Embedded systems are increasingly leveraging Artificial Intelligence of Things (AIoT) to enable real-time decision-making in critical applications, such as autonomous navigation and medical diagnostics. In these contexts, Random Forests (RFs) have been widely adopted due to their inherent parallelism. However, RFs rely on axis-aligned splits, which limit their ability to model complex decision boundaries. Oblique Random Forests (ORFs), which employ hyperplane-based splits, offer a more expressive alternative by improving classification accuracy. Despite their advantages, inference of ORFs is resource-consuming, prohibiting the implementation of such models on resource-constrained hardware devices.In this work, we present Shakan, a novel framework for Oblique Decision Trees (ODTs) inference on embedded systems. We introduce a new training technique designed to mitigate both training complexity and overfitting while enabling low-latency inference in hardware, along with a new architecture that maximizes performance and optimizes resource usage. Shakan enables, on resource-constrained devices, the inference of several ORFs configurations that can provide either a significant increase in accuracy or a notable speedup in terms of inference latency compared to state-of-the-art accelerators for traditional RFs on embedded devices. The most accurate configurations provide average accuracy improvements above 5% with similar latency, while the fastest configurations achieve speedups of 1140×, 214×, and 29× for tree depths of 5, 7, and 9, respectively, with comparable accuracy.

Post-quantum public key encryption (PKE) schemes employing Quasi-Cyclic (QC) sparse parity-check matrix codes are enjoying significant success, thanks to their good performance profile and significant reduction in the keypair size. However, there is no formal proof that the hardness of decoding random QC codes is related to the decoding hardness of random codes, which is known to be NP-hard, nor that changing the (constant in the length) rate of the employed QC codes does not change the nature of the underlying hard problem. In this work, we address and solve these challenges, answering both of them in the affirmative. First, we prove computational equivalences among hard problems from coding theory and the corresponding problems for QC codes. Then, we provide a systematization of hard problems and security assumptions underlying QC-MDPC-based cryptosystems, proving that fixing the rate of the QC codes does not change the hardness of key recovery attacks. These results allow the design of Niederreiter-style QC-MDPC PKEs, with the additional flexibility granted by freely choosing the code rate, leading to code-based public key encryption schemes which are both secure and can be fit in very demanding scenarios, such as embedded systems.

Cryptography based on error correction codes has gained significant interest due to its ability to provide security against both classical and quantum adversaries. In 2025, the U.S. National Institute of Standards and Technology selected the Hamming Quasi-Cyclic (HQC) key encapsulation mechanism for standardization. A key aspect of HQC is the possibility of decryption failures, which reveal information about the private key. To address this issue, the HQC authors developed a probabilistic model for the decoding failure rate (DFR) of the underlying error-correcting code, and adjusted the cryptosystem parameters to thwart attacks based on decryption failures. However, the DFR model relies on the assumption of independence between coordinates of the error vector, which does not hold in HQC. This approximation yields conservative DFR estimates in regimes where failure probabilities can be simulated, and it is hypothesized to remains conservative for cryptographic-grade parameter sets. In this work, we eliminate the independence assumptions and derive a new closed-form DFR model for HQC. We demonstrate that the previous approximation remains conservative in the cryptographic regime and that HQC’s current decoding failure rates are lower than the required ones. We describe optimization techniques that enable our probabilistic model to serve as a parameter-tuning tool, and demonstrate how the size of HQC public keys and ciphertexts can be slightly reduced without compromising security.

In the Noisy Intermediate-Scale Quantum (NISQ) era, the topological constraints present in many of the currently available quantum devices pose a physical limit on the feasible interactions between qubits. To comply with such limitations, the compilation of quantum circuits requires solving the Qubit Routing Problem (QRP), by inserting SWAP operations among qubits. The State of the Art provides heuristic algorithms addressing this task, yet the depth of the output circuits is often incompatible with the current limits of quantum hardware. Therefore, we propose DDRoute, a novel heuristic algorithm to solve QRP, designed to reduce the depth overhead introduced by the routing process in the compiled circuits. Our experimental evaluation proves the efficiency of our approach, with a depth reduction of up to 70% with respect to the state-of-the-art routing procedures.

Providing closed form estimates of the Decoding Failure Rates (DFR) of iterative decoder for low- and moderate-density parity check codes has attracted significant interest in the research community over the years. This interest has raised due to the use of iterative decoders in post-quantum cryptosystems, where the desired DFRs are impossible to estimate via Monte Carlo simulations. In this work, we propose a new technique to provide accurate estimates of the DFR of a two-iterations (parallel) bit-flipping decoder, which is also employable for cryptographic purposes. In doing so, we successfully tackle the estimation of the bit-flipping probabilities at the first and second decoder iteration, and provide a fitting estimate for the syndrome weight distribution. We numerically validate our results, providing comparisons of the modeled and simulated weight of the syndrome, incorrectly-guessed error bit distribution at the end of the first iteration, and two-iteration DFR, both in the floor and waterfall regime. Finally, we apply our method to estimate the DFR of LEDAcrypt, a post-quantum cryptosystem, improving by factors larger than$2^{70}$, with respect to the previous estimation techniques.