How do ESP32s get boot entropy when RNG, radio bursts, and SRAM are all shaky, and does an 'entropy capsule' help cold starts?
trending30
01 02 Fraud control gets a blockchain audit trail here, with synthetic telecom/IoT requests, federated meta-learning, and QLoRA LLMs.03 Embeds SD-JWT signatures into read-only files, so authenticity stops depending on whatever IAM duct tape is around.04 A Seed for Privacy -- semi-automatic privacy-revealing data reminder in databases and data streams arxiv.orgArXiv paper on pArborist, semi-automatic query growth for privacy-revealing events in DBs and streams, because privacy loves paperwork.05 One source, two GPU vendors, and memory bandwidth still gets the last word in this OpenMP LWE KEM paper.06 One Token Is Enough: Fingerprinting and Verifying Large Language Models from Single-Token Output Distributions arxiv.orgInstead of probing internals, this fingerprints LLMs from trivial one-token distributions and still tracks lineage through API chains.07 ArXiv paper on image encryption via E91 keys, Lorenz chaos, and reservoir computing, with the usual NPCR/UACI ritual.08 ARMOR-IMC: Adaptive Resource Mapping for Operational Robustness via Secure In-Memory Computing arxiv.orgPost-training defense for analog in-memory accelerators, maps around process faults and power leakage without retraining.09 Aggregate PM counters verify 6G RAN intents, letting 3GPP/O-RAN pretend privacy and assurance coexist.10 When Routes Run Out: Adversarial Co-Learning and Explainable Robustness in Quantum Repeater Networks arxiv.orgAdversarial bandit routing for quantum repeaters, then a decision tree and LLM explain why the agent survives Eve.11 Can federated learning keep its accuracy once you bolt on HE and DP, or does client heterogeneity still wreck it?12 PromptGraph: Graph-Guided Prompt Sanitization for Balancing Privacy and Utility in LLM Inference arxiv.orgCan a graph over prompt spans scrub PII without torching LLM utility?13 Paper tests watermark-like logit perturbations against black-box LLM theft, then measures how much utility gets shaved off.14 Front-running meets a power-weighted lottery here, and the paper claims that makes sandwiching uneconomic.15 Devil in the Lens: Analyzing and Defending Physical Prompt Injection Against Vision-Language Models on Wearable Devices arxiv.orgPhysical prompt injection on AI glasses works across VLMs; the paper also benchmarks a mask filter and semantic detector.16 Byzantine Accountability Without Consensus: Strong Eventual Consistency for Non-Associative, Stochastic, Robust Aggregation arxiv.orgACFA: consensus-free Byzantine replication for robust aggregators like multi-Krum, by CRDTing signed contributions and equivocations.17 LLM-guided evolution breaks perceptual hashes like pHash, PDQ, PhotoDNA, and NeuralHash in black-box mode, because apparently those work...18 Can you trust Opal2 SEDs on Linux? This black-box study of 38 drives says firmware bugs and incompatibilities say no.19 Cardano's governance paper is a full spec plus research wish list for proving the thing doesn't deadlock.20 Survey of LLM watermarking, from biased sampling to adaptive-removal attacks, because even text wants DRM.21 Toward Stronger Code Watermarking: A Grammar-Driven Approach to Optimizing the Trade-off Between Quality and Detectability arxiv.orgGrammar-driven code watermarking, using syntax-aware masking and role biasing instead of the usual token twitching.22 What if digital credentials were designed around verifier steps, not issuer wishful thinking?23 The gateway moves into the client, which then proves it deserves access in a zkVM; RISC Zero demo for XACML and JWTs.24 QUBO Modeling of Module Learning With Errors: Stability and Scaling in Post-Quantum Cryptography arxiv.orgQUBO-encodes small MLWE instances for annealers, then checks noise stability and embedding blowup as dimensions grow.25 Degree-Constrained Interval Optimization for Minimax Polynomial Approximation in Homomorphic Encryption arxiv.orgHow do you pick the approximation interval for HE minimax polynomials without getting wrecked by out-of-range inputs?26 Can you split a zkML inference proof into independent layer chunks and get real prover speedups? zkComposer says yes.27 Replaces threshold signatures with ordinary approvals plus a Shamir-sealed enforcement key, so custody survives PQ migration.28 QR codes get a PKI makeover here: EdDSA, CBOR certs, ZLIB, JWKS, and a trust registry squeezed into scan-size limits.29 TRACE: A Two-Channel Robust Attribution Watermark via Complementary Embeddings for LLM-Agent Trajectories arxiv.orgTRACE watermarks LLM-agent trajectories with two keyed channels, trying to survive log deletion and rewriting, which is the annoying part.30 Mini-Programs, Mega-Problems: Unveiling OAuth-based Authentication Misuses in Mini-Programs via Dynamic Analysis arxiv.orgPaper on MINIAUTH, a dynamic crawler for mini-program OAuth bugs, with tens of thousands of apps and one delightful Baidu key-bruteforce...