每日AI论文精选-2026-10-26

AI 论文精选速递

2026年10月26日 · 追踪前沿研究

1

Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory

Prateek Chhikara, Dev Khant 等5人

Mem0, a memory-centric architecture with graph-based memory, enhances long-term conversational coherence in LLMs by efficiently extracting, consolidating, and retrieving information, outperforming existing memory systems in terms of accuracy and computational efficiency.

#Mem0#memory-centric architecture#graph-based memory

https://arxiv.org/abs/2504.19413

https://hjfy.top/arxiv/2504.19413

https://github.com/mem0ai/mem0 (67799)

2

HuggingFace's Transformers: State-of-the-art Natural Language Processing

Thomas Wolf, Lysandre Debut 等22人

Transformers library provides state-of-the-art Transformer architectures and pretrained models for natural language processing tasks with a unified API and emphasis on extensibility and robust deployment.

#Transformer architectures#pretrained models#unified API

https://arxiv.org/abs/1910.03771

https://hjfy.top/arxiv/1910.03771

https://github.com/huggingface/transformers (168166)

3

ARIS: Autonomous Research via Adversarial Multi-Agent Collaboration

Ruofeng Yang, Yongcan Li 等3人

ARIS is an open-source research harness that uses cross-model adversarial collaboration to ensure reliable long-term research outcomes through coordinated execution, orchestration, and assurance layers.

#LLMs#agent systems#model weights

https://arxiv.org/abs/2605.03042

https://hjfy.top/arxiv/2605.03042

https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep (18060)

4

Raven: The Harness of Harnesses for Composable Agentic Intelligence

EverMind AI

Raven is an open-source multi-agent ecosystem that autonomously builds and evolves modular harnesses, enabling cross-domain collaboration and expanding reliable task coverage through composable agent units.

#multi-agent ecosystem#modular harnesses#All-Domain Collaboration Network

https://arxiv.org/abs/2609.33439

https://hjfy.top/arxiv/2609.33439

https://github.com/EverMind-AI/Raven (5823)

5

PaddleOCR-VL-1.6: Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement ...

Zelun Zhang, Hongen Liu 等15人

PaddleOCR-VL-1.6 enhances document parsing performance through targeted data optimization and progressive post-training techniques, achieving state-of-the-art results on OmniDocBench v1.6.

#document parsing#data optimization#post-training

https://arxiv.org/abs/2606.03264

https://hjfy.top/arxiv/2606.03264

https://github.com/PaddlePaddle/PaddleOCR (91670)

6

UniAVC: Unified Audio-Video Creation via Progressive Training and Asymmetric Scheduling

Kien T. Pham, Yuhao Liu 等9人

UniAVC trains separate audio and video branches on abundant unimodal data before aligning them with scarce paired data, using independent noise schedules and asymmetric supervision to enable joint synthesis, cross-modal generation, and instruction-guided editing.

#cross-modal generation#noise schedule#denoising

https://arxiv.org/abs/2610.24791

https://hjfy.top/arxiv/2610.24791

https://github.com/HKUST-LongGroup/UniAVC (247)

7

RAG-Anything: All-in-One RAG Framework

Zirui Guo, Xubin Ren 等5人

RAG-Anything is a unified framework that enhances multimodal knowledge retrieval by integrating cross-modal relationships and semantic matching, outperforming existing methods on complex benchmarks.

#Retrieval-Augmented Generation#RAG#Large Language Models

https://arxiv.org/abs/2510.12323

https://hjfy.top/arxiv/2510.12323

https://github.com/HKUDS/RAG-Anything (24129)

8

OpenVLHarness: Real World Visual Agentic Tool-Use with Multimodal Memory

Yu Zhou, Cheng-Fu Yang 等11人

OpenVLHarness improves multimodal reasoning by integrating persistent memory, heterogeneous tool interfaces, and iterative evidence composition across 22 benchmarks.

#multimodal large language models#visual reasoning#tool interfaces

https://arxiv.org/abs/2610.26860

https://hjfy.top/arxiv/2610.26860

https://github.com/uclanlp/OpenVLHarness (255)

9

OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use

Hanzhang Zhou, Long Li 等15人

OpenGCUA introduces an open generalist computer-use agent that integrates visual and programmatic interfaces, harness-to-model decision transfer, and code execution to improve task performance and efficiency across models.

#generalist computer-use agent#visual and programmatic interfaces#harness-to-model decision transfer

https://arxiv.org/abs/2610.22200

https://hjfy.top/arxiv/2610.22200

https://github.com/Tongyi-MAI/OpenGCUA (497)

论文摘要由 AI 生成 · 仅供学习参考

生成时间: 2026-10-26 15:36

© 2026 AI 论文精选 · 共收录 9 篇

posted @ 2026-10-26 16:43  Tockm  阅读(28)  评论(0)    收藏  举报