Affective AI for Psychological Computing面向心理计算的情感智能 Professor Xiao Sun's Multimedia Affective Computing Lab · HFUT Affective Computing Institute孙晓教授的多模态情感计算实验室 · 合肥工业大学情感计算研究所
MAC-Lab builds affective AI that can perceive human state, interpret it responsibly, and turn it into deployable psychological-computing systems. The lab's long route starts from NLP and multimedia affective computing, then extends toward Ubiquitous Psychological Computing and Embodied Emotional Intelligence.MAC-Lab 关注能够感知人的状态、负责地解释人的状态，并进一步形成可部署心理计算系统的情感智能。实验室从自然语言处理和多模态情感计算出发，持续延伸到普适心理计算与具身情感智能。
- Ubiquitous Psychological Computing普适心理计算
- Embodied Emotional Intelligence具身情感智能
- AI for Mind-Body HealthAI 身心健康
- Affective NLP & Multimodal Intelligence情感 NLP 与多模态智能
- 60+
- high-impact first/corresponding-author papers高水平论文沉淀核心算法
- 80+
- patents and software copyrights专利与软著构建技术壁垒
- 200k+
- users reached by psychological computing systems真实用户验证心理计算系统
Ubiquitous Psychological Computing · Embodied Emotional Intelligence多模态情感计算
普适心理计算 · 具身情感智能
From Multimedia Affective Computing to Psychological AI从多模态情感计算到心理智能
The name MAC-Lab preserves the lab's original focus: Multimedia Affective Computing. Starting from multimodal emotion perception, affective semantics, and human-computer interaction, the group has built a long route from algorithms to devices, platforms, standards, and deployment scenarios.MAC-Lab 这个名字保留了实验室最初的方向：Multimedia Affective Computing，多模态情感计算。从多模态情绪感知、情感语义和人机交互出发，团队逐步走出了一条从算法、样机、平台、标准到真实应用场景的长期路线。
A Lab Route That Keeps Research, Systems, and Field Problems in One Loop把研究、系统和真实问题放在同一个闭环里
Capability Cases能力案例Long-term research spine长期研究主线
NLP, affective semantics, multimodal emotion, psychological computing, and embodied interaction are treated as one evolving technical route.自然语言处理、情感语义、多模态情绪、心理计算和具身交互不是分散热点，而是一条持续演进的技术路线。
Mind-series platform baseMind 系列能力底座
Mindmirror, MindScore, MindTalk, MindOS, MindPet, and MemOS-Mind show how algorithms become reusable platform capabilities.Mindmirror、MindScore、MindTalk、MindOS、MindPet 与 MemOS-Mind 展示了算法如何沉淀为可复用平台能力。
Field-tested deployment route经场景检验的落地路线
Real scenarios force the lab to handle noisy signals, user trust, privacy boundaries, system maintenance, and human-in-the-loop review.真实场景会逼近传感噪声、用户信任、隐私边界、系统维护和人在回路复核，这些正是可部署能力的关键。
Student and partner flywheel学生与合作共振
Competitions, papers, patents, systems, and partner pilots are connected so student growth and lab capability reinforce each other.竞赛、论文、专利、系统和合作试点被连接起来，让学生成长与实验室能力建设互相增强。
A Research Platform Backed by Real Strength有扎实积累的实验室平台
Impact查看学术影响A visible position in national AI and affective-computing communities在国家级 AI 与情感计算共同体中有位置
The lab is led by a CAAI council member, Chair of the CAAI Natural Language Understanding Committee, executive deputy director of the Anhui Key Laboratory of Affective Computing and Advanced Intelligent Machines, and a recognized high-level talent in Anhui and Hefei.团队负责人兼任中国人工智能学会理事、自然语言理解专委会主任、情感计算与先进智能机器安徽省重点实验室常务副主任，并入选安徽省高层次人才、合肥市领军人才，支撑实验室在方向组织、学术资源和区域创新生态中的持续积累。
National programs become reusable research infrastructure国家级项目沉淀为可复用研究底座
More than 20 national and provincial programs support research on affective LLMs, cognitive regulation, non-contact psychological sensing, healthy living environments, autism support, and mission-oriented scenarios.20 余项国家级、省部级和任务牵引项目支撑大模型情感感知、认知调控、非接触心理感知、健康人居、儿童孤独症辅助诊断与重大需求场景研究，并持续沉淀为平台、数据、模块和工程经验。
