Knowledge catalogue
arXiv:2605.12831v1 Announce Type: new Abstract: Inverse reinforcement learning (IRL), which infers reward functions from demonstrations, is a valuable tool for modeling and understanding decision-maki
arXiv:2605.13369v1 Announce Type: cross Abstract: Large language models (LLMs) are typically deployed with fixed parameters, and their performance is often improved by allocating more computation at i
arXiv:2605.13153v1 Announce Type: new Abstract: Temporal Knowledge Graph Reasoning (TKGR) aims at inferring missing (especially future) events from historical data. Current evaluation in TKGR uniforml
arXiv:2605.13384v1 Announce Type: new Abstract: Most models of machine teaching and learning assume the learner makes no errors in its internal deductive inference. However, humans and large language
arXiv:2605.13316v1 Announce Type: new Abstract: Action diffusion excels at high-fidelity action generation but incurs heavy computational costs owing to its iterative denoising nature. Despite current
arXiv:2605.12940v1 Announce Type: cross Abstract: Probabilistic Circuits (PCs) are deep generative models that support exact and efficient probabilistic inference. Yet in autoregressive language model
arXiv:2605.13470v1 Announce Type: new Abstract: Recent advances in AI have been primarily driven by large-scale neural architectures that excel at function approximation, rather than by tailored induc
arXiv:2605.13560v1 Announce Type: new Abstract: This work studies lung tumor growth prediction from sparse and irregular longitudinal computed tomography (CT) observations with measurement variability
arXiv:2605.13067v1 Announce Type: cross Abstract: As end-to-end robotic policies are progressively deployed in the real world to solve real tasks, they face a gap between the training and inference co
arXiv:2605.11436v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed on long-horizon tasks in partially observable environments, where they must act while inferring a
arXiv:2605.11727v1 Announce Type: cross Abstract: Vision-language models typically reason over post-ISP RGB images, although RGB rendering can clip, suppress, or quantize sensor evidence before infere
b9129 is a build release of llama.cpp, a C/C++ implementation for LLM inference . As part of the llama.cpp project's release cycle, this build represents an intermediate development version containing
arXiv:2605.11723v1 Announce Type: new Abstract: In this paper, we propose Concentrate and Concentrate (CaC), a coarse-to-fine anomaly reward model based on Vision-Language Models. During inference, it
arXiv:2605.11159v1 Announce Type: new Abstract: Knowledge graph completion (KGC) aims to automatically infer missing facts in multi-relational data by mapping entities and relations into continuous re
arXiv:2605.11683v1 Announce Type: new Abstract: Vision Transformers (ViTs) incur significant computational overhead due to the quadratic complexity of self-attention relative to the token sequence len
arXiv:2605.11567v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models predominantly adopt action chunking, i.e., predicting and committing to a short horizon of consecutive low-level act
arXiv:2602.09725v3 Announce Type: replace-cross Abstract: Remote KV cache reuse fetches KV cache for identical contexts from remote storage, avoiding recomputation, accelerating LLM inference. While i
arXiv:2602.01682v2 Announce Type: replace Abstract: We study online inverse linear optimization, also known as contextual recommendation, where a learner sequentially infers an agent's hidden objectiv
arXiv:2605.11414v1 Announce Type: new Abstract: While traditional time-series classifiers assume full sequences at inference, practical constraints (latency and cost) often limit inputs to partial pre
arXiv:2605.11179v1 Announce Type: cross Abstract: Many three-dimensional spatial fields are anisotropic, with directions of rapid and slow variation that need not align with the coordinate axes. Stand
arXiv:2605.12218v1 Announce Type: new Abstract: Bird's-eye-view (BEV) representations derived from multi-camera input have become a central interface for online high-definition (HD) map construction.
arXiv:2605.11402v1 Announce Type: new Abstract: Deep learning-based website fingerprinting has emerged as an effective technique for inferring the websites users visit. Although existing methods achie
arXiv:2605.11803v1 Announce Type: new Abstract: As Video Large Language Models (Video-LLMs) scale to longer and more complex videos, their inference cost grows rapidly due to the large volume of visua
arXiv:2605.12144v1 Announce Type: new Abstract: In visual localization, Absolute Pose Regression (APR) enables real-time 6-DoF camera pose inference from single images, yet critically depends on fine-
arXiv:2605.12464v1 Announce Type: new Abstract: Quantization has emerged as a standard technique for accelerating inference for generative models by enabling faster low-precision computations and redu
arXiv:2605.12208v1 Announce Type: cross Abstract: Approximate Bayesian inference typically revolves around computing the posterior parameter distribution. In practice, however, the main object of inte
arXiv:2605.12389v1 Announce Type: new Abstract: Segmenting small and sparse structures in large-scale images is fundamentally constrained by voxel-level, lattice-bound computation and extreme class im
arXiv:2605.11111v1 Announce Type: cross Abstract: Scientific Machine Learning (SciML) faces unique challenges for extreme-resolution data, with mitigations that often fail to scale or degrade the accu
arXiv:2605.11361v1 Announce Type: new Abstract: Inference-time reward alignment asks how to turn a pre-trained diffusion model with base law p into a sampler that favors a reward r while remaining clo
arXiv:2605.08692v1 Announce Type: cross Abstract: Post-training weight-only quantization to 4 bits is widely used to reduce the memory and compute costs of large language model inference. Existing PTQ
