We present SumSig, a code-based digital signature scheme that leverages sum-check protocols to reduce the reliance on repetition in Fiat–Shamir-based constructions. Instead of repeating a constant-soundness $\Sigma$-protocol many times, our approach verifies algebraic consistency of the entire witness via a single sum-check over an extension field, achieving negligible soundness error without repetition. Our construction introduces three main ideas: (1) a representation of the syndrome decoding witness as a multilinear polynomial suitable for sum-check verification; (2) a degree-doubling binarity enforcement technique based on power-sum constraints $S_1 = S_2 = S_4 = w$ to ensure binary witnesses; and (3) a linearization helper polynomial that enables efficient simulation in the random oracle model. For 128-bit security, SumSig yields signatures of approximately 5–8 KB with public keys of 50–100 KB, depending on the polynomial commitment scheme. This offers a different trade-off compared to existing code-based signatures such as Wave and LESS, which achieve either very small signatures with large public keys or moderate public keys with larger signatures. The resulting scheme features deterministic signing with no aborts and admits a quasi-tight reduction to the Syndrome Decoding problem in the random oracle model.

Many important applications of fully homomorphic encryption (FHE) require *high-precision* arithmetic, e.g., plaintext rings $\mathbb{Z}_p$ for a huge prime or power-of-two modulus $p$. The classic FHE schemes are poorly suited to this, because the inverse error rate of fresh ciphertexts, and the error growth under homomorphic multiplication, are both larger than $p$, which results in large and inefficient parameters. While there are now several works addressing this problem, the landscape for *exact* (as opposed to approximate) FHE is highly fragmented: known solutions either work only for certain rare plaintext moduli having very special forms (sometimes using non-standard ciphertext rings that lack other important features for FHE), or have quite complicated and high-latency constructions. This work gives a very simple and general technique for high-precision exact FHE, in which the error rates and growth match those of prior schemes with *exponentially smaller* precision, and which has good practical efficiency. In contrast to all prior works, our scheme works for *any integer modulus*, and also over *any underlying (number) ring*---or even with no structured ring at all, making it the first solution that can be based on plain LWE. It is also *fully compatible with prior FHE techniques* for fast ring arithmetic, plaintext packing and SIMD operations, bootstrapping, etc. For plaintext ring $\mathbb{Z}_{2^{64}}$, our (preliminary, unoptimized, single-threaded) implementation does homomorphic multiplication in just tens of milliseconds, and supports a two- to three-fold increase in multiplicative depth versus classic FHE schemes at typical security levels.

Thieves executed a meticulously planned heist, making off with a estimated $35 million in valuables from a Sparkasse vault in Gelsenkirchen

The 25 forecasts we made in 2025, revisited.

A growing adoption of transformer-based machine learning models is raising concerns about sensitive data exposure. Nonetheless, current secure inference solutions incur substantial overhead due to their extensive reliance on non-linear protocols, such as softmax and Gaussian error linear unit (GELU). Driven by numerical stability needs, softmax approximations (e.g., NeurIPS 2021) typically extract the maximum element of an input vector, incurring logarithmic rounds (in the input length). Existing GELU protocols (e.g., S&P 2024) use piecewise approximations with high-degree polynomials that rely heavily on secure multiplications and comparisons, which are expensive. Such complexities also hinder model owners unfamiliar with cryptography from deploying custom models. SHAFT, our proposed system, provides a secure, handy, accurate, and fast transformer inference framework for deployment. Highlights of our contributions include 1) the first constant-round (independent of sequence length) softmax protocol for transformers, using input clipping and an ordinary differential equation characterization, and 2) a highly accurate GELU protocol on a novel characterization designed for Fourier series approximation. Extending to broader contexts, our new protocols also apply to general neural networks that use softmax as the final layer and to transformer architectures with different activation functions. Remarkably, SHAFT outperforms state-of-the-art SIGMA (PETS 2024), which uses secret sharing, and BumbleBee (NDSS 2025), which additionally uses RLWE-based homomorphic encryption. More specifically, SHAFT reduces communication by 62–70% and is 1.8–2.4× faster than SIGMA, while also surpassing BumbleBee in terms of running time by 2.6–3.7× under LAN settings. Alongside these improvements, SHAFT attains accuracy comparable to plaintext models, confirming its numerical stability. Next in this progression, SHAFT provides an accessible open-source framework for secure and handy deployment by smoothly integrating with the Hugging Face library (EMNLP Demos 2020).

A recent interstellar visitor crossed our inner Solar System.  We are told it is “probably natural.”  We are offered models…

These ten movies got the worst Metacritic scores of the year. Did they deserve the critical drubbings?

Canadian computer scientist Yoshua Bengio warns against granting legal rights to cutting-edge technology

While approximately 4.3% of Oklahoma farmland is foreign-owned, according to the USDA’s most recent filings, most of that is held by Canadian and European companies for renewable energy projects. Less than 1% of that share is Chinese.

In 2026, C programming conferences like CppCon, Programming Conference, and LambdaConf will highlight the language's vital role in embedded systems, AI, cybersecurity, and high-performance computing. These hybrid events foster innovation through workshops, networking, and discussions on emerging trends. Attendees can expect breakthroughs in low-level programming amid technological advancements.

Blog com notícias sobre, Linux, Android, Segurança , etc

2025 was a big year for cybersecurity, with cyberattacks, data breaches, threat groups reaching new notoriety levels, and, of course, zero-day flaws exploited in breaches. Some stories, though, were more impactful or popular with our readers than others. This article explores 15 of the biggest cybersecurity stories of 2025.

A deep-sea search for Malaysia Airlines Flight 370 has begun in the Indian Ocean, reviving efforts to solve one of aviation’s greatest mysteries.

Scientific Reports - Reversible data hiding in encrypted images using LWE-based secret sharing

Learn more about the “glue sniffer” protein, which is able to detect brain cells’ incoming chemical signals, and what that means for neuroscience.

A China-linked APT used DNS poisoning to deliver the MgBot backdoor in targeted cyber-espionage attacks in Türkiye, China, and India.

Looking for a Linux distribution to challenge you? AerynOS delivers that - plus atomic-powered security and stability.

The U.S. is playing it safe when it comes to chipmaking tools entering China.

Recently, studies exemplified by Hyper-Connections (HC) have extended the ubiquitous residual connection paradigm established over the past decade by expanding the residual stream width and...

The bloodletting will hit hardest in back-office operations, risk management, and compliance.

Bork!Bork!Bork!: Do you want bork with that?

Flock left at least 60 of its people-tracking Condor PTZ cameras live streaming and exposed to the open internet.

VVS stealer (or VVS $tealer) is a Python-based infostealer targeting Discord users. It employs Pyarmor for obfuscation, contributing to its efficacy.