On Sequences with Non-Learnable Subsequences
| dc.creator | V'yugin, Vladimir V. | |
| dc.date | 2008-06-26 | |
| dc.date.accessioned | 2026-07-07T09:46:56Z | |
| dc.date.available | 2026-07-07T09:46:56Z | |
| dc.description | The remarkable results of Foster and Vohra was a starting point for a series of papers which show that any sequence of outcomes can be learned (with no prior knowledge) using some universal randomized forecasting algorithm and forecast-dependent checking rules. We show that for the class of all computationally efficient outcome-forecast-based checking rules, this property is violated. Moreover, we present a probabilistic algorithm generating with probability close to one a sequence with a subsequence which simultaneously miscalibrates all partially weakly computable randomized forecasting algorithms. %subsequences non-learnable by each randomized algorithm. According to the Dawid's prequential framework we consider partial recursive randomized algorithms. | |
| dc.identifier | https://arxiv.org/abs/0806.4341 | |
| dc.identifier | http://arxiv.org/abs/0806.4341 | |
| dc.identifier | LNCS 5010, pp. 302-313, 2008 | |
| dc.identifier.uri | http://salesiana.dossiersoluciones.com/handle/123456789/163700 | |
| dc.subject | Artificial Intelligence | |
| dc.subject | Machine Learning | |
| dc.subject | F.4.1; I.2.6 | |
| dc.title | On Sequences with Non-Learnable Subsequences | |
| dc.type | text |