Learning an Adversary's Actions for Secret Communication

Mehrdad Tahmasbi, Matthieu R. Bloch, Aylin Yener

Research output: Contribution to journalArticlepeer-review

4 Scopus citations


Secure communication over a wiretap channel is investigated, in which an active adversary modifies the state of the channel and the legitimate transmitter has the opportunity to sense and learn the adversary's actions. The adversary has the ability to switch the channel state and observe the corresponding output at every channel use while the encoder has causal access to observations that depend on the adversary's actions. A joint learning/transmission scheme is developed in which the legitimate users learn and adapt to the adversary's actions. For some channel models, it is shown that the achievable rates, defined precisely for the problem, are arbitrarily close to those obtained with hindsight, had the transmitter known the actions ahead of time. This initial study suggests that there is much to exploit and gain in physical-layer security by learning the adversary, e.g., monitoring the environment.

Original languageEnglish (US)
Article number8836089
Pages (from-to)1607-1624
Number of pages18
JournalIEEE Transactions on Information Theory
Issue number3
StatePublished - Mar 2020

All Science Journal Classification (ASJC) codes

  • Information Systems
  • Computer Science Applications
  • Library and Information Sciences


Dive into the research topics of 'Learning an Adversary's Actions for Secret Communication'. Together they form a unique fingerprint.

Cite this