Knowledge catalogue
This is what production RL infra looks like. ICYMI: RL training at scale separates into two distinct problems. - Tight collective comms for the trainer. - Distributed async inference for rollout. Kudo
arXiv:2607.01469v2 Announce Type: replace Abstract: Agentic Video Question Answering (VideoQA) systems invoke tools during inference, but their tool libraries are fixed, so recurring procedures are re
arXiv:2412.12640v2 Announce Type: replace Abstract: The increasing demand for data privacy, alongside the benefits of aggregating data from networked devices, has catalyzed the emergence of federated
arXiv:2607.05153v1 Announce Type: cross Abstract: Scientists often seek to draw causal inferences from structured data that is not independently and identically distributed, such as spatial data, netw
arXiv:2607.03971v1 Announce Type: cross Abstract: Causality has become an increasingly important tool for gaining a deeper understanding of complex systems. Among various causal analysis methods, caus
arXiv:2204.13492v5 Announce Type: replace Abstract: We present StreamDEQ, a method that aims to infer frame-wise representations on videos with minimal per-frame computation. Conventional deep network
arXiv:2607.00363v1 Announce Type: cross Abstract: Flow Matching (FM) has emerged as a powerful paradigm for speech generation but remains constrained by high inference latency and timbre leakage. To a
arXiv:2607.01179v1 Announce Type: cross Abstract: Scaling inference compute, by generating many parallel attempts per problem, is a costly but reliable lever for improving language model capabilities.
arXiv:2607.00325v1 Announce Type: cross Abstract: A growing body of literature suggests that training data membership inference problems are fundamentally hard tasks in modern language modeling settin
arXiv:2510.09340v2 Announce Type: replace Abstract: Recent language models exhibit significant logical reasoning abilities, yet the mechanisms supporting deductive inference remain poorly understood.
arXiv:2606.30217v1 Announce Type: new Abstract: Large multimodal models have achieved strong reasoning on complex visual tasks, but their inference efficiency is often restricted by long chains of tho
arXiv:2511.17038v4 Announce Type: replace Abstract: From a Bayesian perspective, score-based diffusion solves inverse problems through joint inference, embedding the likelihood with the prior to guide
arXiv:2512.07287v3 Announce Type: replace-cross Abstract: As intents unfold and environments change, multi-turn agents face continuously shifting decision contexts. Although reusing past experience is
arXiv:2606.30352v1 Announce Type: new Abstract: Recent advances in 3D scene reconstruction have highlighted the intricate trade-offs among rendering quality, inference efficiency, and data dependency.
arXiv:2606.28831v1 Announce Type: cross Abstract: Long-context LLM inference faces a fundamental conflict: head-adaptive compression algorithms (e.g., Top-p nucleus sampling) offer superior accuracy b
arXiv:2606.30017v1 Announce Type: new Abstract: Recent advances in 3D Gaussian Splatting have demonstrated unprecedented success in novel view synthesis. However, the substantial inference and storage
arXiv:2512.06208v3 Announce Type: replace-cross Abstract: Inference of standard convolutional neural networks (CNNs) on FPGAs often incurs high latency and a long initiation interval due to the deep n
arXiv:2606.27771v1 Announce Type: cross Abstract: Reinforcement learning (RL) post-training improves the reward alignment of flow-based generators, but often degrades perceptual quality in ways that a
arXiv:2606.27978v1 Announce Type: cross Abstract: Pixel-space continuous-token autoregressive (AR) generation directly models images as sequences of raw pixel patches, avoiding discrete tokenization o
arXiv:2606.26657v1 Announce Type: new Abstract: Identifying high-utility candidates from massive discrete spaces under expensive evaluations is a recurring challenge across the sciences, with structur
arXiv:2606.25985v1 Announce Type: new Abstract: Vision-language-action (VLA) models have shown strong potential for general-purpose robot manipulation, but their inference latency remains a major obst
arXiv:2606.24450v1 Announce Type: cross Abstract: Perceiving physical contact is fundamental to dexterous manipulation. While robots often rely on dedicated hardware tactile sensors, humans exhibit a
arXiv:2606.23370v1 Announce Type: cross Abstract: Device-side Large Language Models (LLMs) have grown explosively, offering stronger privacy and higher availability than their cloud-side counterparts.
