All entries

Position: Life-Logging Video Streams Make the Privacy-Utility Trade-off Inevitable

arXiv:2605.10404v1 Announce Type: new Abstract: With the growing prevalence of always-on hardware such as smart glasses, body cameras, and home security systems, life-logging visual sensing is becomin

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Position: Stop Evaluating AI with Human Tests, Develop Principled, AI-specific Tests instead

DGX agent

arXiv:2507.23009v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have achieved remarkable results on a range of standardized tests originally designed to assess human cognitive a

model-releasesarxiv-cs-ai

12 May 2026

Positional Encoding via Token-Aware Phase Attention

arXiv:2509.12635v3 Announce Type: replace-cross Abstract: We prove under practical assumptions that Rotary Positional Embedding (RoPE) introduces an intrinsic distance-dependent bias in attention scor

Positional LSH: Binary Block Matrix Approximation for Attention with Linear Biases

arXiv:2605.09472v1 Announce Type: new Abstract: Positional encoding in transformers is commonly implemented through positional embeddings, attention masks, or bias terms, but formal connections betwee

safetyarxiv-cs-lg

Positive Alignment: Artificial Intelligence for Human Flourishing

arXiv:2605.10310v1 Announce Type: new Abstract: Existing alignment research is dominated by concerns about safety and preventing harm: safeguards, controllability, and compliance. This paradigm of ali

Post-hoc Selective Classification for Reliable Synthetic Image Detection

arXiv:2605.08574v1 Announce Type: new Abstract: As synthetic images become increasingly realistic, reliable synthetic image detection techniques are of pressing need to prevent their misuse. Despite s

Power Reinforcement Post-Training of Text-to-Image Models with Super-Linear Advantage Shaping

arXiv:2605.10937v1 Announce Type: new Abstract: Recently, post-training methods based on reinforcement learning, with a particular focus on Group Relative Policy Optimization (GRPO), have emerged as t

local-aiarxiv-cs-cv

PowerStep: Memory-Efficient Adaptive Optimization via ell_p-Norm Steepest Descent

arXiv:2605.10335v1 Announce Type: cross Abstract: Adaptive optimizers, most notably Adam, have become the default standard for training large-scale neural networks such as Transformers. These methods

PPU-Bench:Real World Benchmark for Personalized Partial Unlearning in Vision Language Models

arXiv:2605.08800v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) may memorize sensitive cross-modal information during pretraining. However, existing MLLM unlearning benchmar

Practical Scaling Laws: Converting Compute into Performance in a Data-Constrained World

arXiv:2605.09189v1 Announce Type: new Abstract: The scaling laws guiding modern model training were calibrated for a single regime: data-rich, single-epoch pretraining. The dominant such scaling law f

Practical Wi-Fi-based Motion Recognition Under Variable Traffic Patterns

arXiv:2605.08308v1 Announce Type: cross Abstract: Wi-Fi sensing detects human motions and activities by analysing the channel state information (CSI) derived from Wi-Fi transmissions. However, the imp

PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning

arXiv:2602.03190v3 Announce Type: replace-cross Abstract: Reinforcement learning algorithms such as group-relative policy optimization (GRPO) have shown strong potential for improving the mathematical

Tutorials

Predicting 3D structure by latent posterior sampling

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

tutorialsarxiv-cs-cv

Predicting Large Model Test Losses with a Noisy Quadratic System

arXiv:2605.09154v1 Announce Type: new Abstract: We introduce a predictive model that estimates the pre-training loss of large models from model size (N), batch size (B) and number of weight updates (K

Predicting Plasticity in Deep Continual Learning: A Theoretical Perspective

arXiv:2605.09044v1 Announce Type: new Abstract: Deep continual learning requires models to adapt to new tasks without retraining from scratch. However, neural networks can lose their ability to adapt

Prediction Bottlenecks Don't Discover Causal Structure (But Here's What They Actually Do)

arXiv:2605.09169v1 Announce Type: cross Abstract: A Mamba state-space model trained only for next-step prediction appears to recover Granger-causal structure through a simple readout S = |W_{out} W_{i

Predictive Radiomics for Evaluation of Cancer Immune SignaturE in Glioblastoma: the PRECISE-GBM study

arXiv:2605.10278v1 Announce Type: new Abstract: Background: Radiogenomics allows identification of radiological biomarkers for genomic phenotypes. In glioblastoma, these biomarkers could potentially c

PrepBench: How Far Are We from Natural-Language-Driven Data Preparation?

