Nuova ricerca

MATTEO BOSCHINI

Dottorando presso: Dipartimento di Ingegneria "Enzo Ferrari"


Home |


Pubblicazioni

2020 - Dark Experience for General Continual Learning: a Strong, Simple Baseline [Relazione in Atti di Convegno]
Buzzega, Pietro; Boschini, Matteo; Porrello, Angelo; Abati, Davide; Calderara, Simone
abstract

Continual Learning has inspired a plethora of approaches and evaluation settings; however, the majority of them overlooks the properties of a practical scenario, where the data stream cannot be shaped as a sequence of tasks and offline training is not viable. We work towards General Continual Learning (GCL), where task boundaries blur and the domain and class distributions shift either gradually or suddenly. We address it through mixing rehearsal with knowledge distillation and regularization; our simple baseline, Dark Experience Replay, matches the network's logits sampled throughout the optimization trajectory, thus promoting consistency with its past. By conducting an extensive analysis on both standard benchmarks and a novel GCL evaluation setting (MNIST-360), we show that such a seemingly simple baseline outperforms consolidated approaches and leverages limited resources. We further explore the generalization capabilities of our objective, showing its regularization being beneficial beyond mere performance.


2020 - Rethinking Experience Replay: a Bag of Tricks for Continual Learning [Relazione in Atti di Convegno]
Buzzega, Pietro; Boschini, Matteo; Porrello, Angelo; Calderara, Simone
abstract

In Continual Learning, a Neural Network is trained on a stream of data whose distribution shifts over time. Under these assumptions, it is especially challenging to improve on classes appearing later in the stream while remaining accurate on previous ones. This is due to the infamous problem of catastrophic forgetting, which causes a quick performance degradation when the classifier focuses on learning new categories. Recent literature proposed various approaches to tackle this issue, often resorting to very sophisticated techniques. In this work, we show that naïve rehearsal can be patched to achieve similar performance. We point out some shortcomings that restrain Experience Replay (ER) and propose five tricks to mitigate them. Experiments show that ER, thus enhanced, displays an accuracy gain of 51.2 and 26.9 percentage points on the CIFAR-10 and CIFAR-100 datasets respectively (memory buffer size 1000). As a result, it surpasses current state-of-the-art rehearsal-based methods.