arXiv:2603.01743v2 Announce Type: replace Abstract: Anticipating future actions in videos is challenging, as the observed frames provide only evidence of past activities, requiring the inference of la
As enterprises push AI beyond the pilot stage, the cost and complexity of running inference at scale are forcing a fundamental rethink of how infrastructure is designed, governed and sourced, putting
arXiv:2605.09992v1 Announce Type: cross Abstract: Speculative decoding accelerates LLM inference by drafting future tokens with a small model, but drafter models degrade sharply under template perturb
Build b9114 introduces a NCCL-free AllReduce implementation for tensor-parallel inference that pipelines device-to-host copy, cross-GPU communication, and reduction in a single CUDA kernel. The releas
Build b9119 of llama.cpp introduces CUDA optimizations including an NCCL-free AllReduce implementation for tensor-parallel inference and updates to the llama-bench tool for managing reduction provider
arXiv:2505.24859v3 Announce Type: replace-cross Abstract: Steering vectors are a lightweight method for controlling text properties by adding a learned bias to language model activations at inference
arXiv:2605.10586v1 Announce Type: new Abstract: Learning a physical model from video data that can comprehend physical laws and predict the future trajectories of objects is a formidable challenge in
arXiv:2512.00175v3 Announce Type: replace-cross Abstract: Identifying causal effects in the presence of unmeasured variables is a fundamental challenge in causal inference, for which proxy variable me
arXiv:2605.10793v1 Announce Type: new Abstract: Large language models (LLMs) are costly to deploy due to their large memory footprint and high inference cost. Weight-activation quantization can reduce
arXiv:2602.05243v2 Announce Type: replace-cross Abstract: Transformers achieve strong accuracy but incur high compute and memory cost. Structured pruning reduces inference cost, but most methods rely
arXiv:2605.08177v1 Announce Type: cross Abstract: Parameter-efficient fine-tuning (PEFT) has become a practical route for adapting large language models to downstream tasks, with LoRA-style methods be
arXiv:2601.20251v3 Announce Type: replace-cross Abstract: Exhaustively evaluating many large language models (LLMs) on a large suite of benchmarks is expensive. We cast benchmarking as finite-populati
arXiv:2506.11578v4 Announce Type: replace Abstract: Recently, large language models (LLMs) have demonstrated strong performance, ranging from simple to complex tasks. However, while large models achie
arXiv:2605.08645v1 Announce Type: cross Abstract: Energy-based models (EBMs) provide a powerful and flexible way of learning a joint probability distribution over data by constructing an energy surfac
arXiv:2605.08267v1 Announce Type: cross Abstract: Enterprise AI backends increasingly admit heterogeneous execution requests across model deployment, inference, evaluation, data movement, and agentic
arXiv:2605.08398v1 Announce Type: cross Abstract: In this work, we show that Latent Flow-Matching (LFM) models are robust to different types of perturbations, including data reduction and model capaci
arXiv:2605.09218v1 Announce Type: cross Abstract: 3D scene understanding spans reasoning about free space, object grounding, hypothetical object insertions, complex geometric relationships, and integr
arXiv:2605.09438v1 Announce Type: new Abstract: Many features in pretrained Transformers span multiple layers: they emerge through stages of inference, persist in the residual stream, or are built joi
arXiv:2605.08392v1 Announce Type: new Abstract: Practical diffusion sampling is a numerical approximation problem: under a fixed inference budget, one must simulate a reverse-time ODE or SDE using onl
arXiv:2605.10108v1 Announce Type: new Abstract: Joint named entity recognition (NER) and relation extraction (RE) is a fundamental task in natural language processing for constructing knowledge graphs
arXiv:2605.09408v1 Announce Type: new Abstract: Link prediction (inferring missing or future connections between nodes in a graph) is a fundamental problem in network science with widespread applicati
arXiv:2605.09100v1 Announce Type: new Abstract: Text embedding and generative tasks are usually trained separately based on large language models (LLMs) nowadays. This causes a large amount of trainin
arXiv:2605.08832v1 Announce Type: new Abstract: Neural surrogate models for computational fluid dynamics (CFD) are typically trained as forward operators that map explicit problem specifications, such
arXiv:2605.09132v1 Announce Type: new Abstract: Vision--language models (VLMs) show promise for clinical decision support in radiology because they enable joint reasoning over radiological images and
arXiv:2602.04549v2 Announce Type: replace Abstract: 3D Gaussian Splatting (3DGS) revolutionized novel view rendering. Instead of inferring from dense spatial points, as implicit representations do, 3D
arXiv:2510.20797v2 Announce Type: replace-cross Abstract: Context compression reduces Transformer inference costs by replacing lengthy inputs with shorter pre-computed representations. It carries sign
arXiv:2605.09235v1 Announce Type: cross Abstract: One-step generative modeling has emerged as a leading approach to amortize the inference cost of diffusion and flow-matching models. Among distillatio
arXiv:2605.10830v1 Announce Type: new Abstract: The remarkable achievements of both generative models of 2D images and neural field representations for 3D scenes present a compelling opportunity to in
arXiv:2605.10664v1 Announce Type: cross Abstract: Activation steering controls language model behavior by adding directions to internal representations at inference time, but standard residual-stream
arXiv:2605.08840v1 Announce Type: new Abstract: Large language models (LLMs) face growing challenges in efficient generative inference due to the increasing memory demands of Key-Value (KV) caches, es