Knowledge catalogue
arXiv:2609.30797v1 Announce Type: new Abstract: Text-based memory and context compression support reuse of past interactions. Resizing continuous memory changes the input to a frozen LLM, coupling cap
arXiv:2609.30962v1 Announce Type: new Abstract: Low-light image enhancement (LLIE) remains challenging for lightweight models because illumination restoration and color fidelity are difficult to optim
In July, AI agents running a security test escaped their sandbox and ended up inside @huggingface's servers. So today we're happy to be among @nvidia and @JensenHuang's partners on the release of the
arXiv:2609.30729v1 Announce Type: cross Abstract: Intrusion detectors for small Internet-of-Things (IoT) devices are usually compressed by pruning and judged by overall accuracy. We show that this hid
arXiv:2609.30765v1 Announce Type: new Abstract: Insurance reserve estimation is a fundamental actuarial task supporting premium pricing, solvency assessment, financial reporting, capital planning, and
arXiv:2602.11229v3 Announce Type: replace Abstract: Reliable physics simulation demands two capabilities that today's neural PDE solvers do not deliver together: generalization across heterogeneous PD
arXiv:2609.30539v1 Announce Type: cross Abstract: Split Gibbs sampling enables diffusion posterior inference for general nonlinear inverse problems by decoupling prior and likelihood computations, all
arXiv:2609.30734v1 Announce Type: new Abstract: Multi-agent LLM workflows use planning, execution, verification, and summarization to improve task performance, yet the value of each component depends
arXiv:2609.30943v1 Announce Type: new Abstract: Long-form technical text generation underpins knowledge-intensive workflows, yet remains challenging for large language models (LLMs) due to the need fo
arXiv:2609.30934v1 Announce Type: new Abstract: Rapid advances in AI-generated video (AIGV) have increased the risks posed by deceptive video manipulation. Unlike fully synthetic videos, manipulated v
arXiv:2502.19784v3 Announce Type: replace-cross Abstract: With over 500 languages in Nigeria, three languages - Hausa, Yoruba and Igbo spoken by more than 175 million people, account for about 65% of
arXiv:2609.30296v1 Announce Type: new Abstract: Neural networks are becoming popular in solving mathematical problems, but stochastic models do not provide mathematical exactness by themselves. This s
arXiv:2609.31458v1 Announce Type: cross Abstract: Transformers have become a central architecture for in-context learning (ICL), particularly through their state-of-the-art performance in large langua
Parsing forms still trips up the latest frontier VLMs. A form isn't text on a page. It's a set of fields, grouped into sections, each tied to a specific box. That's why forms need purpose-built parsin
arXiv:2609.31135v1 Announce Type: cross Abstract: Spatio-Temporal Video Grounding (STVG) aims to localize the spatio-temporal tube in a video corresponding to a natural language query. While recent me
arXiv:2609.30404v1 Announce Type: new Abstract: We present POIL, a point-based one-shot imitation learning framework with stable dynamical systems. While one-shot imitation avoids collecting extensive
arXiv:2609.30661v1 Announce Type: new Abstract: The main result presented is a formula for the population loss in the student-teacher kernel model that is applicable to shallow ReLU networks with bias
arXiv:2609.30469v1 Announce Type: new Abstract: Pretrained ASR systems perform poorly on noisy Broadcast Police Communication (BPC), hindering efforts to understand police decision-making. Pseudo-labe
arXiv:2609.30554v1 Announce Type: new Abstract: Large language models (LLMs) have recently demonstrated promising capabilities as in-context policy optimizers for Reinforcement Learning (RL), enabling
arXiv:2609.31235v1 Announce Type: new Abstract: Artificial neural networks (ANNs) usually represent neural transmission with fixed trainable weights during a training batch, which omits short-term cha
To say that we were surprised at the jump and suddenness of the capabilities of our models when it came to “cyber” or “swarming” or “message boards” or anything else related to the incidents is an und
arXiv:2602.00158v3 Announce Type: replace-cross Abstract: Probing studies what information is encoded in a frozen LLM's layer representations by training a lightweight predictor on top of them. Beyond
arXiv:2609.31558v1 Announce Type: new Abstract: Contrastive Language--Image Pretraining (CLIP) has emerged as a dominant vision backbone due to its strong transferability and zero-shot capabilities. H
retrieval is a decision problem, not just a similarity problem cool experiment from @assaf_elovic swapping embeddings for Jev in GPT Researcher - 73% vs 46% relevant context decision models will show
arXiv:2609.31201v1 Announce Type: new Abstract: Scaling studies typically represent training data by a single count of samples. For hierarchically and spatially structured data, however, the same numb
arXiv:2609.30861v1 Announce Type: new Abstract: Language-model agents increasingly improve by converting execution experience into reusable external skills. Yet repeated skill updates form a learning
arXiv:2609.30633v1 Announce Type: new Abstract: Successful training of deep neural networks is highly dependent on the distribution of the initial weights. If the weights are too large, network traini
arXiv:2509.25685v3 Announce Type: replace Abstract: We propose Structured-Diffuser, a diffusion planner that embeds task and motion structure directly into the noise model. Unlike standard diffusion-b
arXiv:2609.30478v1 Announce Type: new Abstract: Convolutional neural networks trained on ImageNet are known to exhibit a strong preference for local high-frequency texture, an inductive bias that tran
arXiv:2609.30957v1 Announce Type: new Abstract: The empirical success of pretraining large language models has inspired a deeper investigation into the underlying loss landscapes and the optimization