arXiv:2605.08687v1 Announce Type: cross Abstract: Data preparation is a central and time-consuming stage in data analysis workflows. Traditionally, commercial tools have relied on graphical user inter

Preserve and Personalize: Personalized Text-to-Image Diffusion Models without Distributional Drift

model-releasesarxiv-cs-cv

arXiv:2505.19519v3 Announce Type: replace Abstract: Personalizing text-to-image diffusion models involves integrating novel visual concepts from a small set of reference images while retaining the mod

Preserving Foundational Capabilities in Flow-Matching VLAs through Conservative SFT

model-releasesarxiv-cs-ro

arXiv:2605.08879v1 Announce Type: new Abstract: Unconstrained fine-tuning of flow-matching Vision-Language-Action (VLA) models drives dense parameter overwrites, degrading pre-trained capabilities. We

Pretraining large language models with MXFP4

arXiv:2605.09825v1 Announce Type: cross Abstract: Why does full-pipeline FP4 training of large language models often diverge, even when forward activations and activation gradients remain stable? We a

Agents

Preventing Prompt Injection with Type-Directed Privilege Separation

arXiv:2509.25926v2 Announce Type: replace-cross Abstract: Modern language models have enabled the development of agentic systems that achieve strong performance on reasoning-intensive tasks. Unfortuna

agentsarxiv-cs-lg

Preventing Rank Collapse in Federated Low-Rank Adaptation with Client Heterogeneity

arXiv:2602.13486v2 Announce Type: replace-cross Abstract: Federated low-rank adaptation (FedLoRA) has facilitated communication-efficient and privacy-preserving fine-tuning of foundation models for do

local-aiarxiv-cs-ai

Price of Quality: Sufficient Conditions for Sparse Recovery using Mixed-Quality Data

arXiv:2605.10713v1 Announce Type: cross Abstract: We study sparse recovery when observations come from mixed-quality sources: a small collection of high-quality measurements with small noise variance

PRIM: Meta-Learned Bayesian Root Cause Analysis

model-releasesarxiv-cs-lg

arXiv:2605.08786v1 Announce Type: new Abstract: Root cause analysis (RCA) in complex systems is challenging due to error propagation across multiple variables, the need for structural causal knowledge

Primal-Dual Guided Decoding for Constrained Discrete Diffusion

arXiv:2605.09749v1 Announce Type: new Abstract: Discrete diffusion models generate structured sequences by progressively unmasking tokens, but enforcing global property constraints during generation r

PrimeKG-CL: A Continual Graph Learning Benchmark on Evolving Biomedical Knowledge Graphs

arXiv:2605.10529v1 Announce Type: new Abstract: Biomedical knowledge graphs underwrite drug repurposing and clinical decision support, yet the upstream ontologies they depend on update on independent

Priming: Hybrid State Space Models From Pre-trained Transformers

arXiv:2605.08301v1 Announce Type: cross Abstract: Hybrid State-Space models combine Attention with recurrent State-Space Model (SSM) layers, balancing eidetic memory from Attention with compressed fad

Princeton faculty votes to require proctoring in all in-person exams starting this summer, reversing an 1893 policy amid concerns about AI-fueled cheating (Douglas Belkin/Wall Street Journal)

Douglas Belkin / Wall Street Journal: Princeton faculty votes to require proctoring in all in-person exams starting this summer, reversing an 1893 policy amid concerns about AI-fueled cheating — The c

safetytechmeme

Priority-Driven Control and Communication in Decentralized Multi-Agent Systems via Reinforcement Learning

model-releasesarxiv-cs-lg

arXiv:2605.10482v1 Announce Type: cross Abstract: Event-triggered control provides a mechanism for avoiding excessive use of constrained communication bandwidth in networked multi-agent systems. Howev

Tutorials

PriorVLA: Prior-Preserving Adaptation for Vision-Language-Action Models

arXiv:2605.10925v1 Announce Type: new Abstract: Large-scale pretraining has made Vision-Language-Action (VLA) models promising foundations for generalist robot manipulation, yet adapting them to downs

tutorialsarxiv-cs-ro

Agents

PRISM: Fast Online LLM Serving via Scheduling-Memory Co-design

arXiv:2605.08581v1 Announce Type: new Abstract: Modern online large language model (LLM) services, such as Retrieval-Augmented Generation (RAG) and agent systems, increasingly expose two prominent cha

agentsarxiv-cs-lg

PRISM: Generation-Time Detection and Mitigation of Secret Leakage in Multi-Agent LLM Pipelines

arXiv:2605.10614v1 Announce Type: new Abstract: Multi-agent LLM systems introduce a security risk in which sensitive information accessed by one agent can propagate through shared context and reappear

