Summer internships at the department in 2025
10. January 2025
The summer internship application period has started and you will find the job ad
here.
The deadline is on 26 January at 23:59 EET (UTC+2).
Public defence in Mathematics, M.Sc. (Tech) Pihla Karanko Sept. 19, 2024
9. September 2024
Doctoral student: Pihla Karanko
Opponent: Assistant Professor Pavel Hubáček, Czech Academy of Sciences, Czech Republic
Custos: Associate Professor Chris Brzuska, Aalto University School of Science, Department of Mathematics and Systems Analysis
Cryptography uses mathematical models to simulate real-world scenarios involving secrecy. Since we cannot know what an adversary might do (e.g. use supercomputers to break encryption), researchers try to model worst-case scenarios. Security is defined through "games" where a powerful unspecified adversarial algorithm attempts to break the system under controlled conditions. For example, in studying pseudorandom functions (PRFs), the adversary tries to distinguish between true PRF outputs and random bitstrings. If the adversary cannot guess correctly more than 50 % of the time, the PRF in question is considered secure.
Currently, no algorithm is proven to satisfy the rigorous security definitions for PRFs or other cryptographic tools. Instead, real-life systems rely on plausible candidates that have withstood extensive scrutiny. This reliance on unproven assumptions motivates efforts to reduce and better understand them. E.g. a PRF can be built from a one-way function (OWF), a tool with a simpler security definition.
Main Results:
- The thesis studies ways to get a OWF from a weaker (i.e. easier to break) OWF. We show that the existing methods are likely optimal in efficiency, highlighting the importance of the input distribution in such security amplification techniques. - We transform a weak PRF into a strong one using a special technique. This approach has practical applications in secure password handling, allowing more efficient authentication, where the user does not need to reveal the password to the server it is authenticating to.
- We propose a modification to a popular public key encryption mechanism, Fujisaki-Okamoto (FO) Transform, used in post-quantum secure encryption schemes. Our modification provides a more robust security proof.
- We propose a new security definition for garbling which is a method that allows outsourcing computations to untrusted servers without revealing the function or data. Our definition ensures strong security and efficient input encoding, overcoming current limitations when garbling cryptographic functions, useful for maintaining security when multiple servers encrypt a message jointly and one server becomes corrupted.
Key words: theoretical cryptography, one-way function, pseudorandom function
Thesis available for public display 10 days prior to the defence at: https://aaltodoc.aalto.fi/doc_public/eonly/riiputus/
Contact information:
Email pihla.karanko@aalto.fi
Doctoral theses at the School of Science: https://aaltodoc.aalto.fi/handle/123456789/52
Armaan Hooda spent one day as a math professor in Aalto
30. August 2024
Here you can read the whole article with Armaan's interview on Aalto's website.
Public defence in Systems and Operations Research, M.Sc. (Tech) Olli Herrala Sept. 6, 2024
22. August 2024
Mathematical optimization models for supporting climate decision making.
Doctoral student: Olli Herrala
Opponent: Associate Professor Giovanni Pantuso, University of Copenhagen, Denmark
Custos: Professor Fabricio Oliveira, Aalto University School of Science, Department of Mathematics and Systems Analysis
During the past 10 years, we have seen stronger and more frequent impacts of climate change. It is increasingly evident that action should be taken to at least slow down this change. However, making decisions about what exactly should be done is challenging, largely because of the substantial uncertainty in many parts of the problem.
Decision-making under uncertainty has been researched widely in the past 70 years, but most of the research assumes the uncertainty to be independent of our decisions. This is however not the case when making decisions in researching new climate change mitigation techniques such as carbon capture and storage, as these decisions have an impact on the highly uncertain future costs of climate change mitigation. To address this research gap, this thesis presents results on solving problems with decision-dependency in both probabilities and information structures.
This thesis also considers a hierarchical decision-making setting where an international policymaker wants to reduce emissions from electricity production by setting a carbon tax, while avoiding significant decreases in the total production that would increase electricity prices. However, the problem also includes transmission system operators and electricity producers, each with their own goals. The interactions between these players result in a complex problem where a carbon tax might have unexpected consequences such as shifting production from one country to another or simply increasing the price of electricity.
The methods presented in this dissertation allow decision-makers to model and anticipate the effects of decision-dependent uncertainty and hierarchical decision-making processes. The solution methods are based on mixed-integer optimization, leveraging the substantial developments in solving such models during the past 20 years. The case studies presented in the dissertation illustrate the capabilities of the proposed methods and show how they could be used to support decision-making in these complex systems.
Key words: mixed-integer optimization, climate change mitigation
hesis available for public display 10 days prior to the defence at: https://aaltodoc.aalto.fi/doc_public/eonly/riiputus/
Page content by: webmaster-math [at] list [dot] aalto [dot] fi