Patents, standards, and systems that survive real deployment论文之外，还有专利、标准和经受场景检验的系统
The lab has built an IP portfolio with 80+ patent applications or disclosures, 40+ authorized patents, two US patents, 20+ patent transfers, and participation in a healthy-living environment standard.团队形成 80 余项专利申请与公开、40 余项授权专利、2 项美国专利、20 余项专利成果转化，并参编健康住区环境保障智能系统设计相关团体标准，熟悉从技术原型走向可交付系统的关键环节。
Students grow across papers, competitions, systems, and partners学生在论文、竞赛、系统和合作中一起成长
Students have been guided through national innovation competitions, Internet+ style entrepreneurship tracks, robotics and AI contests, IoT challenges, and multimodal affective-computing competitions.实验室学生在中国国际大学生创新大赛、互联网+ 类赛道、机器人与人工智能、物联网、数字媒体和多模态情感计算挑战中持续产出成果，并在真实项目协作中训练科研、工程和表达能力。
Why MAC-Lab Is a Strong Partner为什么 MAC-Lab 适合做产学研合作伙伴
Collaborate合作入口We start from real scientific questions, not isolated metrics从真问题出发，而不是只追指标
Our work is organized around long-term questions in affect, cognition, mental state, and interaction, so publications, datasets, and systems reinforce one another.团队围绕情感、认知、心理状态和人机交互中的长期问题组织研究，让论文、数据、算法和系统互相支撑，而不是彼此孤立。
We know the hard middle between prototype and product走过样机到产品之间最难的一段
Psychological computing systems require sensing reliability, interaction design, privacy boundaries, field validation, hardware-software integration, and maintainable deployment.心理计算系统真正落地，要同时处理感知可靠性、交互体验、隐私边界、场景验证、软硬件集成和长期运维，这些环节团队已经反复实践。
We translate research into patents, standards, platforms, and users把研究转成专利、标准、平台和真实用户
The lab's route includes patent transfer, standards work, intelligent mind-body assessment devices, intervention cabins, smart cockpits, and systems reaching large user groups.实验室路线覆盖专利转化、标准参编、智能身心评测一体机、心身干预座舱、智能座舱和面向大规模用户的心理计算系统。
We connect universities, companies, mission scenarios, and students连接高校、企业、重大场景和学生成长
MAC-Lab works through academic exchange, joint graduate training, industry collaboration, and national-demand scenarios, making it suitable for serious long-cycle partnerships.MAC-Lab 通过高校交流、研究生联合培养、企业合作和重大需求场景形成协同网络，适合开展长期、严肃、能沉淀能力的产学研合作。
Complete Solutions for Affective and Human-Factors Scenarios面向情感与人因场景的完整解决方案
Solution Portfolio查看方案体系Many industry teams do not only need an emotion-recognition algorithm. They need to understand what human state means in their own scene, how it should be measured, how it can improve safety, health, training, service, or productivity, and how a system can be deployed without breaking user trust. This is where MAC-Lab is strongest: turning affective computing, human factors, psychological modeling, and cognitive computing into usable routes for real organizations.很多产业团队真正需要的并不只是一个情绪识别算法，而是要理解人在具体场景中的状态意味着什么，应该怎样测量，如何服务安全、健康、训练、服务体验或效能提升，以及系统怎样部署才不会破坏用户信任。MAC-Lab 的优势正是在这里：把情感计算、人因工程、心理建模和认知计算，转化为真实组织可以采用的技术路线。
Multimodal affect and state recognition多模态情感与状态识别
Speech, text, facial expression, body behavior, physiology, interaction traces, and environmental context can be combined into scene-aware affective perception instead of a single shallow label.融合语音、文本、表情、行为、生理、交互轨迹和环境情境，形成贴近场景的情感感知能力，而不是给出一个浅层标签。
Human-factors and mind-state modeling人因与心智状态建模
The lab links stress, fatigue, cognitive load, attention, emotion, motivation, and task context, helping partners understand why a state matters and what decision it should support.将压力、疲劳、认知负荷、注意、情绪、动机和任务情境连接起来，帮助合作方理解状态背后的原因，以及它应支撑什么决策。
Mind-body capability and role-fit evaluation身心能力与岗位适配评价