Privacy Auditing Synthetic Data Release through Local Likelihood Attacks

model-releasesarxiv-cs-lg

arXiv:2508.21146v2 Announce Type: replace Abstract: Auditing the privacy leakage of synthetic data is an important but unresolved problem. Existing privacy auditing frameworks for synthetic data rely

Privacy-Aware Video Anomaly Detection through Orthogonal Subspace Projection

arXiv:2605.08651v1 Announce Type: cross Abstract: Video anomaly detection (VAD) systems often prioritize accuracy while overlooking privacy concerns, limiting their suitability for real-world deployme

Privacy-Preserving Distributed Learning in IoT Systems: A Unified Threat Model and Evaluation Framework

arXiv:2605.09232v1 Announce Type: cross Abstract: The increasing deployment of Internet-of-Things (IoT) devices has accelerated the use of distributed learning frameworks, where data remains local whi

local-aiarxiv-cs-lg

Privacy-Preserving Federated Learning: Integrating Zero-Knowledge Proofs in Scalable Distributed Architectures

arXiv:2605.08152v1 Announce Type: cross Abstract: The intersection of Artificial Intelligence (AI) and distributed systems has given rise to Federated Learning (FL), a paradigm that enables decentrali

local-aiarxiv-cs-ai

Private Vertical Federated Inference for Time-Series

arXiv:2605.08343v1 Announce Type: new Abstract: Institutions may benefit from collaborative inference on time-series data. In settings where privacy is necessary, multi-party computation (MPC) is a st

ProactBench: Beyond What The User Asked For

arXiv:2605.09228v1 Announce Type: cross Abstract: Most LLM benchmarks score how well a model responds to explicit requests. They leave unmeasured a different conversational ability: noticing and actin

Probability-Flow Distillation: Exact Wasserstein Gradient Flow for High-Fidelity 3D Generation

arXiv:2605.09071v1 Announce Type: new Abstract: Score Distillation Sampling (SDS) and its variants have been widely used for text-to-3D generation by distilling 2D image diffusion priors. However, the

Probing Cross-modal Information Hubs in Audio-Visual LLMs

arXiv:2605.10815v1 Announce Type: new Abstract: Audio-visual large language models (AVLLMs) have recently emerged as a powerful architecture capable of jointly reasoning over audio, visual, and textua

Probing Routing-Conditional Calibration in Attention-Residual Transformers

arXiv:2605.09850v1 Announce Type: cross Abstract: Post-hoc calibration is usually evaluated as a function of logits or softmax confidence alone, even as routing-augmented architectures increasingly ac

Probing the Critical Point (CritPt) of AI Reasoning: a Frontier Physics Research Benchmark

arXiv:2509.26574v4 Announce Type: replace Abstract: While large language models (LLMs) with reasoning capabilities are progressing rapidly on high-school math competitions and coding, can they reason

Probing the Impact of Scale on Data-Efficient, Generalist Transformer World Models for Atari

arXiv:2605.08578v1 Announce Type: cross Abstract: Developing generalist systems that retain human-like data efficiency is a central challenge. While world models (WMs) offer a promising path, existing

Process Matters more than Output for Distinguishing Humans from Machines

arXiv:2605.06524v2 Announce Type: replace Abstract: Reliable human-machine discrimination is becoming increasingly important as large language models and autonomous agents are deployed in online setti

ProcVLM: Learning Procedure-Grounded Progress Rewards for Robotic Manipulation

arXiv:2605.08774v1 Announce Type: cross Abstract: Long-horizon robotic manipulation requires dense feedback that reflects how a task advances through its procedural stages, not merely whether the fina

safetyarxiv-cs-lg

ProDG: Prototypes for Data-Free Generative Post-Hoc Explainability

arXiv:2605.08858v1 Announce Type: new Abstract: Ante-hoc interpretability methods based on prototypes provide highly accurate explanations by utilizing the intuitive 'this looks like that' reasoning p

Product-of-Gaussian-Mixture Diffusion Models for Joint Nonlinear MRI Reconstruction

model-releasesarxiv-cs-cv

arXiv:2605.10629v1 Announce Type: new Abstract: Recently, diffusion models have attracted considerable attention for magnetic resonance image reconstruction due to their high sample quality. However,