Knowledge catalogue
arXiv:2605.18654v1 Announce Type: cross Abstract: A fraud scorer needs to answer in under 2 ms. The best tabular foundation models (TFMs) take 151-1,275 ms on GPU. We close this gap by distilling the
arXiv:2605.17428v1 Announce Type: cross Abstract: Our preliminary experiments on gym-DSSAT maize irrigation tasks revealed that +/-2 degrees C temperature noise causes an 11.9% reduction in economic r
arXiv:2605.16311v1 Announce Type: new Abstract: Distributed training of large neural networks is bottlenecked by full-precision gradient communication and by coordinatewise optimizers that ignore the
arXiv:2605.17156v1 Announce Type: cross Abstract: Quantum error correction (QEC) is essential for building fault-tolerant quantum computers, requiring decoders that are simultaneously accurate, fast,
arXiv:2605.18389v1 Announce Type: new Abstract: Optimal transport provides a powerful framework for comparing measures while respecting the geometry of their support, but comes with an expensive compu
arXiv:2605.17019v1 Announce Type: new Abstract: Streaming video effect generation is highly desirable for live human-centric applications such as e-commerce streaming, entertainment, and vlogging, yet
arXiv:2605.16259v1 Announce Type: cross Abstract: While real-time image generation using diffusion models has advanced rapidly on NVIDIA GPUs, systematic optimization research on non-CUDA platforms su
arXiv:2502.20969v4 Announce Type: replace-cross Abstract: Retrieval-augmented generation (RAG) extends large language models (LLMs) with external data sources to enhance factual correctness and domain
arXiv:2605.17821v1 Announce Type: cross Abstract: Large Language Model (LLM) training is frequently interrupted by a heterogeneous spectrum of failures, from common GPU crashes to catastrophic cluster
arXiv:2605.18177v1 Announce Type: new Abstract: Query-based Vision Transformer segmentation models typically reconstruct dense spatial feature maps to predict masks, inheriting design patterns from co
arXiv:2605.17170v1 Announce Type: new Abstract: Agentic workloads have emerged as a major workload for LLM inference. They differ significantly from chat-only workloads, requiring long-context process
arXiv:2510.06388v2 Announce Type: replace Abstract: Calibrated predictions are useful because their numerical values can be interpreted as probabilities. Calibration errors are therefore widely used t
arXiv:2605.17613v1 Announce Type: cross Abstract: The large size of the KV cache has become a major bottleneck for serving LLMs with increasing context lengths. In response, many KV cache compression
Vultr expanded its global cloud infrastructure by opening a new data center region in Milan, Italy, marking the company's 33rd cloud data center location worldwide. This expansion provides European cu
We are releasing Carbon: a crazy fast DNA model Carbon is 275x faster than the next best model. So fast you can process the whole human genome on a single GPU in <2 days. Here are the tricks we used:
arXiv:2605.15617v1 Announce Type: cross Abstract: Large language model (LLM) training today runs on clusters spanning thousands of GPUs. While this scale enables rapid model advances, developing, debu
arXiv:2605.15475v1 Announce Type: new Abstract: We introduce an empowered transposed Fully Connected Weighted (t-FCW) graph representation to embed point clouds into a metric space. While original t-F
arXiv:2605.16076v1 Announce Type: cross Abstract: Plant disease detection is still largely manual in Bangladesh, where extension workers eyeball leaf samples across millions of smallholdings. We built
B9208 is a build release of llama.cpp, an open-source C/C++ project that enables efficient large language model inference on diverse hardware platforms. The llama.cpp project focuses on optimized LLM
b9221 is an intermediate build release from the llama.cpp project, which is a C/C++ implementation enabling efficient LLM inference on consumer hardware. The release includes platform-specific binarie
arXiv:2605.15652v1 Announce Type: cross Abstract: Vector-HaSH and the Tolman-Eichenbaum Machine (TEM) propose that the hippocampal-entorhinal circuit factorizes content from a prestructured grid-cell
Robbie Whelan / Wall Street Journal: Decart, which offers real-time generative video and GPU optimization tech, raised 300M at a ~4B valuation, up from 3.1B after raising 153M in August 2025 — Decart'
arXiv:2605.15824v1 Announce Type: new Abstract: Human-centric video customization, particularly at the garment level, has shown significant commercial value. However, existing approaches cannot suppor
This guide demonstrates how to fine-tune NVIDIA's Cosmos Predict 2.5 video generation model using parameter-efficient techniques like LoRA (Low-Rank Adaptation) and DoRA (Mixture of Experts-based adap
Flux-genotype is a self-evolving AI kernel designed to run on CPU hardware using Ollama, featuring the capability to mutate and adapt its own architecture dynamically. This project demonstrates an app
arXiv:2604.04310v2 Announce Type: replace Abstract: In robot control, planning, and learning, there is a need for rigid-body dynamics libraries that are highly performant, easy to use, and compatible
Alessandra Ram / Wired: How the AI industry's intense pressure to keep up and its financial rewards are creating “walls of resentment” between Silicon Valley workers and their spouses — Are you marrie
Agentic AI inference at one-tenth the cost per token with NVIDIA Vera Rubin NVL72. Agent sandboxes run 50% faster on NVIDIA Vera than traditional CPUs — while enterprise data queries are up to 3x fast
Eduard Kovacs / SecurityWeek: Pwn2Own Berlin 2026: participants earned a total of $1,298,250 for 47 vulnerabilities, with successful exploits of AI products like Codex, Cursor, and LM Studio — Partici
Together with SpaceXAI, we’re training a significantly larger model from scratch, using 10x more total compute. With Colossus 2’s million H100-equivalents and our combined data and training techniques
