Underspecification Presents Challenges for Credibility in Modern Machine Learning.
Alexander D'AmourKatherine A. HellerDan MoldovanBen AdlamBabak AlipanahiAlex BeutelChristina ChenJonathan DeatonJacob EisensteinMatthew D. HoffmanFarhad HormozdiariNeil HoulsbyShaobo HouGhassen JerfelAlan KarthikesalingamMario LucicYi-An MaCory Y. McLeanDiana MincuAkinori MitaniAndrea MontanariZachary NadoVivek NatarajanChristopher NielsonThomas F. OsborneRajiv RamanKim RamasamyRory SayresJessica SchrouffMartin SeneviratneShannon SequeiraHarini SureshVictor VeitchMax VladymyrovXuezhi WangKellie WebsterSteve YadlowskyTaedong YunXiaohua ZhaiD. SculleyPublished in: CoRR (2020)
Keyphrases
- machine learning
- artificial intelligence
- lessons learned
- real world
- databases
- key issues
- computer vision
- website
- machine learning approaches
- inductive logic programming
- computational intelligence
- text mining
- information extraction
- knowledge discovery
- knowledge representation
- information technology
- natural language
- database systems
- case study
- decision trees
- feature selection
- social networks
- data mining