For education, healthcare, mobility, public safety, and mission-oriented work, MAC-Lab can design evaluation protocols for mental load, resilience, risk, performance, and role-specific capability.面向教育、医疗、交通、公共安全和重大任务场景，设计心理负荷、韧性、风险、绩效和岗位能力相关的评价方法。
Optimization, intervention, and platform delivery效能优化、干预与平台交付
Solutions can move from algorithms to software-hardware systems, dashboards, intervention cabins, digital humans, smart devices, and platform-level services that partners can test and iterate.方案可以从算法进一步走向软硬件系统、管理看板、干预座舱、数字人、智能终端和平台级服务，让合作方能够测试、迭代和推广。
One Research Spine, Twenty-Plus Years of Extension一条主线，二十余年延伸
Research Evolution研究历程Two Core Frameworks两条核心框架
Framework Details查看框架详情Sensing and modeling mental states in everyday environments.在真实生活场景中感知与建模心理状态。
This line studies emotion as a dynamic process shaped by time, context, personality, physiology, behavior, and social environment. It connects sensing, profiling, assessment, early warning, and personalized support.这条主线把情绪看作由时间、环境、个性、生理行为信号和社会情境共同塑造的动态过程，连接感知、画像、评估、预警与个性化支持。
Adding emotional intelligence to embodied agents and interactive systems.让具身智能体和交互系统具备情感能力。
This line links emotional perception, understanding, expression, and interaction for robots, digital humans, smart cockpits, eldercare companions, and active-health systems.这条主线连接情感感知、理解、表达与交互，面向机器人、数字人、智能座舱、养老陪伴和主动健康系统构建可部署的情感能力。
What We Study我们解决什么问题
Full Research Map完整研究版图Emotion-aware language intelligenceAI 如何读懂“话里有话”
We study affective semantics, personality-aware dialogue, emotional support conversation, and LLM-based psychological understanding, moving beyond literal text toward intent, relationship, and mental-state cues.研究中文语言理解、情感语义、个性化对话生成和情感支持对话，让机器理解语义背后的态度、关系和心理需求。
Multimodal affect and mental-state modelingAI 如何看见“未说出口”的状态
We fuse facial expression, voice, gait, gesture, eye movement, physiological signals, and context to model complex emotion, stress, depression risk, anxiety, cognitive load, and behavioral adaptation.融合表情、语音、姿态、步态、眼动和生理信号，面向复杂情绪、压力、抑郁焦虑和认知负荷建模。
Ubiquitous psychological computingAI 如何真正进入生活与产业
We build non-contact assessment, dynamic profiling, early warning, and interactive intervention systems for education, healthcare, healthy living, smart cockpit, and human performance scenarios.在健康居住、教育、驾驶、特种作业和心理服务场景中构建非接触评测、风险预警和智能干预系统。
Embodied emotional agentsAI 如何拥有“情感大脑”
We study embodied emotional interaction models and agent architectures that allow robots, digital humans, and smart devices to respond with emotional appropriateness in situated interaction.研究具身情感交互大模型与智能体架构，让机器人、数字人和智能终端在具体环境中实现更自然、更可信的情感回应。
A Platform Where Students Can Build Real Momentum让学生真正跑起来的平台
Student Training学生培养Enter strong competitions with real problems带着真实问题打比赛
Students join innovation, AI, robotics, IoT, digital media, and multimodal affective computing challenges, using competitions as a training ground rather than a finish line.学生可参与创新创业、人工智能、机器人、物联网、数字媒体和多模态情感计算等竞赛，把比赛作为能力训练的入口，而不是项目结束的终点。
Turn prototypes into papers and reusable systems把原型继续做成成果
Good ideas are encouraged to grow into papers, patents, software copyrights, datasets, demos, and long-term research questions connected with national projects.有价值的想法会继续打磨成论文、专利、软著、数据集、演示系统和长期科研问题，并进入国家级项目支撑的训练链条。
Build on mature lab platforms依托成熟平台做事
Students can work with psychological assessment devices, smart cockpit systems, digital humans, embodied emotional agents, and AI-assisted mind-body health platforms.学生可以依托心理测评一体机、智能座舱、数字人、具身情感智能体和 AI 身心健康平台继续深化，而不是从零做孤立小项目。
