Annex 7: Habilitation thesis reviewer's report
Masaryk University
Faculty Faculty of Informatics, MU
Habilitation field Informatics
Applicant RNDr. Jan Bouda, Ph.D.
Unit Faculty of Informatics Masaryk University, Brno
Habilitation thesis Randomness in (Quantum) Information Processing
Reviewer Prof. Dr. Renato Renner
Unit Institute for Theoretical Physic, ETH Zurich, Switzerland
Reviewer's report (extent of text up to the reviewer)
The topic of the habilitation thesis of Dr. Jan Bouda is concerned with the use of randomness
in quantum information processing. In particular, it addresses the question of how (and under
which conditions) low-quality randomness can be turned into high-quality randomness.
Furthermore, he investigates the possibility of using low-quality randomness directly in
cryptographic applications.
Bouda has made several solid contributions to this area of research. Among them is a careful
analysis of the number of unitaries needed for a two-design, a concept that plays an important
role in quantum information theory. Another is the investigation of the security of private
quantum channels, where Bouda proposed and analysed an attack that may be regarded as a
quantum analogue of the known plaintext attack.
One particularly important result is the finding that the use of weak randomness in a quantum
key distribution protocol can lead to security loopholes. Because, in realistic implementations,
one cannot usually assume that the randomness used by the legitimate parties is perfect
(although perfect randomness is assumed in most known security proofs), this is of high
practical relevance.
Conversely, however, Bouda has also been able to show that, while classical encryption
schemes cannot usually tolerate weak randomness, this problem can be circumvented by
employing a quantum ciphertext. In other words, even if the key used for encryption is
sampled from a weak random source, secrecy can be guaranteed in most cases.
Another important line of results achieved by Bouda concerns the generation of randomness
using physical devices (such as the camera of a mobile device). In his work, he showed that
such randomness generation is possible with state-of-the-art equipment and appropriate
software for post-processing the raw randomness. Here, randomness extractors play a crucial
role, and Bouda has also made significant contributions to the development of such extractors.
In his very recent work, Bouda extended these considerations to a device-independent setting.
This means that the quality of the generated randomness can be guaranteed even if the devices
that produce them are only partially trusted. This, again, is of high practical importance, as
realistic devices do not usually meet the theoretical specifications perfectly.
The contributions of Bouda, as described above, have appeared in reasonable physics and
computer science journals, such as Physical Review A and the Journal on Quantum
Information and Computation, as well as in conference proceedings. Taken together, they
show that Bouda has a well-defined and well motivated research program. He is a respected
member of the quantum information research community and he entertains various
international collaborations.
Summarising, Jan Bouda has made a large number of relevant contributions in the area of
quantum information theory. He has clearly demonstrated his expertise in this field and shown
his ability to lead a well-motivated and productive research program. Given these
achievements, I am happy to recommend acceptance of his habilitation thesis.
Reviewer's questions for the habilitation thesis defence (number of questions up to the
reviewer)
1. There are obviously many ways to characterise weak randomness. How universal are
the various results on weak randomness? In other words, to what extent do they
depend on the precise definition of what one calls “weak”?
2. In practice, it is usually possible to extract almost perfect random bits from weak
randomness. Why is it still relevant to study the workings of cryptographic schemes
when used directly with weak randomness?
3. From a conceptual point of view, what is the difference in using “classical” physical
devices (such as random number generators based on thermal noise) from quantum
random number generators (e.g., those based on a beam splitter)? Is there a clear
advantage when using the latter?
4. What are the most important open research problems concerning the use of weak
randomness in quantum information processing?
Conclusion
The habilitation thesis submitted by Jan Bouda entitled “Randomness in (Quantum)
Information Processing” meets the requirements applicable to habilitation theses in the field
of Informatics.
In Zurich on October 30, 2015 Renato Renner (signature)