Communication with Contextual Uncertainty
Communication with Contextual Uncertainty
复制标题
与情境不确定性进行沟通
DOI:
10.1007/s00037-017-0161-3
复制
发表时间:
2018
影响因子:
1.4
通讯作者:
Sudan, Madhu
中科院分区:
文献类型:
--
作者:
Ghazi, Badih;Komargodski, Ilan;Kothari, Pravesh K.;Sudan, Madhu
We introduce a simple model illustrating the utility of context in compressing communication and the challenge posed by uncertainty of knowledge of context. We consider a variant ofdistributionalcommunication complexity where Alice gets some informationand Bob gets, where (X,Y) is drawn from a known distribution, and Bob wishes to compute some functiong(X,Y) or some close approximation to it (i.e., the output isg(X,Y) with high probability over (X,Y)). In our variant, Alice does not knowg, but only knows some functionfwhich is a very close approximation tog. Thus, the function being computed forms the context for the communication. It is an enormous implicit input, potentially described by a truth table of size 2n. Imprecise knowledge of this function models the (mild) uncertainty in this context.We show that uncertainty can lead to a huge cost in communication. Specifically, we construct a distributionoverand a class of function pairs (f,g) which are very close (i.e., disagree witho(1) probability when (X,Y) are sampled according to), for which the communication complexity offorgin the standard setting isone bit, whereas the (two-way) communication complexity in the uncertain setting is at leastbits even when allowing a constant probability of error.It turns out that this blow-up in communication complexity can be attributed in part to the mutual information betweenXandY. In particular, we give an efficient protocol for communication under contextual uncertainty that incurs only a small blow-up in communication if this mutual information is small. Namely, we show that ifghas a communication protocol with complexitykin the standard setting and the mutual information betweenXandYisI, thenghas a one-way communication protocol with complexityin the uncertain setting. This result is an immediate corollary of an even stronger result which shows that ifghas one-way communication complexityk, then it has one-way uncertain-communication complexity at most. In the particular case where the input distribution is a product distribution (and soI= 0), the protocol in the uncertain setting only incurs aconstant factorblow-up in one-way communication and error.
登录
查看更多内容
DOI:
10.1145/2160158.2160161
发表时间:
2012-04
期刊:
Electron. Colloquium Comput. Complex.
影响因子:
--
作者:
Oded Goldreich;Brendan Juba;M. Sudan
通讯作者:
Oded Goldreich;Brendan Juba;M. Sudan
影响因子:
0.5
作者:
Huang, Wei;Shi, Yaoyun;Zhu, Yufan
通讯作者:
Zhu, Yufan
DOI:
10.1007/978-3-662-43948-7_13
发表时间:
2014
期刊:
Electron. Colloquium Comput. Complex.
影响因子:
--
作者:
Mohammad Bavarian;Dmitry Gavinsky;Tsuyoshi Ito
通讯作者:
Tsuyoshi Ito
影响因子:
1.1
作者:
Elad Haramaty;M. Sudan
通讯作者:
M. Sudan
影响因子:
2.5
作者:
Andrej Bogdanov;Elchanan Mossel
通讯作者:
Elchanan Mossel