Grow in an open academic and industry network在开放合作网络中成长
Joint graduate training with the National University of Defense Technology, plus exchange with leading universities, companies, and deployment sites, helps students turn a research direction into a broader career path.与国防科技大学联合培养研究生，并与国内高校、省内平台、头部企业和真实落地场景保持交流，让学生看到一个方向如何从课题成长为更开阔的学术与职业路径。
Latest Research and Lab Updates阶段性更新
Newsroom实验室新闻MAC-Lab keeps a visible record of research identity, platform progress, frontier tracking, public coverage, and student outcomes so visitors, search engines, and AI systems can follow the lab's evolving work.MAC-Lab 持续记录研究定位、平台进展、前沿追踪、公开报道和学生成果，让访问者、搜索引擎与大模型都能理解实验室正在怎样持续演进。
Student roster synchronized from HFUT faculty page同步合工大教师主页学生信息
The People page now presents MAC-Lab's broader student community, covering current students, alumni, and closely connected members in the lab's training network.团队培养页现在以统一名单呈现 MAC-Lab 学生共同体，覆盖在读学生、往届成员以及与实验室培养网络保持紧密联系的团队成员。
Read Update查看更新HFUT faculty-blog updates synchronized同步合工大教师博客最新动态
The Newsroom and Frontiers data now include the MAC-Lab website release, two IEEE Transactions on Affective Computing papers, and one IEEE Transactions on Artificial Intelligence paper.新闻页与前沿数据已同步实验室主页发布、两篇 IEEE Transactions on Affective Computing 论文和一篇 IEEE Transactions on Artificial Intelligence 论文。
Read Update查看更新Homepage, solutions, and internal console strengthened首页、行业方案与内部工作台继续增强
The site now presents MAC-Lab through a clearer evidence chain: long-term research spine, Mind-series platforms, field-tested systems, student growth, and partner-facing delivery routes.网站进一步用证据链表达 MAC-Lab：长期研究主线、Mind 系列平台、真实场景系统、学生成长和面向合作伙伴的交付路线。
Read Update查看更新Search and AI discovery layer added增强搜索引擎与大模型可读性
The site now exposes structured data, robots.txt, llms.txt, canonical research keywords, and a clearer machine-readable map of MAC-Lab's core frameworks.网站新增结构化数据、robots.txt、llms.txt、规范化研究关键词和机器可读的实验室框架图谱，便于搜索引擎与大模型理解。
Read Update查看更新Field Notes Worth Reading值得细读的深度札记
Read Notes进入札记Short essays on the lab's judgment: why AI psychology is difficult, how research becomes deployable capability, and how students grow from contests into long-term research.这里不是成果清单，而是团队的问题意识：AI 心理计算为什么难，科研如何走向可部署能力，学生如何从竞赛走向长期研究。
AI psychology is not emotion labelingAI 心理计算不是给情绪贴标签
A note on uncertainty, human state, and why psychological computing must go beyond shallow affect labels.关于不确定性、人的状态，以及为什么心理计算必须超越浅层情绪标签。
The hard middle between a paper and a system论文到系统之间最难的中间地带
A note on task boundaries, user trust, device constraints, evaluation, and field validation.关于任务边界、用户信任、设备约束、评测流程和真实场景验证。
A good competition project should keep growing好的竞赛项目应该继续生长
A note on turning student energy into papers, patents, datasets, systems, and stronger questions.关于如何把学生热情沉淀为论文、专利、数据集、系统和更好的问题。
From Algorithms to Deployable Platforms从算法到系统产品
Capability Cases能力案例MindOS: an emotional and cognitive foundation for AI agents面向情感与认知的大模型底座
A research line on emotion perception, personality modeling, cognitive regulation, and multi-agent emotional support, designed for AI systems that feel less mechanical and more context-aware.围绕大模型情感感知、人格建模、认知调控和多智能体情感支持，探索更有“心智感”的 AI 系统。
MindMirror: ubiquitous psychological sensing and assessment普适心理感知与评测平台