The first NVIDIA Vera CPUs arrived at three of the world's leading AI labs on Friday — Anthropic in San Francisco, OpenAI in Mission Bay, SpaceXAI in Palo Alto — followed by a delivery to Oracle Cloud
Cerebras Systems filed for a $60 billion IPO, marking a significant milestone for the AI hardware startup that specializes in large-scale AI chip systems. The article likely discusses Cerebras' journe
Setting a GPU power limit using nvidia-smi reduces heat output by approximately 20% with only a 5-8% inference speed loss. Undervolting the GPU can reduce power consumption by 5-15% with zero performa
Deedy / @deedydas: SF vibes are frenetic over the huge divide in outcomes and career uncertainty for software engineers; over 5 years ~10K people in AI attained retirement wealth — The vibes in SF fee
Zinnia Lee / Forbes: Shenzhen-listed RoboTechnik, which claims to be the largest silicon photonics tool maker and whose stock is up 340% over the past year, files for a HK listing — RoboTechnik Intell
Rebecca Bellan / TechCrunch: A profile of AI video generation startup Runway, which is training models directly on observational data, is now valued at 5.3B, and added 40M in ARR in Q2 — Every major A
arXiv:2605.14438v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) architectures enhance the efficiency of large language models by activating only a subset of experts per token. However, standa
arXiv:2605.14526v1 Announce Type: cross Abstract: Differentiable simulation of soft bodies is a foundation for system identification, trajectory optimization, and Real2Sim transfer. Yet, existing meth
arXiv:2605.13942v1 Announce Type: new Abstract: Machine learning (ML) is increasingly applied to optimize system performance in tasks such as resource management and network simulation. Unlike traditi
arXiv:2605.14249v1 Announce Type: new Abstract: We present EnergyLens, an end-to-end framework for energy-aware large language model (LLM) inference optimization. As LLMs scale, predicting and reducin
GridCare Inc., a startup that helps data center operators more quickly connect their facilities to the electrical grid, has raised 64 million in funding. Early Nvidia Corp. backer Sutter Hill Ventures
arXiv:2603.11045v2 Announce Type: replace-cross Abstract: Inverse problems for stiff parabolic partial differential equations (PDEs), such as the inverse heat conduction problem (IHCP), are severely i
Meaghan Tobin / New York Times: Nvidia's future in China remains unclear after the Trump-Xi Summit; Trump says China “chose not to” buy Nvidia chips as “they want to try to develop their own” — The st
arXiv:2605.14277v1 Announce Type: new Abstract: Parallelization has played an instrumental role in the field of artificial intelligence (AI), drastically reducing the time taken to train and evaluate
arXiv:2605.14188v1 Announce Type: cross Abstract: What does a book look like to a quantum computer? This paper takes eight classical works of the Renaissance and its late-antique inheritance -- from A
Mauro Orru / Wall Street Journal: Semiconductor stocks fell globally on Friday after the Trump-Xi summit concluded without major chip deals; Nvidia closed down 4.42% and AMD closed down 5.69% — Beijin
Lee Jaewoon / The Elec: Seoul-based WIRobotics, which develops wearable and humanoid robots and is collaborating with Nvidia and AWS, raised a ~$68M Series B led by JB Investment — Company to accelera
arXiv:2605.14889v1 Announce Type: cross Abstract: Online surgical phase recognition (SPR) underpins context-aware operating-room systems and requires committing to a prediction at every frame from pas
arXiv:2605.14090v1 Announce Type: cross Abstract: We put forth a critical theoretical framework for analyzing generative models both descriptively and normatively. Our thesis is that generative models
arXiv:2605.13981v1 Announce Type: cross Abstract: The rise in deployment of large language models has driven a surge in GPU demand and datacenter scaling, raising concerns about electricity use, grid
arXiv:2605.14578v1 Announce Type: new Abstract: Partial Dependence Plots (PDPs) visualize how changes in a single feature affect the average model prediction. They are widely used in practice to inter
almost all LLM “thought” is derivative. below just sounds like hype to me 📁 Jensen Huang, CEO of NVIDIA, says GPT was never just about generating images or text but about generating thought itself. Th
U.K.-based artificial intelligence inference chip startup Fractile Ltd. said today it has closed on a 220 million Series B round of funding. The company was founded in 2022 by the Oxford University-tr
arXiv:2605.12649v1 Announce Type: new Abstract: Dataset distillation aims to synthesize a compact proxy dataset that is unreadable or non-raw from the original dataset for privacy protection and highl
This post appears to show Elon Musk in a photo with NVIDIA CEO Jensen Huang, Apple CEO Tim Cook, U.S. President Trump, Chinese President Xi Jinping, and members of a U.S. delegation, though the specif
arXiv:2605.13247v1 Announce Type: new Abstract: Sparse Mixture-of-Experts (MoE) models offer a powerful way to scale model size without increasing compute, as per-token FLOPs depend only on k active e
arXiv:2603.15854v2 Announce Type: replace-cross Abstract: Sampling from a categorical distribution is mathematically simple, but in large-vocabulary decoding, it often triggers extra memory traffic an
arXiv:2605.13108v1 Announce Type: new Abstract: Face presentation attack detection (FacePAD) remains challenging under diverse spoofing representation, including 2D print and replay, 3D mask-based spo
The NVIDIA Vera Rubin Platform is a rack-scale AI supercomputer designed to power agentic AI and reasoning models at scale by eliminating bottlenecks in communication and memory movement for efficient
A utility tool was created to streamline sharing GPU profile traces by uploading them to Hugging Face buckets using the command `uvx trace-tuil -b `. This simplifies the process of distributing perfor