FUNCTION-SPACE PARAMETERIZATION OF NEURAL NETWORKS FOR SEQUENTIAL LEARNING

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussaPosterScientificvertaisarvioitu

1 Sitaatiot (Scopus)

Abstrakti

Sequential learning paradigms pose challenges for gradient-based deep learning due to difficulties incorporating new data and retaining prior knowledge. While Gaussian processes elegantly tackle these problems, they struggle with scalability and handling rich inputs, such as images. To address these issues, we introduce a technique that converts neural networks from weight space to function space, through a dual parameterization. Our parameterization offers: (i) a way to scale function-space methods to large data sets via sparsification, (ii) retention of prior knowledge when access to past data is limited, and (iii) a mechanism to incorporate new data without retraining. Our experiments demonstrate that we can retain knowledge in continual learning and incorporate new data efficiently. We further show its strengths in uncertainty quantification and guiding exploration in model-based RL. Further information and code is available on the project website.

AlkuperäiskieliEnglanti
Sivumäärä29
TilaJulkaistu - 2024
OKM-julkaisutyyppiEi sovellu
TapahtumaInternational Conference on Learning Representations - Messe Wien Exhibition and Congress Center, Vienna, Itävalta
Kesto: 7 toukok. 202411 toukok. 2024
Konferenssinumero: 12
https://iclr.cc/

Conference

ConferenceInternational Conference on Learning Representations
LyhennettäICLR
Maa/AlueItävalta
KaupunkiVienna
Ajanjakso07/05/202411/05/2024
www-osoite

Sormenjälki

Sukella tutkimusaiheisiin 'FUNCTION-SPACE PARAMETERIZATION OF NEURAL NETWORKS FOR SEQUENTIAL LEARNING'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Siteeraa tätä