arXiv:2606.21670v1 Announce Type: cross Abstract: We describe our entry to the efficiency track of the Academic Text-to-Music (ATTM) Grand Challenge at ICME 2026. Beyond the challenge protocol's FAD-C
arXiv:2510.04646v2 Announce Type: replace Abstract: Flow matching models generate high-fidelity molecular geometries but incur significant computational costs during inference, requiring hundreds of n
arXiv:2606.11831v1 Announce Type: cross Abstract: Neural relational inference (NRI) methods discover interaction graphs from trajectories through variational reasoning on discrete potential edges. How
arXiv:2512.22219v2 Announce Type: replace-cross Abstract: We introduce Mirage Persistent Kernel (MPK), the first compiler and runtime system that automatically transforms multi-GPU model inference int
arXiv:2606.09316v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) enables agents to access external knowledge at inference time, but it primarily retrieves fragmented declarative ev
arXiv:2601.22574v2 Announce Type: replace-cross Abstract: Although Video Large Multimodal Models have achieved strong performance in video understanding, they still suffer from hallucination. Existing
arXiv:2606.06574v1 Announce Type: new Abstract: Large language models (LLMs) perform inference by following a fixed depth and order, non-recurrent execution of all layers. We reveal the wide existence
BitNet b1.58 uses ternary weights (-1, 0, 1) and achieves performance comparable to full-precision transformers , enabling efficient LLM inference on CPUs and edge devices. While research into efficie
arXiv:2606.05936v1 Announce Type: new Abstract: Modern language models rely on pretraining filters to remove undesirable content from training corpora and inference-time guardrails to suppress undesir
arXiv:2606.05533v1 Announce Type: cross Abstract: Existing robot planning systems rely on appearance-based reasoning, where visual observations are encoded into latent spaces organized around object a
arXiv:2606.06467v1 Announce Type: new Abstract: Long-context inference in modern LLMs is increasingly constrained by decoding efficiency, especially in reasoning-heavy settings where models generate l
arXiv:2606.04434v1 Announce Type: new Abstract: Multimodal In-Context Learning (ICL) has emerged as a practical inference paradigm for Multimodal Large Language Models, where a small set of interleave
arXiv:2606.04351v1 Announce Type: cross Abstract: Processing video in vision-language models is expensive: each frame occupies hundreds of tokens, and inference cost scales with every frame and every
arXiv:2602.03024v2 Announce Type: replace-cross Abstract: Deep Equilibrium Models (DEQs) have emerged as a powerful paradigm in deep learning, offering the ability to model infinite-depth networks wit
arXiv:2606.02483v1 Announce Type: cross Abstract: Tool-augmented language agents speculatively issue likely future tool calls to hide latency, but those calls leak inferred user intent to external ser
arXiv:2606.00716v1 Announce Type: new Abstract: Inference and control in engineered physical systems pay a heavy physics cost at deployment: state estimators, inverse-problem solvers, model-predictive
arXiv:2606.00240v1 Announce Type: new Abstract: Effective real-world assistance requires AI agents with robust Theory of Mind (ToM): inferring human mental states from their behavior. Despite recent a
arXiv:2606.01595v1 Announce Type: new Abstract: Bayesian inference provides a principled framework for modeling epistemic uncertainty in neural networks by treating predictions as distributions rather
arXiv:2605.31314v1 Announce Type: new Abstract: The diffusion based robot navigation world models are typically trained using parallel supervision, while autoregressive inference is employed during pa
arXiv:2605.30637v1 Announce Type: new Abstract: Clinical decision-making (CDM) is central to real-world clinical workflows, where clinicians infer diagnoses, select treatments, or anticipate future he
arXiv:2605.30728v1 Announce Type: new Abstract: Machine learning (ML) training and inference often process data sets far exceeding GPU memory capacity, forcing them to rely on PCIe for on-demand tenso
arXiv:2603.09453v3 Announce Type: replace-cross Abstract: Foundation models are increasingly being deployed in contexts where understanding the uncertainty of their outputs is critical to ensuring res
B9441 is a release from llama.cpp, which provides LLM inference in C/C++ . The release follows the project's versioning pattern used by the ggml-org/llama.cpp repository on GitHub, typically containin
arXiv:2605.30131v1 Announce Type: new Abstract: Radiology report generation (RRG) is commonly formulated as a single-path generation task, where a multimodal large language model (MLLM) produces one d
arXiv:2605.29005v1 Announce Type: cross Abstract: Diffusion-based neural solvers for combinatorial optimization repeatedly re-evaluate dense edge/factor interactions, making inference expensive in wal
arXiv:2605.30102v1 Announce Type: cross Abstract: The design space of agentic AI inference spans two extremes: frontier large language models (LLMs), typically hosted in the cloud and offering strong
B9381 is a build/release tag for llama.cpp, a C/C++ framework for LLM inference. This release likely includes bug fixes, performance optimizations, and feature improvements to the project, which enabl
arXiv:2510.02329v2 Announce Type: replace-cross Abstract: Speculative decoding accelerates LLM inference by verifying candidate tokens from a draft model against a larger target model. Recent judge de
arXiv:2605.28422v1 Announce Type: cross Abstract: Latent reasoning enables reasoning over continuous hidden states rather than explicit tokens, avoiding the language bottleneck and inference overhead
arXiv:2605.25156v1 Announce Type: cross Abstract: Prediction models trained under the source distribution do not generalize well to a different target distribution. A valid inference about an unseen d
arXiv:2605.25537v1 Announce Type: new Abstract: Real-time chunking (RTC) lets chunked action policies operate under inference delay by conditioning a newly generated action chunk on actions already co
arXiv:2510.01384v4 Announce Type: replace Abstract: A natural desideratum for generative models is self-correction--detecting and revising low-quality tokens at inference. While Masked Diffusion Model
Over the last 25 years of building Google’s global network, we’ve navigated major architectural eras — from the Internet, to streaming, and the cloud. Today, we are squarely in the midst of a fourth:
arXiv:2605.24597v1 Announce Type: new Abstract: Many applications of large language models (LLMs) require deductive reasoning, yet models frequently produce incorrect or redundant inference steps. We
arXiv:2605.25107v1 Announce Type: cross Abstract: Existing work on population dynamics inference often focuses on flows arising from vector fields that are the gradients of scalar potentials. Among al
arXiv:2602.20191v2 Announce Type: replace-cross Abstract: Dynamic runtime latency and memory constraints necessitate flexible large language model (LLM) deployment, where an LLM can be inferred with v
b9296 is a build release of llama.cpp , the C/C++ implementation for efficient large language model inference. As an intermediate build in the llama.cpp release cycle, it likely includes recent bug fi