arXiv:2609.31298v1 Announce Type: cross Abstract: Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder for which early and accurate diagnosis is critical to improving long-term deve
arXiv:2609.30866v1 Announce Type: cross Abstract: In scanning microscopy, drift causes the specimen to be sampled at positions displaced from the nominal probe positions. This displacement alters the
We’re partnering with Nvidia and 100+ industry partners to build infrastructure that contains rogue AI agents. In this research, we gave 9 AI models root access inside SPACE and told them to break out
arXiv:2609.31193v1 Announce Type: new Abstract: Current Audio-Visual LLMs (AVLLMs) struggle with reasoning over videos featuring multi-speaker dialogues. In such videos, resolving 'who says what' is c
arXiv:2609.31293v1 Announce Type: cross Abstract: Speech-based screening is a promising, non-invasive approach for detecting Alzheimer's disease and related cognitive risks. However, models trained on
Hello everyone, I made a calculator for the theoretical HW roofline for decoding / prefill based on several parameters (LLM model architecture, quants, GPU, Memory, ..). It still a theoretical bound,
Tomek Korbak is a full time OpenAI employee working on misalignment. In his message here, he says AI continues to break out of their best secure environments. In reply to a comment, he says we all are
Distributed Training and Inference both involves having a fundamental understanding of how distributed systems work in general Distributed Parallelism Tensor Parallelism Pipeline Parallelism Model Par
i don’t get it fully but this looks cool We achieved a breakthrough at BTL. AI models are very linear. They think in one direction at a time, and when they're wrong they scramble and start again. For
Reasoning from scratch, round number 5! This time, talking about log-probability scoring (also a great fundamental concept for loss functions like cross-entropy in pre-training and distillation) and s
arXiv:2512.06607v2 Announce Type: replace-cross Abstract: Applying LLMs to predictive tasks in finance is challenging due to look-ahead bias resulting from their training on long time-series data. Thi
arXiv:2609.28955v1 Announce Type: new Abstract: Current Vision-Language-Action (VLA) models often struggle with high-precision robotic manipulation. We attribute this limitation primarily to their vis
arXiv:2609.29204v1 Announce Type: new Abstract: Vision-language-action (VLA) models offer strong local control and instruction following but often struggle with long-horizon tasks requiring persistent
arXiv:2609.29816v1 Announce Type: new Abstract: Recent years have witnessed major progress in joint audio-video generation. Existing models still suffer from limited per-modality fidelity, insufficien
blank cells in tables can mess up agent decision making many document OCR tools struggle with reading multiple blank cells in a document and cause values to be shifted from the original column and row
arXiv:2609.29773v1 Announce Type: new Abstract: Many real-world tasks (e.g., office workflows, scientific experimentation) require LLM agents to interact repeatedly with their environments for context
arXiv:2609.28599v1 Announce Type: cross Abstract: Flipping 60% of training labels from a single Byzantine client using label-flipping model poisoning self-degrades an attacker's own federated detectio
Can an LM, starting from random init (!!), learn to generate all of its pretraining data? Introducing Self-Play Pretraining with Zero Data. Two models start from random initialization: a generator pro
arXiv:2609.29536v1 Announce Type: new Abstract: Chest radiographs are routinely used to verify the position of catheters, tubes, and other support devices. Existing image models often return labels or
arXiv:2609.29556v1 Announce Type: new Abstract: Multimodal emotion recognition is a key research area in affective computing, with applications in sentiment analysis, intelligent customer service, and
arXiv:2609.29261v1 Announce Type: new Abstract: Obstacle-aided locomotion is a fundamental capability for snake robots to traverse complex environments. However, conventional rigid-link snake robots o
arXiv:2609.28796v1 Announce Type: cross Abstract: Current automated gait analysis for clinical applications relies on uninterpretable black-box classifiers. Although Vision-Language Models (VLMs) offe
arXiv:2609.29847v1 Announce Type: cross Abstract: The Brinkman penalisation method embeds boundary-value problems on complex domains into a simple computational box by modeling the solid region as a s
arXiv:2609.29016v1 Announce Type: cross Abstract: AI-driven scientific discovery accelerates research by autonomously developing solutions and designs. Large language model (LLM) agents support this p
arXiv:2609.29029v1 Announce Type: new Abstract: Onboard vision-language models could enable satellites to answer queries directly, but exhaustive tiled inference over high-resolution imagery is slow a
arXiv:2609.29327v1 Announce Type: cross Abstract: Selective classification enables a model to abstain from predictions on uncertain instances, but existing approaches typically reject them through con
arXiv:2609.29999v1 Announce Type: cross Abstract: Post-training quantization of vision--language models (VLMs) is typically assessed through aggregate task accuracy and memory savings, but preserving
arXiv:2609.29668v1 Announce Type: cross Abstract: Large language model agents increasingly automate data workflows, but end-to-end cloud data engineering and analytical execution require reliable coor
arXiv:2609.29995v1 Announce Type: cross Abstract: AI teaching assistants (AI TAs) backed by large language models (LLMs) and pedagogical guardrails are increasingly being integrated into programming c
arXiv:2609.28732v1 Announce Type: cross Abstract: Primary bone tumors are rare but clinically aggressive neoplasms whose diagnosis from radiographs is challenged by heterogeneous morphology, subtle le