wovepaper — the arXiv observatory

The paper presents Metis, a memory foundation model that embeds a persistent, dynamically updated memory state within the model backbone, allowing it to store and retrieve informat…

#memory-augmented models#foundation models#native memory#multimodal learning

eess.IV2025

Medical Artificial Intelligence for Early Detection of Lung Cancer: A Survey

Guohui Cai, Ying Cai, Zeyu Zhang +5

cs.CL2025

Role-Play Paradox in Large Language Models: Reasoning Performance Gains and Ethical Dilemmas

Jinman Zhao, Zifan Qian, Linbo Cao +5

cs.RO2024

Flight Structure Optimization of Modular Reconfigurable UAVs

Yao Su, Ziyuan Jiao, Zeyu Zhang +4

cs.CL2025

Generative Meta-Learning for Zero-Shot Relation Triplet Extraction

Wanli Li, Tieyun Qian, Yi Song +4

cs.CV2026

AnyDepth: Depth Estimation Made Easy

Zeyu Ren, Zeyu Zhang, Wukai Li +2

cs.RO2021

Reconstructing Interactive 3D Scenes by Panoptic Mapping and CAD Model Alignments

Muzhi Han, Zeyu Zhang, Ziyuan Jiao +4

cs.RO2022

Understanding Physical Effects for Effective Tool-use

Zeyu Zhang, Ziyuan Jiao, Weiqi Wang +3

cs.AI2026

MemOps: Benchmarking Lifecycle Memory Operations in Long-Horizon Conversations

Xixuan Hao, Zeyu Zhang, Zehao Lin +6

The paper introduces MemOps, a benchmark that evaluates long‑term conversational memory by tracking explicit lifecycle operations (remember, forget, update, etc.) rather than only…

Publications (266)

MWM: Mobile World Models for Action-Conditioned Consistent Prediction

A Multidimensional Self-Adaptive Numerical Simulation Framework for Semiconductor Boltzmann Transport Equation

EvoVLA: Self-Evolving Vision-Language-Action Model

Enhancing Signed Graph Neural Networks through Curriculum-Based Training

PresentAgent-2: Towards Generalist Multimodal Presentation Agents

Metis: Memory Foundation Model

Medical Artificial Intelligence for Early Detection of Lung Cancer: A Survey

Role-Play Paradox in Large Language Models: Reasoning Performance Gains and Ethical Dilemmas

Flight Structure Optimization of Modular Reconfigurable UAVs

Generative Meta-Learning for Zero-Shot Relation Triplet Extraction

AnyDepth: Depth Estimation Made Easy

Reconstructing Interactive 3D Scenes by Panoptic Mapping and CAD Model Alignments

Understanding Physical Effects for Effective Tool-use

MemOps: Benchmarking Lifecycle Memory Operations in Long-Horizon Conversations

ESA: Annotation-Efficient Active Learning for Semantic Segmentation

Temporal Leakage in LLM Backtesting: Measurement, Validation, and Adjusted Scores

FPSAttention: Training-Aware FP8 and Sparsity Co-Design for Fast Video Diffusion

Characterizing Demand Graphs for (Fixed-Parameter) Shallow-Light Steiner Network

How Order-Sensitive Are LLMs? OrderProbe for Deterministic Structural Reconstruction

Gyral-Sulcal-Net: An Integrated Network Representation of Brain Folding Patterns

VaseMuseum: Digital Intelligent Museum for Ancient Greek Pottery

Robust Core-Periphery Constrained Transformer for Domain Adaptation

SSS: Semi-Supervised SAM-2 with Efficient Prompting for Medical Imaging Segmentation

LLMTaxo: Leveraging Large Language Models for Constructing Taxonomy of Factual Claims from Social Media

Disorder-assisted Robustness of Ultrafast Cooling in High Doped CVD-Graphene

Dynamic Execution Commitment of Vision-Language-Action Models

DELTAMEM: Incremental Experience Memory for LLM Agents via Residual Trees

LLM3:Large Language Model-based Task and Motion Planning with Motion Failure Reasoning

DeepPlantCRE: A Transformer-CNN Hybrid Framework for Plant Gene Expression Modeling and Cross-Species Generalization

