Automatic Recall Machines: Internal Replay, Continual Learning and the Brain

By Xu Ji, João F. Henriques, Tinne Tuytelaars, Andrea Vedaldi. 2020

11 citations recorded by the source corpus.

Replay in neural networks involves training on sequential data with memorized samples, which counteracts forgetting of previous behavior caused by non-stationarity. We present a method where these auxiliary samples are generated on the fly, given only the model that is being trained for the assessed objective, without extraneous buffers or generator networks. Instead the implicit memory of learned samples within the assessed model itself is exploited. Furthermore, whereas existing work focuses on reinforcing the full seen data distribution, we show that optimizing for not forgetting calls for the generation of samples that are specialized to each real training batch, which is more efficient and scalable. We consider high-level parallels with the brain, notably the use of a single model for inference and recall, the dependency of recalled samples on the current environment batch, top-down modulation of activations and learning, abstract recall, and the dependency between the degree to which a task is learned and the degree to which it is recalled. These characteristics emerge naturally from the method without being controlled for.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Always Be Dreaming: A New Approach for Data-Free Class-Incremental Learning (2021, background): “Automatic-Recall Machines (ARM) [24] perturb training data from the current task into images which maximize “forgetting” from the past tasks.”
New Insights on Reducing Abrupt Representation Change in Online Continual Learning (2021, methodology): “Our evaluation includes the metrics and experimental settings used in previous works on online continual learning with a single-head (Aljundi et al., 2019a; Ji et al., 2020; Shim et al., 2020).”
GDumb: A Simple Approach that Questions Our Progress in Continual Learning (2020, methodology): “We present results in Table 4 and show that GDumb outperforms parallel works like HAL [35], QMED [12], ARM [41] in addition to outperforming recent works GSS [37], MIR [11], and ADI [47], consistently across datasets.”
Reducing Representation Drift in Online Continual Learning (2021, methodology): “We focus our evaluation on replay-based methods, as they have been shown to outperform other approaches in the online continual learning setting (Chaudhry et al., 2019; Caccia et al., 2019; Ji et al., 2020).”