• Corpus ID: 220633085

Probabilistic Active Meta-Learning

  title={Probabilistic Active Meta-Learning},
  author={Jean Kaddour and Steind{\'o}r S{\ae}mundsson and Marc Peter Deisenroth},
Data-efficient learning algorithms are essential in many practical applications where data collection is expensive, e.g., in robotics due to the wear and tear. To address this problem, meta-learning algorithms use prior experience about tasks to learn new, related tasks efficiently. Typically, a set of training tasks is assumed given or randomly chosen. However, this setting does not take into account the sequential nature that naturally arises when training a model from scratch in real-life… 
Bayesian Meta-Learning for Few-Shot Policy Adaptation Across Robotic Platforms
The proposed method can successfully adapt a trained policy to different robotic platforms with novel physical parameters and the superiority of the meta-learning algorithm compared to state-of-the-art methods for the introduced few-shot policy adaptation problem is demonstrated.
Near-Optimal Task Selection for Meta-Learning with Mutual Information and Online Variational Bayesian Unlearning
This paper exploits the submodularity property of the new criterion for devising the first active task selection algorithm for meta-learning with a near-optimal performance guarantee and proposes an online variant of the Stein variational gradient descent to perform fast belief updates of the meta-parameters.
BAMLD: Bayesian Active Meta-Learning by Disagreement
To decrease the number of labeling requests for meta-training tasks, this paper introduces an information-theoretic active task selection mechanism which quantifies the epistemic uncertainty via disagreements among the predictions obtained under different inductive biases.
Towards Reliable and Efficient AI for 6G: Bayesian Active Meta-Learning for Few Pilot Demodulation and Equalization
Bayesian active meta-learning is seen in experiments to significantly reduce the number of frames required to obtain efficient adaptation procedure for new frames, and the capacity to quantify uncertainty in the model parameter space is further leveraged by extending Bayesian meta- learning to an active setting.
Not All Tasks are Equal-Task Attended Meta-learning for Few-shot Learning
  • Computer Science
  • 2022
This work introduces a training curriculum called task attended meta-training to learn a meta-model from weighted tasks in a batch, and introduces a comparisons of the task-attended ML models with their non-task-attENDED counterparts on complex datasets like miniImagenet, FC100 and tieredImagenets.
Auto-λ: Disentangling Dynamic Task Relationships
This work proposes an automated weighting framework, named Auto- λ, that explores continuous, dynamic task relationships via task-specific weightings, and can optimise any choice of combination of tasks through the formulation of a meta-loss; where the validation loss automatically influences task weightings throughout training.
Multidimensional Belief Quantification for Label-Efficient Meta-Learning
This work proposes a novel uncertainty-aware task selection model for label efficient meta-learning that formulates a multidimensional belief measure, which can quantify the known uncertainty and lower bound the unknown uncertainty of any given task.
Auto-Lambda: Disentangling Dynamic Task Relationships
This work proposes a gradient-based meta learning framework, Auto- λ, which explores continuous, dynamic task relationships via task-specific weightings, and can optimise any choice of combination of tasks through the formula-tion of a meta-loss; where the validation loss automatically influences task weightings throughout training.
FLAR: A Unified Prototype Framework for Few-sample Lifelong Active Recognition
This paper proposes a unified framework towards Few-sample Lifelong Active Recognition (FLAR), which aims at performing active recognition on progressively arising novel categories that only have few training samples.
Improving Generalization in Meta-RL with Imaginary Tasks from Latent Dynamics Mixture
LDM is proposed that trains a reinforcement learning agent with imaginary tasks generated from mixtures of learned latent dynamics that significantly outperforms standard meta-RL methods in test returns on the gridworld navigation and MuJoCo tasks where the authors strictly separate the training task distribution and the test task distribution.


Meta Reinforcement Learning with Latent Variable Gaussian Processes
This paper frame meta learning as a hierarchical latent variable model and infer the relationship between tasks automatically from data and shows that this results in up to a 60% reduction in the average interaction time needed to solve tasks compared to strong baselines.
Unsupervised Curricula for Visual Meta-Reinforcement Learning
The algorithm allows for unsupervised meta- learning that both transfers to downstream tasks specified by hand-crafted reward functions and serves as pre-training for more efficient meta-learning of test task distributions.
Unsupervised Meta-Learning for Reinforcement Learning
The experimental results indicate that unsupervised meta-reinforcement learning effectively acquires accelerated reinforcement learning procedures without the need for manual task design and these procedures exceed the performance of learning from scratch.
Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
We propose an algorithm for meta-learning that is model-agnostic, in the sense that it is compatible with any model trained with gradient descent and applicable to a variety of different learning
Meta-Learning surrogate models for sequential decision making
This work introduces a unified probabilistic framework for solving sequential decision making problems ranging from Bayesian optimisation to contextual bandits and reinforcement learning, and explores the use of Neural processes due to statistical and computational desiderata.
Meta-Learning Probabilistic Inference for Prediction
VERSA is introduced, an instance of the framework employing a flexible and versatile amortization network that takes few-shot learning datasets as inputs, with arbitrary numbers of shots, and outputs a distribution over task-specific parameters in a single forward pass, amortizing the cost of inference and relieving the need for second derivatives during training.
Domain randomization for transferring deep neural networks from simulation to the real world
This paper explores domain randomization, a simple technique for training models on simulated images that transfer to real images by randomizing rendering in the simulator, and achieves the first successful transfer of a deep neural network trained only on simulated RGB images to the real world for the purpose of robotic control.
Automatic Curriculum Learning For Deep RL: A Short Survey
The ambition of this work is to present a compact and accessible introduction to the Automatic Curriculum Learning literature and to draw a bigger picture of the current state of the art in ACL to encourage the cross-breeding of existing concepts and the emergence of new ideas.
Solving Rubik's Cube with a Robot Hand
It is demonstrated that models trained only in simulation can be used to solve a manipulation problem of unprecedented complexity on a real robot, made possible by a novel algorithm, which is called automatic domain randomization (ADR), and a robot platform built for machine learning.
Challenges of Real-World Reinforcement Learning
A set of nine unique challenges that must be addressed to productionize RL to real world problems are presented and an example domain that has been modified to present these challenges as a testbed for practical RL research is presented.