Pretraining frontier-scale LLMs in FP8 is now standard practice, but moving to 4-bit floating point has remained an open research problem because narrower formats compress dynamic range and amplify quantization error at long token horizons. A new research from NVIDIA describes a pretraining methodology built around NVFP4, a 4-bit microscaling format supported natively by Blackwell Tensor Cores, and validates it by pretraining a 12-billion-parameter hybrid Mamba-Transformer on 10 trillion tokens. The research team state this is the longest […]

Ver mais

arXiv:2602.12284v1 Announce Type: new Abstract: Timely classification of humanitarian information from social media is critical for effective disaster response. However, deploying large language models (LLMs) for this task faces challenges in resource-constrained emergency settings. This paper develops a lightweight, cost-effective framework for disaster tweet classification using parameter-efficient fine-tuning. We construct a unified experimental corpus by integrating and normalizing the HumAID dataset (76,484 tweets across 19 disaster events) into a dual-task benchmark: humanitarian information categorization and event type identification. […]

Ver mais

arXiv:2602.20166v1 Announce Type: new Abstract: In many applications involving intelligent agents, the overwhelming volume of alerts (mostly false) generated by the agents may desensitize users and cause them to overlook critical issues, leading to the so-called ”alert fatigue”. A common strategy is to train a reflection model as a filter to intercept false alerts with labelled data collected from user verification feedback. However, a key challenge is the noisy nature of such data as it is often collected […]

Ver mais

arXiv:2603.25963v1 Announce Type: new Abstract: Localization in GNSS-denied and GNSS-degraded environments is a challenge for the safe widespread deployment of autonomous vehicles. Such GNSS-challenged environments require alternative methods for robust localization. In this work, we propose BEVMapMatch, a framework for robust vehicle re-localization on a known map without the need for GNSS priors. BEVMapMatch uses a context-aware lidar+camera fusion method to generate multimodal Bird’s Eye View (BEV) segmentations around the ego vehicle in both good and adverse weather […]

Ver mais

The Axios team have published a full postmortem on the supply chain attack which resulted in a malware dependency going out in a release the other day, and it involved a sophisticated social engineering campaign targeting one of their maintainers directly. Here’s Jason Saayman’a description of how that worked: so the attack vector mimics what google has documented here: https://cloud.google.com/blog/topics/threat-intelligence/unc1069-targets-cryptocurrency-ai-social-engineering they tailored this process specifically to me by doing the following: they reached out masquerading as the founder […]

Ver mais

arXiv:2604.16914v1 Announce Type: new Abstract: Clinical ultrasound analysis demands models that generalize across heterogeneous organs, views, and devices, while supporting interpretable workflow-level analysis. Existing methods often rely on task-wise adaptation, and joint learning may be unstable due to cross-task interference, making it hard to deliver workflow-level outputs in practice. To address these challenges, we present USTri, a tri-stage ultrasound intelligence pipeline for unified multi-organ, multi-task analysis. Stage I trains a universal generalist USGen on different domains to learn […]

Ver mais

arXiv:2602.20323v1 Announce Type: new Abstract: Reliable object manipulation requires understanding physical properties that vary across objects and environments. Vision-language model (VLM) planners can reason about friction and stability in general terms; however, they often cannot predict how a specific ball will roll on a particular surface or which stone will provide a stable foundation without direct experience. We present PhysMem, a memory framework that enables VLM robot planners to learn physical principles from interaction at test time, without […]

Ver mais

Six months ago I implemented a NumPy environment for generals.io and trained an agent that hit top 20 on human leaderboards. I reached 5k fps with that setup. In the last couple of days I rewrote everything in JAX with help from opus4.5 (here we go again) and got 1.4M FPS on single H200, which is a 300x speedup! I’m confident that with so much more fps going super-human is much more attainable! For those interested in coding […]

Ver mais

arXiv:2601.15442v1 Announce Type: cross Abstract: The unification of neural and symbolic approaches to artificial intelligence remains a central open challenge. In this work, we introduce a tensor network formalism, which captures sparsity principles originating in the different approaches in tensor decompositions. In particular, we describe a basis encoding scheme for functions and model neural decompositions as tensor decompositions. The proposed formalism can be applied to represent logical formulas and probability distributions as structured tensor decompositions. This unified treatment […]

Ver mais

We introduce $k$-lazyGD, an online learning algorithm that bridges the gap between greedy Online Gradient Descent (OGD, for $k=1$) and lazy GD/dual-averaging (for $k=T$), creating a spectrum between reactive and stable updates. We analyze this spectrum in Smoothed Online Convex Optimization (SOCO), where the learner incurs both hitting and movement costs. Our main contribution is establishing that laziness is possible without sacrificing hitting performance: we prove that $k$-lazyGD achieves the optimal dynamic regret $mathcal{O}(sqrt{(P_T+1)T})$ for any laziness slack […]

Ver mais