research

Is Child-Directed Language Optimized for Word Learning? A Computational Study of Verb Meaning Acquisition

arXiv:2605.12047v1 Announce Type: new Abstract: Is child-directed language (CDL) optimized to support language learning, and which aspects of linguistic development does it facilitate? We investigate

Research

Is Monotonic Sampling Necessary in Diffusion Models?

DGX agent

arXiv:2605.11773v1 Announce Type: new Abstract: Diffusion models generate samples by iteratively denoising a Gaussian prior, traversing a sequence of noise levels that, in every published sampler, dec

researcharxiv-cs-lg

13 May 2026

Kairos: A Scalable Serving System for Physical AI

arXiv:2605.11381v1 Announce Type: new Abstract: Physical AI is experiencing rapid growth with frontier foundation models increasing its capabilities across general environments. Physical AI tasks are

researcharxiv-cs-ro

Language Modeling with Hyperspherical Flows

arXiv:2605.11125v1 Announce Type: new Abstract: Discrete Diffusion Language Models progressed rapidly as an alternative to autoregressive (AR) models, motivated by their parallel generation abilities.

Large Language Models for Causal Relations Extraction in Social Media: A Validation Framework for Disaster Intelligence

arXiv:2605.11348v1 Announce Type: new Abstract: During disasters, extracting causal relations from social media can strengthen situational awareness by identifying factors linked to casualties, physic

Latent Chain-of-Thought Improves Structured-Data Transformers

arXiv:2605.11262v1 Announce Type: new Abstract: Chain-of-thought and more broadly test-time compute are known to augment the expressive capabilities of language models and have led to major innovation

Learning Action Manifold with Multi-view Latent Priors for Robotic Manipulation

arXiv:2605.11832v1 Announce Type: new Abstract: This paper tackles spatial perception and manipulation challenges in Vision-Language-Action (VLA) models. To address depth ambiguity from monocular inpu

researcharxiv-cs-ro

Learning Ego-Centric BEV Representations from a Perspective-Privileged View: Cross-View Supervision for Online HD Map Construction

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.

Learning U-Statistics with Active Inference

arXiv:2605.11638v1 Announce Type: cross Abstract: U-statistics play a central role in statistical inference. In many modern applications, however, acquiring the labels required for U-statistics is cos

Learning Weakly Communicating Average-Reward CMDPs: Strong Duality and Improved Regret

arXiv:2605.11586v1 Announce Type: new Abstract: We study infinite-horizon average-reward constrained Markov decision processes (CMDPs) under the weakly communicating assumption. Our contributions are

LiBaGS: Lightweight Boundary Gap Synthesis for Targeted Synthetic Data Selection

arXiv:2605.11231v1 Announce Type: new Abstract: Synthetic data is useful only when the added samples fill missing parts of the training distribution that matter for the downstream task. We introduce L

Limits of Learning Linear Dynamics from Experiments

arXiv:2605.12010v1 Announce Type: new Abstract: Learning governing dynamics from data is a common goal across the sciences, yet it is only well-posed when the underlying mechanisms are identifiable. I

LoopUS: Recasting Pretrained LLMs into Looped Latent Refinement Models

arXiv:2605.11011v1 Announce Type: new Abstract: Looped computation shows promise in improving the reasoning-oriented performance of LLMs by scaling test-time compute. However, existing approaches typi

Lower bounds for one-layer transformers that compute parity

arXiv:2605.12171v1 Announce Type: new Abstract: This note shows that no self-attention layer post-processed by a rational function can sign-represent the parity function unless the product of the numb

Machine Learning for neutron source distributions

arXiv:2605.12165v1 Announce Type: cross Abstract: In light of the recent advancements in machine learning, we propose a novel approach to neutron source distribution estimation through the utilisation

MambaNetBurst: Direct Byte-level Network Traffic Classification without Tokenization or Pretraining

arXiv:2605.11034v1 Announce Type: cross Abstract: We present MambaNetBurst, a compact tokenizer-free byte-level sequence classifier for network burst classification based on a Mamba-2 backbone. In con

Manifold Sampling via Entropy Maximization

arXiv:2605.12338v1 Announce Type: new Abstract: Sampling from constrained distributions has a wide range of applications, including in Bayesian optimization and robotics. Prior work establishes conver

Mapping Embodied Affective Touch Strategies on a Humanoid Robot

arXiv:2605.11825v1 Announce Type: new Abstract: Affective touch in human-robot interaction is shaped not only by emotional intent, but also by robot embodiment, including touch location, physical cons

researcharxiv-cs-ro

Martingale-Consistent Self-Supervised Learning

arXiv:2605.11846v1 Announce Type: new Abstract: Self-supervised learning (SSL) is often deployed under changing information, such as shorter histories, missing features, or partially observed images.