Partial Symbol Recovery for Interference Resilience in Low-Power Wide Area Networks

UniMesh: Unifying 3D Mesh Understanding and Generation

On the Emergence of Symmetrical Reality

V-Retrver: Evidence-Driven Agentic Reasoning for Universal Multimodal Retrieval

A Reconfigurable Data Glove for Reconstructing Physical and Virtual Grasps

Towards Error-Free Long Video Generation

Generalizing Knowledge Graph Embedding with Universal Orthogonal Parameterization

DOEI: Dual Optimization of Embedding Information for Attention-Enhanced Class Activation Maps

FDG-Diff: Frequency-Domain-Guided Diffusion Framework for Compressed Hazy Image Restoration

SegStitch: Multidimensional Transformer for Robust and Efficient Medical Imaging Segmentation

Multi-Relation Graph-Kernel Strengthen Network for Graph-Level Clustering

CoV: Chain-of-View Prompting for Spatial Reasoning

Motion-R1: Enhancing Motion Generation with Decomposed Chain-of-Thought and RL Binding

Efficient Learning With Sine-Activated Low-rank Matrices

Lineage Evolution Reinforcement Learning

KnowTrace: Bootstrapping Iterative Retrieval-Augmented Generation with Structured Knowledge Tracing

An STREL-based Formulation of Spatial Resilience in Cyber-Physical Systems

TrustMap: Mapping Truthfulness Stance of Social Media Posts on Factual Claims for Geographical Analysis

Mitigating Degree Bias in Signed Graph Neural Networks

Less Detail, Better Answers: Degradation-Driven Prompting for VQA

In Situ Answer Sentence Selection at Web-scale

X-VoE: Measuring eXplanatory Violation of Expectation in Physical Events

MedLoCoMo: A Long-Context Multi-Session Medical Dialogue Benchmark for Large Language Models

Trustworthy GNNs with LLMs: A Systematic Review and Taxonomy

MARL-MambaContour: Unleashing Multi-Agent Deep Reinforcement Learning for Active Contour Optimization in Medical Image Segmentation

MemSim: A Bayesian Simulator for Evaluating Memory of LLM-based Personal Assistants

Hazards in Daily Life? Enabling Robots to Proactively Detect and Resolve Anomalies

GeneralVLA-2: Geometry-Aware Reconstruction and Governed Memory for Robot Planning

PresentAgent: Multimodal Agent for Presentation Video Generation

Prompt and Parameter Co-Optimization for Large Language Models

Human-Robot Interaction in a Shared Augmented Reality Workspace

StereoAdapter-2: Globally Structure-Consistent Underwater Stereo Depth Estimation

FDC: Fast KV Dimensionality Compression for Efficient LLM Inference

ManipLVM-R1: Reinforcement Learning for Reasoning in Embodied Manipulation with Large Vision-Language Models

Evaluation of OpenAI o1: Opportunities and Challenges of AGI

BIFE: Better Interaction, Fewer Errors for Minute-Long Video Generation

RecBole 2.0: Towards a More Up-to-Date Recommendation Library

HACK: Homomorphic Acceleration via Compression of the Key-Value Cache for Disaggregated LLM Inference

Robust Deep Signed Graph Clustering via Weak Balance Theory

SuperMag: Vision-based Tactile Data Guided High-resolution Tactile Shape Reconstruction for Magnetic Tactile Sensors

3D CoCa: Contrastive Learners are 3D Captioners

HiLoRA: Adaptive Hierarchical LoRA Routing for Training-Free Domain Generalization

Vision-EKIPL: External Knowledge-Infused Policy Learning for Visual Reasoning

TMCIR: Token Merge Benefits Composed Image Retrieval

MemBench: Towards More Comprehensive Evaluation on the Memory of LLM-based Agents

SegTTA: Training-Free Test-Time Augmentation for Zero-Shot Medical Imaging Segmentation

RPD-Diff: Region-Adaptive Physics-Guided Diffusion Model for Visibility Enhancement under Dense and Non-Uniform Haze

Spatial-Temporal Conditional Random Field for Human Trajectory Prediction

MoRL: Reinforced Reasoning for Unified Motion Understanding and Generation

Stealthy-Hyperuniform Wave Dynamics in Two-Dimensional Photonic Crystals