Probabilistic Active Meta-Learning
Authors: Jean Kaddour, Steindor Saemundsson, Marc Deisenroth (he/him)
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In our experiments, we assess whether PAML speeds up learning task domains by learning a meta-model for the dynamics of simulated robotic systems. We test its performance on varying types of task-descriptors. Specifically, we generate tasks within domains by varying configuration parameters of the simulator, such as the masses and lengths of parts of the system. We then perform experiments where the learning algorithm observes: (i) fully observed task parameters, (ii) partially observed task parameters, (iii) noisy task parameters and (iv) high-dimensional image descriptors. We compare PAML to uniform sampling (UNI), used in recent meta-learning work [1, 15] and equivalent to domain randomization [16], Latin hypercube sampling (LHS) of the parameterization interval, and an oracle, i.e., the meta-learning model trained on the test tasks, representing an upper bound on the predictive performance given a fixed model. |
| Researcher Affiliation | Academia | Jean Kaddour Department of Computer Science University College London Steindór Sæmundsson Department of Computing Imperial College London Marc Peter Deisenroth Department of Computer Science University College London |
| Pseudocode | Yes | Algorithm 1 PAML |
| Open Source Code | No | The paper does not contain any explicit statement about providing open-source code for the described methodology, nor does it provide a link to a code repository. |
| Open Datasets | No | In our experiments, we assess whether PAML speeds up learning task domains by learning a meta-model for the dynamics of simulated robotic systems. We test its performance on varying types of task-descriptors. Specifically, we generate tasks within domains by varying configuration parameters of the simulator, such as the masses and lengths of parts of the system. The paper describes generating data through simulation rather than using an explicitly linked or cited public dataset. |
| Dataset Splits | No | The paper mentions '100 test tasks' and 'We start with four initial tasks and then sequentially add 15 more tasks.' and 'Fixed, evenly spaced grids of test task parameters'. However, it does not provide explicit training, validation, or test split percentages or absolute sample counts that would allow full reproduction of data partitioning. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments. It refers to 'simulated robotic systems' but no underlying hardware. |
| Software Dependencies | No | The paper mentions using a 'Gaussian process (GP) [18]' and 'sparse variational GP formulation from [19]', as well as a 'Variational Auto Encoder [21, 22]', but it does not specify any software dependencies with version numbers (e.g., PyTorch 1.9, TensorFlow 2.x, Python 3.x). |
| Experiment Setup | Yes | Observations in these tasks consist of state-space observations, x, x, i.e., position, velocity and control signals u. We start with four initial tasks and then sequentially add 15 more tasks. ... The hyper-parameters of the GP play the role of the global parameters θ and are shared across all tasks. A detailed description of (hyper-)parameters for the experiments is given in the Appendix. The cart-pole tasks differ by varying masses of the attached pendulum and the cart, pm [0.5, 5.0] kg and pl [0.5, 2.0] m, respectively. Pendubot and cart-double-pole tasks have lengths of both pendulums in the ranges, pl1, pl2 [0.6, 3.0] m and pl1, pl2 [0.5, 3.0] m, respectively. |