International audience

Continual learning in dynamic environments is a challenge for large-scale machine learning models. This research addresses Domain Incremental Learning (DIL), a setting where the goal is to incrementally increase the input data scope of a model. More specifically, we investigate the possibility of using foundation models (FMs) as a fixed feature extractor combined with a PPCA that can be sequentially and accurately updated. Focusing on the classification of VHR remote sensing (RS) images, we show on the FLAIR#1 dataset that this simple DIL strategy achieves competitive accuracy compared to memory-based baselines across different pre-trained sources. We also compare different types of foundation models and highlight the importance of data diversity over data specialization to improve the quality of FMs.

Agricultural abandonment is a global trend leading to vegetation succession and Forests expansion. Manual annotations of 1946 and 2019 aerial surveys images from a peri-urban area in Massif Central shows Land Use and Land Cover (LULC) evolution in this period. We propose to use a convolutional neural network trained on labelled years images to predict LULC maps from 13 intermediate years unlabelled images. However, sensors variety used for acquisition during this time induce variability in ground sampling distances and colorimetry. We have shown using transfer between labelled years that sampling distances have to be the same in the training and testing set, and that coarse scaling offer sufficient performances for the considered LULC classes. Colorimetric data augmentations were individually used after sampling unification to make models more robust to sensors and illumination changes, but proved to be inconsistent in transfer on intermediate years.

In open-ended environments, autonomous learning agents must set their own goals and build their own curriculum through an intrinsically motivated exploration. They may consider a large diversity of goals, aiming to discover what is controllable in their environments, and what is not. Because some goals might prove easy and some impossible, agents must actively select which goal to practice at any moment, to maximize their overall mastery on the set of learnable goals. This paper proposes CURIOUS , an algorithm that leverages 1) a modular Universal Value Function Approximator with hindsight learning to achieve a diversity of goals of different kinds within a unique policy and 2) an automated curriculum learning mechanism that biases the attention of the agent towards goals maximizing the absolute learning progress. Agents focus sequentially on goals of increasing complexity, and focus back on goals that are being forgotten. Experiments conducted in a new modular-goal robotic environment show the resulting developmental self-organization of a learning curriculum, and demonstrate properties of robustness to distracting goals, forgetting and changes in body properties.