Topic: meta-learning

Humans have the ability to quickly adapt to new tasks and generalise what they learned to improve the learning process itself. At the core of this capacity for meta-learning lies a difficult credit assignment problem. After a learning episode, the system needs to determine how to change the components of plasticity such that the next time a similar task is encountered, a better learning outcome can be arrived at more quickly. In machine learning, this problem is often approached by storing the entire learning trajectory and revisiting it in reverse-time order, a process that places large deman

Synaptic plasticity is known to be a key player in the brain’s life-long learning abilities. However, due to experimental limitations, the nature of the local changes at individual synapses and their link with emerging network-level computations remain unclear. I will present a numerical, meta-learning approach to deduce plasticity rules from either neuronal activity data and/or prior knowledge about the network's computation. I will first show how to recover known rules, given a human-designed loss function in rate networks, or directly from data, using an adversarial approach. Then I will pr

Animals are presented with noisy and incomplete information from which we learn how to react in novel situations. For example, a given training set could have many explanations, each implying a different behavioural response; acting appropriately requires inferring the most plausible explanation. The way animals choose to wield Occam's razor, selecting the most parsimonious explanation, is called their inductive bias, and it is implicitly built into the operation of animals' neural circuits. This relationship between an observed sensory circuit and its inductive bias is a useful explanatory wi

While meta-learning algorithms are often viewed as algorithms that learn to learn, an alternative viewpoint frames meta-learning as inferring a hidden task variable from experience consisting of observations and rewards. From this perspective, learning to learn is learning to infer. This viewpoint can be useful in solving problems in meta-RL, which I’ll demonstrate through two examples: (1) enabling off-policy meta-learning, and (2) performing efficient meta-RL from image observations. I’ll also discuss how this perspective leads to an algorithm for few-shot image segmentation.