11 June 2019 in PLOS Computational Biology
Popular computational models of decision-making make specific assumptions about learning processes that may cause them to underfit observed behaviours. Here we suggest an alternative method using recurrent neural networks (RNNs) to generate a flexible family of models that have sufficient capacity to represent the complex learning and decision- making strategies used by humans. In this approach, an RNN is trained to predict the next action that a subject will take in a decision-making task and, in this way, learns to imitate the processes underlying subjects’ choices and their learning abilities. We demonstrate the benefits of this approach using a new dataset drawn from patients with either unipolar (n = 34) or bipolar (n = 33) depression and matched healthy controls (n = 34) making decisions on a two-armed bandit task. The results indicate that this new approach is better than baseline reinforcement-learning methods in terms of overall performance and its capacity to predict subjects’ choices. We show that the model can be interpreted using off-policy simulations and thereby provides a novel clustering of subjects’ learning processes—something that often eludes traditional approaches to modelling and behavioural analysis. Computational models of decision-making provide a quantitative characterisation of the learning and choice processes behind human actions. Designing a computational model is often based on manual engineering which an iterative process to examine the consistency between different aspects of the model and the empirical data. In practice, however, inconsistencies between the model and observed behaviours can remain hidden behind examined summary statistics. To address this limitation, we developed a recurrent neural network (RNNs) as a flexible type of model that can automatically characterize human decision-making processes without requiring tweaking and engineering. To show the benefits of this new approach, we collected data on a decision-making task conducted on subjects with either bipolar or unipolar depression, as well as healthy controls. The results showed that, indeed, important aspects of decision-making remained uncaptured by typical computational models and even their enhanced variants, but were captured by RNNs automatically. Further, we were able to show that the nature of such processes can be unveiled by simulating the model under various conditions. This new approach can be used, therefore, as a standalone model of decision-making or as a baseline model to evaluate how well other candidate models fit observed data.
25 July 2019 in IEEE Robotics and Automation Letters
28 July 2018 in Autonomous Robots
23 December 2021 in IEEE Robotics and Automation Letters
Quantification of uncertainty in point cloud matching is critical in many tasks such as pose estimation, sensor fusion, and grasping. Iterative closest point (ICP) is a commonly used pose estimation algorithm which provides a point estimate of the transformation between two point clouds. There are many sources of uncertainty in this process that may arise due to sensor noise, ambiguous environment, initial condition, and occlusion. However, for safety critical problems such as autonomous driving, a point estimate of the pose transformation is not sufficient as it does not provide information about the multiple solutions. Current probabilistic ICP methods usually do not capture all sources of uncertainty and may provide unreliable transformation estimates which can have a detrimental effect in state estimation or decision making tasks that use this information. In this work we propose a new algorithm to align two point clouds that can precisely estimate the uncertainty of ICP’s transformation parameters. We develop a Stein variational inference framework with gradient based optimization of ICP’s cost function. The method provides a non-parametric estimate of the transformation, can model complex multi-modal distributions, and can be effectively parallelized on a GPU. Experiments using 3D kinect data as well as sparse indoor/outdoor LiDAR data show that our method is capable of efficiently producing accurate pose uncertainty estimates.
24 January 2020 in IEEE Robotics and Automation Letters
Sampling-based planning is the predominant paradigm for motion planning in robotics. Most sampling-based planners use a global random sampling scheme to guarantee probabilistic completeness. However, most schemes are often inefficient as the samples drawn from the global proposal distribution, and do not exploit relevant local structures. Local sampling-based motion planners, on the other hand, take sequential decisions of random walks to samples valid trajectories in configuration space. However, current approaches do not adapt their strategies according to the success and failures of past samples. In this work, we introduce a local sampling-based motion planner with a Bayesian learning scheme for modelling an adaptive sampling proposal distribution. The proposal distribution is sequentially updated based on previous samples, consequently shaping it according to local obstacles and constraints in the configuration space. Thus, through learning from past observed outcomes, we maximise the likelihood of sampling in regions that have a higher probability to form trajectories within narrow passages. We provide the formulation of a sample-efficient distribution, along with theoretical foundation of sequentially updating this distribution. We demonstrate experimentally that by using a Bayesian proposal distribution, a solution is found faster, requiring fewer samples, and without any noticeable performance overhead.
14 March 2025 in IEEE Robotics and Automation Letters
We present FORGE, a method for sim-to-real transfer of force-aware manipulation policies in the presence of significant pose uncertainty. During simulation-based policy learning, FORGE combines a force threshold mechanism with a dynamics randomization scheme to enable robust transfer of the learned policies to the real robot. At deployment, FORGE policies, conditioned on a maximum allowable force, adaptively perform contact-rich tasks while avoiding aggressive and unsafe behaviour, regardless of the controller gains. Additionally, FORGE policies predict task success, enabling efficient termination and autonomous tuning of the force threshold. We show that FORGE can be used to learn a variety of robust contact-rich policies, including the forceful insertion of snap-fit connectors. We further demonstrate the multistage assembly of a planetary gear system, which requires success across three assembly tasks: nut threading, insertion, and gear meshing.