Brown University
George Mason University
Hasso Plattner Institute
Brave
ETH Zurich
PQShield Ltd
King's College London
University of Washington
NTT Social Informatics Laboratories
Universität der Bundeswehr München
TBD
TBA
Anonymous credentials are a strongly secure and privacy-preserving solution to user authentication but require users to securely manage cryptographic keys and credentials. In terms of usability, Single Sign-On (SSO) solutions, where the authentication is outsourced to a central identity provider, are significantly more convenient for end users. However, privacy is a major problem in SSO, as the identity provider can track all the user's online authentication attempts. In this talk, we will discuss how techniques from anonymous credentials and OPRFs can improve the privacy of Single Sign-On protocols such as OIDC, while maintaining its usability advantages - as users still don't have to manage any keys or credentials.
In this talk, we will take a look at different advancements in Private Information Retrieval (PIR), with specific emphasis on single-server constructions. We will talk about the needs from real-world deployment perspective, new schemes and what future research in this area will look like.
In this talk, I discuss a proof technique for proving the security of a subclass of blind signatures that have an "alternative secret key" that can be used in a security reduction. Members of this class are the Abe-Okamoto Partially Blind Signature Scheme, Abe's Blind Signature, Anonymous Credentials Light, as well as the recently introduced schemes by Tessaro and Zhu. I will also discuss limitations of the proof technique, some alternatives, and open questions in the area. The talk is based on joint works with Julian Loss, Omar Renawi, and Jiayu Xu.
The 3-round blind signature based on Schnorr signatures is one of the simplest and most efficient blind signatures. Unfortunately, the ROS attack by Benhamouda et al. (Eurocrypt'21) now practically breaks them. This has lead to new ideas to secure blind Schnorr while retaining its simplistic and efficient design. In the post-quantum setting, the landscape is less clear. While the high level construction idea of blind Schnorr has been successfully ported to the post-quantum setting, a new type of ROS attack now breaks many of them. Nevertheless, there still exist schemes where the attack does not apply. Moreover, unlike the relatively simple fixes for blind Schnorr, securing the now-(practically)-insecure post-quantum blind signatures presents challenges, requiring novel approaches.
Due to the significant progress in the area of (non-interactive) zero-knowledge proofs for lattice-related statements, there has been a lot of interest in building lattice-based two-round blind signatures following Fischlin's framework (Crypto 2006). In this talk, we will give an overview of the current state-of-the-art constructions and discuss their trade-offs with respect to signatures sizes, communication complexity and underlying hardness assumptions.
We present the first concurrently-secure blind signatures making black-box use of a pairing-free group for which unforgeability, in the random oracle model, can be proved without relying on the algebraic group model (AGM), thus resolving a long-standing open question. Prior pairing-free blind signatures without AGM proofs have only been proved secure for bounded concurrency or relied on computationally expensive non-black-box use of NIZKs. Our most efficient constructions rely on the chosen-target CDH assumption and can be seen as blind versions of signatures by Goh and Jarecki (EUROCRYPT '03) and Chevallier-Mames (CRYPTO '05). We also give a less efficient scheme with security based on (plain) CDH. The underlying signing protocols consist of four (in order to achieve regular unforgeability) or five moves (for strong unforgeability). All schemes are proved statistically blind in the random oracle model. Joint work with Rutchathon Chairattana-Apirom and Chenzhi Zhu
Digital signature schemes with specific properties have recently seen various real-world applications with a strong emphasis on privacy-enhancing technologies. In particular, signatures with randomizable keys find applications in a wide range of domains such as anonymous credentials, anonymity networks and blockchains. However, the literature on the topic is vast and different terminology is used across contributions, making it difficult to compare related works and understand the range of applications covered by a given construction. In this talk, which is based on a recent systematization of knowledge work from FC ‘24, we will first present a unified view of signatures with randomizable keys, revisiting their security properties. From there, we will discuss related applications and existing challenges in the area.
Equivalence Class Signatures (EQS) are signature schemes for a message space that is partitioned into equivalence classes. They are malleable in that signatures can be publicly randomized and adapted to other message representatives in the same equivalence class of a signed message. Importantly, adapted message-signature pairs are indistinguishable from random message-signature pairs. Together with the Decisional Diffie-Hellman assumption this gives an unlinkability notion and makes EQS a very attractive building block for privacy-preserving primitives such as anonymous credentials. The malleability of EQS has also further been extended to the key space, i.e., to support key randomization. Such schemes are called mercuiral signatures (MS) and further extend potential applications. In this talk we will introduce the aforementioned concepts and present the most efficient constructions. Then, we briefly capture recent results on their instantiability as well as review some (recent) applications. Finally, we will discuss some open problems related to this class of signature schemes.
CISPA Helmholtz Center for Information Security
CISPA Helmholtz Center for Information Security
© Designed and Developed by UIdeck