MetaColloc: Optimization-Free PDE Solving via Meta-Learned Basis Functions

arXiv:2605.12368v1 Announce Type: new Abstract: Solving partial differential equations (PDEs) with machine learning typically requires training a new neural network for every new equation. This optimi

MieDB-100k: A Comprehensive Dataset for Medical Image Editing

arXiv:2602.09587v2 Announce Type: replace Abstract: The scarcity of high-quality data remains a primary bottleneck in adapting multimodal generative models for medical image editing. Existing medical

mimalloc: A new, high-performance, scalable memory allocator for the modern era

researchmicrosoft-research

mimalloc is an open-source, modern, scalable memory allocator that is a drop-in replacement for malloc and free. It is relatively small (~12K lines), with clear internal data structures, and is easy t

Minimax Rates and Spectral Distillation for Tree Ensembles

arXiv:2605.11841v1 Announce Type: cross Abstract: Tree ensembles such as random forests (RFs) and gradient boosting machines (GBMs) are among the most widely used supervised learners, yet their theore

Missingness-MDPs: Bridging the Theory of Missing Data and POMDPs

arXiv:2605.12262v1 Announce Type: cross Abstract: We introduce missingness-MDPs (miss-MDPs), a novel subclass of partially observable Markov decision processes (POMDPs) that incorporates the theory of

MIST: Reliable Streaming Decision Trees for Online Class-Incremental Learning via McDiarmid Bound

arXiv:2605.11617v1 Announce Type: new Abstract: Streaming decision trees are natural candidates for open-world continual learning, as they perform local updates, enjoy bounded memory, and static decis

Mobile Traffic Camera Calibration from Road Geometry for UAV-Based Traffic Surveillance

arXiv:2605.11900v1 Announce Type: new Abstract: Unmanned aerial vehicles (UAVs) can provide flexible traffic surveillance where fixed roadside cameras are unavailable, costly, or impractical. However,

MobileEgo Anywhere: Open Infrastructure for long horizon egocentric data on commodity hardware

arXiv:2605.05945v2 Announce Type: replace-cross Abstract: The recent advancement of Vision Language Action (VLA) models has driven a critical demand for large scale egocentric datasets. However, exist

Modeling Narrative Structure in Latin Epic Poetry with Automatically Generated Story Grammars

arXiv:2502.12276v2 Announce Type: replace Abstract: Computational methods for analyzing prose and poetry utilize word embeddings and other abstract representations that sometimes obscure context-rich

Molecular Design beyond Training Data with Novel Extended Objective Functionals of Generative AI Models Driven by Quantum Annealing Computer

arXiv:2602.15451v3 Announce Type: replace-cross Abstract: Deep generative modeling to stochastically design small molecules is an emerging technology for accelerating drug discovery and development. H

Monitoring access to piped water and sanitation infrastructure in Africa at disaggregated scales using satellite imagery and self-supervised learning

arXiv:2411.19093v4 Announce Type: replace Abstract: Access to drinking water and sanitation services is essential for health and well-being, yet large global disparities persist. Sustainable Developme

more demos on Interaction Models collaboratively doing system design, reading papers, fact-checking with live generative UI

researchsoumith-chintala--x

more demos on Interaction Models collaboratively doing system design, reading papers, fact-checking with live generative UI 1. (System design) - The Interaction Models see your screen and collaborates

Much of Geospatial Web Search Is Beyond Traditional GIS

arXiv:2605.11336v1 Announce Type: cross Abstract: Web search queries concern place far more often than existing labelling schemes suggest, yet the landscape of geospatial web search queries - what peo