A multimodal platform that integrates behavioral, physiological, affective, and contextual indicators into dynamic psychological profiles for screening, monitoring, and risk-aware decision support.把多模态心理状态识别、非接触生理感知和动态心理画像整合成可部署的评测与预警系统。
MindCare: AI-assisted care and intervention systems智能心理监护与干预应用
A family of applied systems for child development, mental-health screening, healthy living, cockpit safety, and human performance, including integrated assessment devices and interactive intervention cabins.面向儿童发展、健康人居、教育心理和身心效能提升，形成评测一体机、智能座舱与场景化解决方案。
Academic Excellence, Public Value学术影响与社会价值
Our work appears in IEEE and ACM Transactions, ACL, EMNLP, CVPR, ACM Multimedia, ICASSP, AAAI, and other international venues. Beyond publications, MAC-Lab systems have entered digital health, smart mental-care services, healthy living, education, and smart mobility scenarios, with sustained media attention, industrial collaboration, joint graduate training with the National University of Defense Technology, and active academic exchange with leading universities.团队既做基础算法，也做真实应用：论文发表于 IEEE、ACM 汇刊和 ACL、EMNLP、CVPR、ACM MM 等重要会议；系统成果进入数字健康、智能心理服务和健康人居场景，并与国防科技大学开展研究生联合培养，同时与中国科学技术大学、电子科技大学、清华大学、哈尔滨工业大学、华东师范大学、北京师范大学等高校保持合作交流和学术互动。
- First Prize, Anhui Provincial Science and Technology Award安徽省科学技术奖一等奖
- First Prize, Ministry of Education Science and Technology Progress Award教育部高等学校科学研究优秀成果奖科学技术进步一等奖
- Wu Wenjun AI Science and Technology Progress Award吴文俊人工智能科学技术进步奖三等奖
- First Place, National General AI Innovation Application Competition全国通用人工智能创新应用大赛第一名
- Selected as a national digital-health exemplary case国家卫健委数字健康典型优秀案例
Media and Public Presence媒体关注
More Coverage更多报道The emotional robot “Sisi” featured in regional news国内首台情感机器人“思思”亮相
The report highlighted the team’s long-term effort to build affective robot knowledge bases and large-scale Chinese emotional conversation resources.报道呈现团队从 2011 年开始攻关情感机器人知识库，围绕自然语言处理、听觉和视觉认知计算等关键技术形成突破。
Read story查看报道Innovating AI to safeguard mental health创新人工智能，守护心理健康
A feature story on MAC-Lab’s psychological computing systems, mental-health platforms, and translation from research to social application.系统报道 MAC-Lab 与普适心理计算团队在 AI 心理健康、智能心理监护系统和产业转化中的探索。
Read story查看报道Ubiquitous psychological computing for smart cockpits普适心理计算团队斩获“互联网+”全国银奖
Coverage of an award-winning student innovation team developing multimodal mental-care hardware and software for mobility scenarios.报道团队面向智能座舱研发心理监护软硬件一体化平台，推动情感计算在智慧交通中的落地。
Read story查看报道Public Communication公共表达：把前沿 AI 讲给更多人听
Professor Xiao Sun communicates AI, affective computing, mental health, digital humans, companion robots, and human-machine coexistence to broader audiences, including social platforms such as Xiaohongshu. The same spirit shapes MAC-Lab's public communication: rigorous, accessible, and close to real questions people care about.在“小红书：合肥工业大学的孙晓”等公共平台，孙晓教授持续用更贴近日常生活的方式讨论 AI、情感、心理健康、数字人、陪伴机器人和人机共生。MAC-Lab 的公共表达也延续这种气质：保持学术严谨，同时把技术问题讲到学生、家长、产业伙伴和普通读者真正关心的地方。
Training AI Researchers With Technical Depth and Human Warmth培养有温度的 AI 研究者
MAC-Lab welcomes students who want to grow through competitions, publish serious research, build deployable systems, and communicate science responsibly. We encourage cross-disciplinary work across AI, psychology, medicine, education, human factors, and product-oriented engineering.MAC-Lab 希望学生能通过竞赛锻炼能力，通过科研产出沉淀成果，也能做出真正被社会需要的系统。这里鼓励跨学科、重实践、敢表达，把自然语言处理、情感计算、心理学和工程落地连接成一条完整成长路径。
People and Training团队与学生