Publication

Lifelong Machine Learning with Data Efficiency and Knowledge Retention

Related publications (208)

Graph Chatbot

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

Nine recommendations for the governance of AI systems

Marie-Valentine Renée Agnès Florin

Some governance functions traditionally performed by humans are increasingly informed and sometimes automatically executed by machine learning algorithms (governance by machine learning) to benefit society. Therefore, it is necessary to think also about th ...

EPFL2022

Humans are Poor Few-Shot Classifiers for Sentinel-2 Land Cover

Devis Tuia, Marc Conrad Russwurm

Learning to predict accurately from a few data samples is a central challenge in modern data-hungry machine learning. On natural images, human vision typically outperforms deep learning approaches on few-shot learning. However, we hypothesize that aerial a ...

IEEE2022

Instance norm improves meta-learning in class-imbalanced land cover classification

Devis Tuia, Marc Conrad Russwurm

Distribution shift is omnipresent in geographic data, where various climatic and cultural factors lead to different representations across the globe. We aim to adapt dynamically to unseen data distributions with model-agnostic meta-learning, where data sa ...

2022

Dif-MAML: Decentralized Multi-Agent Meta-Learning

Ali H. Sayed, Mert Kayaalp, Stefan Vlaski

The objective of meta-learning is to exploit knowledge obtained from observed tasks to improve adaptation to unseen tasks. Meta-learners are able to generalize better when they are trained with a larger number of observed tasks and with a larger amount of ...

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC2022

Loss landscape and symmetries in Neural Networks

Mario Geiger

Neural networks (NNs) have been very successful in a variety of tasks ranging from machine translation to image classification. Despite their success, the reasons for their performance are still not well-understood. This thesis explores two main themes: lo ...

EPFL2021

Systematic selection of field response measurements for excavation back analysis

Ian Smith, Numa Joy Bertola

In a geotechnical excavation, back analyses are routinely performed using the measured field responses to derive the material parameter values for the different soil layers present at the site. For the purpose of back analyses, the engineers will usually m ...

2021

The role of convolutional neural networks in scanning probe microscopy: a review

Georg Fantner

Progress in computing capabilities has enhanced science in many ways. In recent years, various branches of machine learning have been the key facilitators in forging new paths, ranging from categorizing big data to instrumental control, from materials desi ...

BEILSTEIN-INSTITUT2021

Rethinking data browsing through embodied knowledge

In the age of digital and big data, the interaction with machines is central and emphasizes the sophistication of user interfaces and browsing tools. However, in the domain of digital artwork collections, most platforms propose access to images through key ...

2021

Measuring and modeling the motor system with machine learning

Alexander Mathis, Alessandro Marin Vargas

The utility of machine learning in understanding the motor system is promising a revolution in how to collect, measure, and analyze data. The field of movement science already elegantly incorporates theory and engineering principles to guide experimental w ...

2021

Word-level Embeddings for Cross-Task Transfer Learning in Speech Processing

Milos Cernak, Pierre Anton Beckmann

Recent breakthroughs in deep learning often rely on representation learning and knowledge transfer. In recent years, unsupervised and self-supervised techniques for learning speech representation were developed to foster automatic speech recognition. Up to ...

EUROPEAN ASSOC SIGNAL SPEECH & IMAGE PROCESSING-EURASIP2021