Multi-Timescale Conductance Spiking Networks: A Sparse, Gradient-Trainable Framework with Rich Firing Dynamics for Enhanced Temporal Processing

arXiv:2605.11835v1 Announce Type: cross Abstract: Spiking neural networks (SNNs) promise low-power event-driven computation for temporally rich tasks, but commonly used neuron models often trade off g

Not Worth Mentioning? A Pilot Study on Salient Proposition Annotation

arXiv:2603.27358v2 Announce Type: replace Abstract: Despite a long tradition of work on extractive summarization, which by nature aims to recover the most important propositions in a text, little work

On Predicting the Post-training Potential of Pre-trained LLMs

arXiv:2605.11978v1 Announce Type: new Abstract: The performance of Large Language Models (LLMs) on downstream tasks is fundamentally constrained by the capabilities acquired during pre-training. Howev

On the Approximation Complexity of Matrix Product Operator Born Machines

arXiv:2605.11471v1 Announce Type: new Abstract: Matrix product operator Born machines (MPO-BMs) are tractable tensor-network models for probabilistic modeling, but their efficient approximation capabi

One-Step Generative Modeling via Wasserstein Gradient Flows

arXiv:2605.11755v1 Announce Type: cross Abstract: Diffusion models and flow-based methods have shown impressive generative capability, especially for images, but their sampling is expensive because it

Operator Spectroscopy of Trained Lattice Samplers

arXiv:2605.11199v1 Announce Type: cross Abstract: Trained lattice samplers are usually judged by the ensembles they generate. Here we instead analyze the trained field-space function itself: a flow-ma

Optimal Representations for Generalized Contrastive Learning with Imbalanced Datasets

arXiv:2605.11291v1 Announce Type: new Abstract: In this paper, we provide a computable characterization of the geometry of optimal representations in Contrastive Learning (CL) when the classes are imb

Optimistic Dual Averaging Unifies Modern Optimizers

arXiv:2605.11172v1 Announce Type: new Abstract: We introduce SODA, a generalization of Optimistic Dual Averaging, which provides a common perspective on state-of-the-art optimizers like Muon, Lion, Ad

ORBIT: Preserving Foundational Language Capabilities in GenRetrieval via Origin-Regulated Merging

arXiv:2605.12419v1 Announce Type: new Abstract: Despite the rapid advancements in large language model (LLM) development, fine-tuning them for specific tasks often results in the catastrophic forgetti

Oscillators Are All You Need: Irregular Time Series Modelling via Damped Harmonic Oscillators with Closed-Form Solutions

arXiv:2602.12139v2 Announce Type: replace Abstract: Transformers excel at time series modelling through attention mechanisms that capture long-term temporal patterns. However, they assume uniform time

OTT-Vid: Optimal Transport Temporal Token Compression for Video Large Language Models

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

OUI as a Structural Observable: Towards an Activation-Centric View of Neural Network Training

arXiv:2605.11570v1 Announce Type: new Abstract: Activation functions are what make deep networks expressive: without them, the model collapses to a linear map. Yet we still evaluate training mostly fr

Overparametrized models with posterior drift

arXiv:2506.23619v2 Announce Type: replace-cross Abstract: This paper investigates the impact of posterior drift on out-of-sample forecasting accuracy in overparametrized machine learning models. We do

PairDropGS: Paired Dropout-Induced Consistency Regularization for Sparse-View Gaussian Splatting

arXiv:2605.12072v1 Announce Type: new Abstract: Dropout-based sparse-view 3D Gaussian Splatting (3DGS) methods alleviate overfitting by randomly suppressing Gaussian primitives during training. Existi

Paper: http://arxiv.org/abs/2605.06546 HF: http://huggingface.co/papers/2605.06546 Blog: http://nousresearch.com/token-superposition

This paper investigates token superposition, a phenomenon where language models can encode multiple token representations simultaneously in a single position, enabling more efficient use of model capa

researchnous-research--x

Parabolic Position Encoding: Vision-Centric, Principled, Extrapolatable, General

arXiv:2602.01418v2 Announce Type: replace Abstract: We propose Parabolic Position Encoding (PaPE), a parabola-based position encoding for vision modalities in attention-based architectures. Given a se