A point to point hypothesis testing problem involving two parties, one referred to as the observer and the other as the detector, is studied. The observer observes a discrete memoryless source and communicates its observations to the detector over a discrete memoryless channel. The detector performs a binary hypothesis test on the probability distribution of the observer's observation. The trade-off between the type 1 error probability and the type 2 error exponent is explored. We obtain a single-letter characterization of the optimal type 2 error exponent for a given constraint on the type 1 error probability. We also show that a strong converse holds, in the sense that, the optimal type 2 error exponent is independent of the constraint on the type 1 error probability.
Hypothesis Testing over a Noisy Channel / Sreekumar, S.; Gunduz, D.. - 2019-:(2019), pp. 2004-2008. (Intervento presentato al convegno 2019 IEEE International Symposium on Information Theory, ISIT 2019 tenutosi a La Maison de La Mutualite, fra nel 2019) [10.1109/ISIT.2019.8849432].
Hypothesis Testing over a Noisy Channel
D. Gunduz
2019
Abstract
A point to point hypothesis testing problem involving two parties, one referred to as the observer and the other as the detector, is studied. The observer observes a discrete memoryless source and communicates its observations to the detector over a discrete memoryless channel. The detector performs a binary hypothesis test on the probability distribution of the observer's observation. The trade-off between the type 1 error probability and the type 2 error exponent is explored. We obtain a single-letter characterization of the optimal type 2 error exponent for a given constraint on the type 1 error probability. We also show that a strong converse holds, in the sense that, the optimal type 2 error exponent is independent of the constraint on the type 1 error probability.Pubblicazioni consigliate
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris