Message Passing Enables Efficient Reasoning
arXiv:2607.01077v1 Announce Type: new Abstract: While inference-time scaling has improved the reasoning abilities of large language models (LLMs), the need to generate long chains-of-thought (CoTs) is
Knowledge catalogue
arXiv:2607.01077v1 Announce Type: new Abstract: While inference-time scaling has improved the reasoning abilities of large language models (LLMs), the need to generate long chains-of-thought (CoTs) is
arXiv:2607.00724v1 Announce Type: new Abstract: Multilingual fluency often invites a stronger assumption: a model that can speak a user's language must also understand the culture encoded by that lang
arXiv:2607.00461v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) are often constrained by a language-space bottleneck, forcing complex visual reasoning into discrete tokens whi
arXiv:2607.00647v1 Announce Type: new Abstract: Training-free guidance (TFG) steers a pretrained diffusion model toward a desired attribute at inference. To be effective, this guidance must be applied
As AI moves from model development to production inference, compute demand is accelerating and shifting toward continuously operating AI factories that generate tokens at scale. This shift requires ac
arXiv:2604.03118v2 Announce Type: replace Abstract: Distilling video generation models to extremely low inference budgets (e.g., 2--4 NFEs) is crucial for real-time deployment, yet remains challenging
arXiv:2607.00060v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) show strong promise for clinical VQA and radiology report generation, yet inference-time hallucinations still u
TTT-Discover is a method that enables models to learn and discover patterns during test time rather than only during training, allowing for adaptation to new data distributions at inference. The appro
Aurora is a unified training-serving system that integrates reinforcement learning with adaptive speculative training to optimize large language model inference and training efficiency. The system dyn
Untied Ulysses is a memory-efficient technique for context parallelism that processes attention heads in chunks rather than sequences, reducing memory overhead during transformer inference and trainin
arXiv:2606.31915v1 Announce Type: cross Abstract: While conformal prediction provides a general framework for uncertainty quantification in predictive inference, its application is often limited by co
Release b9852 is a build version of llama.cpp, a C/C++ implementation that enables large language model inference with minimal setup and state-of-the-art performance across diverse hardware platforms
arXiv:2606.30919v1 Announce Type: cross Abstract: Edge-cloud inference collaborations are often designed with a routing estimator that decides whether to offload each frame from weak models at the edg
arXiv:2606.31986v1 Announce Type: new Abstract: Chain-of-thought (CoT) reasoning has enabled multi-modal large language models (MLLMs) to tackle complex visual reasoning tasks by generating explicit i
arXiv:2603.06136v2 Announce Type: replace Abstract: Diffusion distillation is central to accelerating image and video generation, yet existing methods are fundamentally limited by the denoising proces
arXiv:2606.31982v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) incur prohibitive inference costs due to long visual token sequences. Training-free visual token reduction prov
Anthropic's agentic coding tool Claude Code has worked with Google Cloud for a while now. An individual developer could easily point CLAUDE_CODE_USE_VERTEX=1 at a Google Cloud (GCP) project, grant the
Flat-rate AI plans are subsidizing agentic workloads. Learn why LLM inference costs are moving to metered pricing and how evals reveal cost per successful task. The post Model subsidies are ending. Wh
Together AI announced a Series C funding round and shared details about their company vision, current capabilities, and future direction in open-source AI development and inference optimization. The a
arXiv:2606.31259v1 Announce Type: cross Abstract: Diffusion-based text-to-audio (TTA) models achieve impressive synthesis quality but suffer from high inference latency due to iterative multi-step den
arXiv:2606.30875v1 Announce Type: cross Abstract: Foundation model pseudo-labeling - labeling data strictly via zero-shot inference - enables massive scale, but performance is undermined by hallucinat
After 18 months of writing, coding, and experimenting, Build a Reasoning Model (From Scratch) is finally out! My first copies just arrived! 📚 440 full-color pages. Inference scaling, reinforcement lea
llama.cpp is a C/C++ implementation for LLM inference with minimal setup and state-of-the-art performance on a wide range of hardware. Release b9844 is an intermediate build of the llama.cpp project f
Release build b9846 is an intermediate build of llama.cpp, an open-source C/C++ project for LLM inference. The llama.cpp project is the main playground for developing new features for the ggml library
b9849 is a build release from llama.cpp, which enables LLM inference in C/C++ . The release represents an intermediate build version of the llama.cpp library, a widely-used open-source project for run
arXiv:2606.30457v1 Announce Type: new Abstract: We study behavior prompting, a paradigm that enables robots to perform new tasks at inference time given a single human demonstration, which we call a b
arXiv:2606.29441v1 Announce Type: cross Abstract: Inference-time safety methods for large language models have proliferated, yet no systematic comparison exists. We evaluate five defense paradigms (no
arXiv:2606.29801v1 Announce Type: new Abstract: Text-to-image diffusion models remain vulnerable to adversarial prompts that elicit disallowed content, motivating reliable inference-time controls. A p
arXiv:2606.28471v1 Announce Type: new Abstract: Model capability is the central variable in LLM pre-training, yet is never observed directly: data shapes it prospectively, while evaluation reveals it
