1
又快又聪明，阿里发布Qwen-Audio-3.0-Realtime：实时语音大模型四项功能升级
阿里实时语音大模型四项功能同步升级，智商与口语抗噪双突破，堪称又快又聪明。
2
Do You Remember? Toward Memory-Centric Multimodal AI
人类记忆是重构的而非忠实记录，这篇论文提出三阶段架构DoYouRemember，让多模态大模型具备类人记忆能力，解决现有模型丢弃内部表征的缺陷。
arXiv:2607.11919v1 Announce Type: cross Abstract: Human memory is reconstructive, not a faithful recording. Current multimodal LLMs (MLLMs) lack this …
3
小米 MiMo 完成大语言、多模态、语音等全系列模型矩阵闭环，端侧模型通过国家大模型备案
IT之家 7 月 15 日消息，促进生成式人工智能服务创新发展和规范应用，网信部门会同有关部门按照《生成式人工智能服务管理暂行办法》要求，有序开展生成式人工智能服务备案工作， 现将新增的“Apple 智能”等 7 款提供手机端侧生成式人工智能服务备案信息予以公告 。 小米官方刚刚发文，介绍了此次通过…
4
Trojan Horse Prompting: Jailbreaking Conversational Multimodal Models by Forging Assistant Message
对话多模态模型安全新漏洞：通过伪造助手消息实现越狱攻击，揭示未被探索的攻击面。
arXiv:2507.04673v2 Announce Type: replace Abstract: The rise of conversational interfaces has greatly enhanced LLM usability by leveraging dialogue hi…
5
AeroRAG: Structured Multimodal Retrieval-Augmented LLM for Fine-Grained Aerial Visual Reasoning
新型多模态检索增强框架让大模型精准理解航拍图像细节，解决细粒度视觉推理难题。
arXiv:2604.17889v2 Announce Type: replace Abstract: Despite recent progress in multimodal large language models (MLLMs), reliable visual question answ…
6
SpurLens: Automatic Detection of Spurious Cues in Multimodal LLMs
提出自动检测多模态大模型中虚假线索的SpurLens方法，助力提升模型可靠性与可解释性。
arXiv:2503.08884v3 Announce Type: replace-cross Abstract: Unimodal vision models are known to rely on spurious correlations, but it remains unclear to…
7
Tool-MCoT: Tool Augmented Multimodal Chain-of-Thought for Content Safety Moderation
多模态链式推理+工具调用，提升内容安全审核的准确性与可解释性。
arXiv:2604.06205v2 Announce Type: replace-cross Abstract: The growth of online platforms and user content requires strong content moderation systems t…
8
Safe responses matter: Output-aware safety guardrail mitigate over-refusal in MLLMs
ECCV 2026新作：输出感知安全护栏，精准解决多模态大模型“宁可错杀一千”的过度拒绝问题。
arXiv:2607.09697v1 Announce Type: new Abstract: Existing safety mechanisms for multimodal large language models (MLLMs) face a fundamental trade-off b…
9
GigaChat Audio: Time-aware Large Audio Language Model
GigaChat Audio 是时间感知音频语言模型，精准理解音频内容与时间关系，提升语音识别与生成质量。
arXiv:2607.10387v1 Announce Type: cross Abstract: Temporal grounding in long recordings remains challenging for audio-conditioned LLMs. We present a t…
10
A Multimodal Dataset for Large Language Model Applications in the Energy Domain
首个面向能源领域的大模型多模态数据集，助力LLM在电力、能源管理等场景的应用突破。
arXiv:2607.11459v1 Announce Type: cross Abstract: This paper presents the mAIEnergy dataset, an open-access, multimodal corpus developed to support La…
11
Building Food Metadata with LLM Juries
DoorDash用LLM陪审团与上下文优化，多模态AI精准构建食品元数据，工程实践亮点。
Article URL: https://careersatdoordash.com/blog/building-food-metadata-with-llm-juries-context-optimization-multimodal-ai/ Comments URL: https://news.…
12
Evidence-Backed Video Question Answering
一篇关于视频问答的创新论文，强调利用证据推理来提升答案的准确性与可解释性。
arXiv:2607.11862v1 Announce Type: cross Abstract: Current Video Large Language Models (Video LLMs) excel in question answering (QA) but largely operat…
13
SigLIP-HD by Fine-to-Coarse Supervision
从粗到细监督实现高效高分辨率视觉表示，SigLIP-HD在处理多模态大模型时兼顾计算效率与细粒度特征。
arXiv:2607.09488v1 Announce Type: new Abstract: High-quality visual representation is a long-standing pursuit in computer vision. In the context of mu…
14
对话Om AI赵天成：多年坚守，押注物理AI原生的「流式」未来
独家对话Om AI创始人：端侧原生流式多模态架构首次打通物理AI闭环，让AI从云端“着陆”真实世界。
一个从未见过监控画面的多模态模型，却比在监控数据上练了多年的小模型“老将”更懂监控。这不是科幻电影，这是2023年Om AI联汇的一场“无心插柳”，也是CEO兼首席科学家赵天成博士更加坚信“多模态训练方式能为物理开放世界带来泛化性”的关键节点。彼时，AI行业正在追求以大语言模型为核心的生成式AI。 …
15
Meta 发布多模态推理模型 Muse Spark 1.1，强化 AI 智能体任务能力
Meta新模型Muse Spark 1.1主打多智能体协作与百万级上下文，让AI能拆解任务并行执行，效率倍增。
IT之家 7 月 12 日消息，Meta 于 7 月 9 日正式发布适用于 AI 智能体的多模态推理模型 Muse Spark 1.1 版本，重点提升了模型在智能体任务中的规划、协同与执行能力，并增强了工具调用、代码开发、应用操作能力。 Meta 表示，Muse Spark 1.1 强化了多智能体协…
16
Muse Spark 1.1 by Meta AI
Meta AI发布多模态推理模型Muse Spark 1.1，专为智能体任务打造，提升AI自主决策能力。
Multimodal reasoning model built for agentic tasks Discussion | Link
17
AT-Attn: Temporal-Aware Cross-Attention for Longitudinal Multimodal Alzheimer's Disease Diagnosis
提出时间感知交叉注意力机制，提升纵向多模态阿尔茨海默病诊断精度，AI医疗前沿突破。
arXiv:2607.07091v1 Announce Type: cross Abstract: In longitudinal Alzheimer's disease (AD) diagnosis support, clinical and imaging information is ofte…
18
Latent graph encoding of multimodal neuroimaging features with generative AI architectures
多模态神经影像与生成式AI结合，用潜在图编码实现跨模态特征融合，前沿研究亮相ICIP 2026。
arXiv:2607.07027v1 Announce Type: cross Abstract: While generative models enable encoding of complex neuroimaging data for feature generation and reco…
19
End-to-End LLM Flight Planning with RAG-based Memory and Multi-modal Coach Agent
将LLM与RAG记忆及多模态教练代理结合，实现端到端航班规划，一个新研究被ICML 2026 Workshop接收。
arXiv:2607.06964v1 Announce Type: cross Abstract: Bridging the gap between human pilot intent and autonomous flight operation is critical for real-wor…
20
HoloCount: A Holistic Visual Counting Benchmark for MLLMs
多模态大模型视觉计数能力新基准，HoloCount全面评测模型在复杂场景下的计数表现，揭示当前MLLMs的瓶颈。
arXiv:2607.06420v1 Announce Type: new Abstract: Visual counting is a fundamental pillar of multimodal intelligence, requiring a seamless integration o…