic anemometer is available which give indication on the average wind speed and direction. In addition to evaluating different olfaction driven path planning strategies, experimental validations also evaluate the classification algorithms and its application to different environments with varying characteristics.

Classification of odours with an array of gas sensors mounted on a mobile robot is a challenging and still relatively unexplored topic. Mobile robots able to classify an odour could navigate to a specific source or isolate high concentration areas in applications such as environmental monitoring. A key aspect to classification is to be able to process the data collected while moving the robot and using a simple and compact sensor system. In order to achieve this, we present a classification algorithm that is based on the transient response from the sensors. An analysis of how classification results vary with regards to the movement of the robot is provided and subsequently the experimental validations show that the classification performance depends more on how the robot traverses the odour plume and the quality of the transient than on the distance from the source location. The experimental validation has been done in a large unmodified indoor environment.

In this paper we present initial experiments towards environmental monitoring with a mobile platform. A prototype of a pollution monitoring robot was set up which measures the gas distribution using an ldquoelectronic noserdquo and provides three dimensional wind measurements using an ultrasonic anemometer. We describe the design of the robot and the experimental setup used to run trials under varying environmental conditions. We then present the results of the gas distribution mapping. The trials which were carried out in three uncontrolled environments with very different properties: an enclosed indoor area, a part of a long corridor with open ends and a high ceiling, and an outdoor scenario are presented and discussed.