Computational chemistry researchers have traditionally faced a frustrating trade-off when simulating molecular interactions: use fast classical force fields that sacrifice precision or rely on accurat
arXiv:2606.29164v1 Announce Type: cross Abstract: Latent reasoning models perform multi-step inference directly in hidden-state space, yet the structure of these latent reasoning trajectories remains
arXiv:2512.05663v3 Announce Type: replace Abstract: Real-time monocular 3D object detection remains challenging due to severe depth ambiguity, viewpoint shifts, and the high computational cost of 3D r
arXiv:2509.02292v2 Announce Type: replace Abstract: What if large language models could not only infer human mindsets but also expose every blind spot in team dialogue such as discrepancies in the tea
arXiv:2511.19119v2 Announce Type: replace Abstract: Spatial reasoning (SR), the ability to infer 3D spatial information from 2D inputs, is essential for real-world applications such as embodied AI and
arXiv:2606.28864v1 Announce Type: new Abstract: Test-time scaling is a paradigm where large models use additional compute at inference to achieve better performance, without changing model weights. Wh
arXiv:2606.29908v1 Announce Type: cross Abstract: Existing world model-based planners for visual navigation typically follow a verification-centric paradigm, decoupling goal intent from trajectory syn
arXiv:2508.02178v3 Announce Type: replace Abstract: Large reasoning models (LRMs) often exhibit overthinking, producing verbose Chain-of-Thought (CoT) traces that increase inference cost and obscure t
arXiv:2606.30355v1 Announce Type: cross Abstract: As real-world prediction systems often face missing modalities at inference, incomplete multimodal learning (IML) remains a practical challenge. While
arXiv:2606.29453v1 Announce Type: cross Abstract: Arbitrary-Scale Super-Resolution (ASR) reconstructs images at continuous magnification factors. Recent methods accelerate inference by replacing compu
arXiv:2511.10480v3 Announce Type: replace-cross Abstract: Optimizing the performance of large language models (LLMs) on large-scale AI training and inference systems requires a scalable and expressive
arXiv:2606.30374v1 Announce Type: cross Abstract: Multimodal MRI is essential for accurate brain tumor segmentation. However, acquiring all modalities at inference is often challenging in practice, wh
arXiv:2606.29565v1 Announce Type: new Abstract: A stateless inference server (vLLM, SGLang, TensorRT-LLM) idles between requests while the accelerator waits; a stateful session reclaims that idle time
arXiv:2606.21401v2 Announce Type: replace-cross Abstract: Agentic AI applications compose multiple model calls and tool executions, creating new scheduling challenges for GPU-CPU clusters. Their infer
arXiv:2606.30109v1 Announce Type: new Abstract: Vision-based tactile sensing converts contact-induced surface deformation into images, enabling robots to infer contact forces and fine surface textures
arXiv:2401.11512v2 Announce Type: replace-cross Abstract: Identifying the most suitable variables to represent the state is a fundamental challenge in Reinforcement Learning (RL). These variables must
arXiv:2510.06096v3 Announce Type: replace-cross Abstract: The objectives that Large Language Models (LLMs) implicitly optimize remain dangerously opaque, making trustworthy alignment and auditing a gr
Together AI presented research at ICML 2026 covering advances across the full technology stack, likely including foundational model improvements, inference optimization, and practical deployment solut
arXiv:2606.30372v1 Announce Type: new Abstract: Quantitative research across the social and behavioral sciences depends on human subject experiments that are expensive, slow, and subject to sampling b
arXiv:2606.30595v1 Announce Type: cross Abstract: Semantic communication (SemCom) aims to preserve semantic meaning and task-oriented information beyond conventional message recovery over wireless cha
b9840 is a release of llama.cpp, a tool for LLM inference in C/C++. The release enables running AI models locally on Windows, Linux, and macOS. This specific build tag likely contains bug fixes, perfo
arXiv:2606.27688v1 Announce Type: cross Abstract: In financial forecasting, predictive performance depends not only on which model is trained, but also on how the trained model is deployed. We study t
arXiv:2603.25168v2 Announce Type: replace Abstract: Previous works based on Segment Anything Model (SAM) have achieved promising performance in unified scene text detection and layout analysis. Howeve
arXiv:2606.27791v1 Announce Type: cross Abstract: Hybrid attention models that mix full and sliding-window attention across layers offer a promising approach to efficient long-context inference, but t
arXiv:2606.27575v1 Announce Type: new Abstract: Predicting how a scene will evolve after a desired 3D transformation from images is a central goal in vision, graphics, and robotics. Yet unlike ideal s
arXiv:2502.16886v4 Announce Type: replace-cross Abstract: To reduce memory consumption during LLM inference, a handful of methods have been proposed for KV cache pruning. While these techniques can ac
arXiv:2606.27807v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have become a dominant paradigm for embodied intelligence. However, most existing approaches are built on large-scal
arXiv:2606.27538v1 Announce Type: cross Abstract: We introduce the context-ready transformer, a new recurrent neural network architecture built from a D-layer transformer block that pre-contextualizes
arXiv:2605.25042v2 Announce Type: replace Abstract: Existing score-based methods for inverse problems often resort to approximate minimization of the KL divergence between the inversion distribution a
arXiv:2606.28301v1 Announce Type: new Abstract: Inference-time scaling is a promising paradigm to improve generative models, especially when outputs must satisfy structural constraints or optimize dow
llama.cpp b9830 is a release of the open-source LLM inference project enabling local language model execution with minimal setup . The release likely includes backend improvements and optimizations fo