#01Oct 21, 2026
cs.CL
LongZero: Enhancing Long-Context Reasoning in LLMs from Zero Data
Zhennan Wan, Jianfei Chen
Reinforcement learning (RL) is an effective way to strengthen the long-context reasoning of large language models (LLMs), but it needs training data that are long, diverse, challenging, and verifiable. Human annotation of such data is expensive, naturally occurring data are too scarce to sustain RL, and existing synthesis methods depend on external corpora or stronger teacher models, which become harder to obtain as the trained models get stronger. Recently, RL from zero data, in which a model learns from tasks it proposes itself, has achieved excellent performance, but existing methods target short-context tasks. Therefore, we propose $\textbf{LongZero}$, a method that lets an LLM improve its own long-context reasoning without any external data. Instead of writing questions and answers over existing documents, LongZero first has the model synthesize the question and answer with its supporting evidence in a compact structured form, and only then separately expands the evidence into a long multi-document context. LongZero combines these self-written multi-hop reasoning tasks with self-written multi-round retrieval tasks and trains the model with GRPO. We evaluate LongZero on two LLMs across five long-context reasoning benchmarks and compare it with RL on data synthesized from external corpora. Experimental results show that LongZero can generalize to real tasks, and outperforms both baselines, suggesting a viable and promising path to improving long-context reasoning of LLMs.
#02Oct 21, 2026
cs.CR
VX-One: Typed, Calibrated Decision Heads for Real-Time Perimeter Triage - An empirical study of unknown-attack honesty, cascade gray zones, and calibration under drift
Juan S. Santillana
Perimeter triage faces a latency/semantics trade-off: signature engines are fast and format-safe but blind to unseen attacks, while generative models are semantic but slow and produce unparseable output. In late 2026 a class of commercial System One decision models (TypeSafe Jev; Microsoft-Decision-1) began selling typed decision primitives - binary gates, closed-set choices, ordinal scores - as product APIs, with no academic treatment. We present VX-One: a small multi-head network (100K-10M parameters) whose outputs are these primitives - a calibrated gate, a choice over a closed candidate set that always includes an explicit unknown option, and an ordinal score, trained multi-task, temperature-calibrated per head (~0.7-1.7 ms p50 on CPU). We evaluate it on UNSW-NB15, CIC-IDS2017, EMBER2024 against calibrated XGBoost, Random-Forest, MLP, Kitsune, and measure a two-phase cascade (Suricata with 52,674 Emerging-Threats rules -> VX-One) over five days of CIC-IDS2017. Findings: (i) VX-One wins detection (AUROC/FPR@95%TPR) on UNSW-NB15 and under temporal drift on CIC-IDS2017 (0.711 vs 0.637), but XGBoost wins multi-class attribution on all three benchmarks; (ii) leave-one-family-out gate detects never-seen families on 9/9 UNSW folds yet drops below chance on CIC recon family, and abstention threshold transfers poorly across benchmarks; (iii) on payload-less reconstructions signature engines resolve ~0.1% of flows - gray zone ~99.9% of traffic, recovered by Phase 2 (combined recall >=0.978); (iv) calibration collapses under temporal drift (ECE 0.04->0.40) for every method while ranking partially survives. We provide a measured account of where typed calibrated decisions help perimeter triage, and where they fail.
#03Oct 21, 2026
cs.DC
The End of Consensus Research?
Clément Burgelin, Rachid Guerraoui, Zicong Hong and 5 more
Consensus is a fundamental problem in distributed computing and at the core of state machine replication (SMR). Despite decades of active human research, discovering practical low-latency solutions to the problem is still open, as it requires combining protocol knowledge, correctness reasoning, implementation, and experimental evaluation. We report on how AI helped us discover A-greement, a new consensus protocol for geo-replicated state machines that significantly outperforms state-of-the-art solutions. Averaged across six deployments with 5 or 9 replicas, A-greement reduces the mean latency of a geo-distributed key-value store over the fastest baseline by 12.2% and 33.3% on write-heavy and read-heavy workloads, respectively, and matches the fastest baseline on read-only workloads. A-greement combines lease-authorized local reads, direct client learning for leader-ordered writes, and a fast fixed-quorum per-key write path, selected according to network topology and workload. At the heart of our discovery lies AI-greement, an agent harness for the "autonomous" discovery, implementation, experimentation, and verification of candidate protocols. AI-greement integrates (1) a concept-oriented protocol knowledge base, (2) a modular implementation skeleton, and (3) a Lean-based verification framework with reusable safety definitions, reference specifications, and proofs. AI-greement is of independent interest: its Lean formalization uncovers several previously unreported vulnerabilities in existing consensus protocols, and it provides guidelines on how domain-specific harnesses can turn general-purpose coding agents into effective tools for (distributed) computing research. Yet we argue that human consensus experts remain essential before, during, and after the search: agents reshape the researcher's role rather than end consensus research.
#04Oct 21, 2026
cs.CL
Looped Uniform Diffusion Language Models
Saiteja Utpala, Chieh-Hsin Lai, Yuki Mitsufuji and 1 more
Uniform diffusion language models (UDLMs) generate text by iteratively denoising sequences of ordinary vocabulary tokens, any of which may be corrupted and subsequently revised. During training, UDLMs must denoise from a context that may itself be unreliable, whereas masked diffusion language models (MDLMs) always condition on a clean context. This challenging training objective makes UDLMs more FLOP-intensive to train than MDLMs, but it yields stronger downstream reasoning performance. Recently, selective looping, which recursively applies a few early Transformer layers within each denoising step, has been shown to improve training efficiency in MDLMs under an iso-FLOP budget. We investigate looping in UDLMs while keeping the training objective and sampler unchanged. We find that looping significantly improves training efficiency: looped UDLMs reach the validation NLL of a non-looped UDLM trained on about 0.5 trillion tokens with up to 2.8x fewer training FLOPs on OpenWebText and 3.3x fewer on FineWeb-Edu. We also find that, at a matched NFE budget, looped UDLMs reach up to 9.8% lower generation perplexity than the non-looped UDLM, with most gains achieved at moderate loop counts. Finally, we study different loop-count distributions and find that sampling smaller loop counts more frequently during training improves inference efficiency: on downstream tasks, this strategy uses 12% fewer loops than uniform loop-count sampling without degrading accuracy.
#05Oct 21, 2026
eess.IV
Geometric Algebra for Neural Networks: A Survey
Tao Zhong, Kaison Fong, Keqin Wang and 1 more
Geometric algebra (GA) unifies geometric primitives and transformations in a single algebraic framework. Neural networks designed around this framework often represent features as multivectors and constrain layer operations to respect the algebraic structure, enabling principled treatment of geometric transformations and, in many cases, built-in equivariance. This survey reviews geometric algebra neural networks, including their building blocks, equivariant properties, application domains, empirical performance, and software support. We provide an algebra-selection guide, an empirical comparison of GA-based and other equivariant architectures, and a failure-mode analysis covering training pathologies, deployment constraints, and performance limitations. Across the surveyed literature, the strongest evidence for GA-based architectures appears in settings with explicit geometric symmetries and heterogeneous geometric entities; evidence for broader advantages in parameter efficiency, interpretability, and dense prediction is comparatively limited and uneven. We therefore conclude that geometric algebra neural networks are not universally superior to well-tuned non-GA equivariant baselines, and outline open problems in algebra selection, numerical stability, scalable implementation, and reproducible evaluation.