[ABE-L] Insper Data Science Seminar - Hybrid

[Hedibert Lopes]


> 
> Data Science
>  
> Academic Seminar – Hybrid
>  
> Title: Split conformal prediction and its extensions to non-exchangeable data.
>  
> Speaker: Paulo Orenstein
> University: IMPA
>  
> Abstract:  Machine learning algorithms offer state-of-the-art predictive performance in a variety of domains, but often lack an associated measure of uncertainty quantification. Split conformal prediction is a leading tool to obtain predictive intervals with virtually no assumptions beyond data exchangeability. This crucial assumption, however, hinders its applicability to many important data, such as time series and spatially dependent processes. In this talk, we will introduce split CP and show how it can be extended to non-exchangeable settings through a small coverage penalty. The proposed framework, based on concentration of measure inequalities, works more generally than traditional split CP, and experiments corroborate our coverage guarantees even under highly dependent data.
> 
> May 30, 2022 - Excepcionalmente segunda-feira
> 
> 12:00 p.m. de São Paulo, Brasil (UTC/GMT -03:00)
> 
>  
> Paulo Renato de Souza room, 2nd floor - Building 1
>  
> Click here to join
>  

Sent from Hedibert's iPhone
-------------- Pr�xima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20220529/55e35f00/attachment-0001.htm>
-------------- Pr�xima Parte ----------
Um anexo não-texto foi limpo...
Nome: image004.jpg
Tipo: image/jpeg
Tamanho: 2383 bytes
Descrição: não disponível
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20220529/55e35f00/attachment-0003.jpg>
-------------- Pr�xima Parte ----------
Um anexo não-texto foi limpo...
Nome: image005.jpg
Tipo: image/jpeg
Tamanho: 2402 bytes
Descrição: não disponível
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20220529/55e35f00/attachment-0004.jpg>
-------------- Pr�xima Parte ----------
Um anexo não-texto foi limpo...
Nome: image006.jpg
Tipo: image/jpeg
Tamanho: 2388 bytes
Descrição: não disponível
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20220529/55e35f00/attachment-0005.jpg>