邀你来挑战｜2023清华技术创新-宝马CHINA IT预见未来联合AI创新挑战营

 #PART 01.   活动背景

Activity Background

2022年，清华技术创新—宝马IT 创新联合挑战项目实验室推出元宇宙创新挑战赛，同学们一起开发元宇宙下的汽车创新业务解决方案，探索了元宇宙与未来人类出行的美好图景。随着ChatGPT在今年初的爆火，AI迎来全民热潮，各行各业都在积极探索AI的应用场景与价值实现。那么AI将如何赋能汽车行业并影响未来汽车产业与人类的出行生活？清华技术创新—宝马China IT 创新联合挑战项目实验室隆重推出预见未来联合AI创新挑战营，旨在针对围绕汽车产品及服务的实际应用场景，鼓励参赛者利用人工智能技术，解决现实问题或创造创新应用。

 #PART 02.   挑战赛道与主题方向

Challenge Course and Theme Direction

挑战赛共有AIGC服务智能、导览智能、车机智能和工厂智能四个挑战赛道，参与者可任选其中一个主题赛道提交相应的创新方案。

{ 主题赛道 }1：AIGC服务智能（金融/ 销售/售后/品牌等部门）

超大规模预训练模型在生成式人工智能领域快速渗透，方便易用的通用化AI能力让品牌、销售、客户服务以及传播的内容、渠道和方式具备更多可能。在当下，用户共创的营销方式、粉丝经济、互动式营销、从质到量飞跃提升的传播物料，这些未来迹象碎片如何整合成让品牌、销售、客户服务等更加智能的未来场景值得我们思考。

同时消费观念与消费方式的变化让汽车金融服务更加活跃，在用户汽车金融服务的全生命周期中，AI通用化语言模型的应用如何能让用户消费和金融服务更加便捷、安全可信赖。在深刻洞察用户需求的前提下，基于人工智能技术的金融领域化场景该如何更好的为用户提供服务是本方向重点。

产出要求：

结合 ①文案和 ②算法生成图片，最终以Demo（生成海报，如下图示例；或视频形式）综合呈现。

①宣传文案（市场调研）：推荐市场上运用较成熟的智能文案生成模型。如：文字输入“宝马iX的宣传文案”，即可获得示例中的高级文案。
②汽车图片（模型开发与训练）：学生需登录BMW IT AI Lab进行模型训练。如：文字输入“宝马iX，红色，海边黄昏”，即可生成如图示例。

{ 主题赛道 }2：导览智能

在AI时代，一个视觉化的数字人助手对于场景导览十分重要，在思考 “如何实现更加智能且富有情感的AI语音对话”前提下，以宝马线下体验空间为实际导览场景，根据实景空间设计符合宝马品牌风格的虚拟讲解助手，可在手机（或其他硬件设备）上辅助参观者进行空间游览、产品介绍或其他业务咨询。

产出要求：

①调研市场可以进行更富有情感的互动实现的算法和相关技术。
②设计具备辅助空间游览、产品介绍或其他业务咨询等功能的讲解数字人。

（图片来自网络）

{ 主题赛道 }3：车机智能

在数字化、智能化和网络化的背景下，车辆作为人们生活空间的延伸，更智能更融合的车机系统是灵魂与核心。在人工智能的支持下，移动出行与办公、娱乐、休息、运动等场景的连接与延伸会更加深入与广泛，在驾驶体验创新、与现有生活场景深度融合、多场景体验串联等方向以现有AI技术能力让车机更智能。以下为车机智能中的三个子方向：

主题方向1. 宝马人工智能乘用体验创新

 'BMW Personal AI Experience' 项目探索人工智能改变车内人机界面 (HMI) 体验的潜力，使其个性化并为个人用户量身定制。利用人工智能，该项目将识别和分析影响客户满意度和体验参与度的关键因素。通过开发动态和适应性强的 HMI 要素，创造一种独特的、定制化体验，与 BMW 的品牌标识相一致，提高客户满意度，并最终培养品牌忠诚度。

项目目的与背景描述：

通过根据个人喜好定制 HMI 体验，最大限度地提高客户满意度和参与度。通过提供独特和个性化的车内体验来增强品牌忠诚度。

HMI 体验是驾驶员满意度的关键因素，但现有系统往往缺乏个性化。当前人工智能提供了新的机会机会，可以根据个人喜好和行为定制 HMI 体验，从而提升整体驾驶体验。

产出要求：

能够集成到 BMW HMI 系统中的一组 AI 驱动的定制功能设计。可响应用户偏好和驾驶习惯的自适应动态 HMI 体验设计。全面评估个性化 HMI 体验对客户满意度和参与度的影响，以及对未来改进的建议。最终产出包含前期调研、设计产出和demo演示效果在内的完整方案呈现。

主题方向2. 车内智能语音系统

针对故障警示、乘员交流表达等实际需求，选取出行场景中语音辅助介入机会点，设计车内语音助手，使用场景可由车内场景延展。

产出要求：

包含前期调研、设计产出和demo演示效果在内的完整方案呈现。

主题方向3. 情绪调节系统

有时人们可能会感到悲伤、愤怒、绝望或其他负面情绪。让我们用车辆来调整这些时刻。车辆能够检测使用的情绪、声音、面部表情、身体姿势、驾驶行为等。如何让车辆可以成为情绪缓解的地方？是否可以借助灯光、音乐、图像和文字使驾驶员和乘客平静下来并振作起来？

产出要求：

针对用户情绪变化，设计构想一个具备情绪调节功能的AI解决方案。需要包含前期调研、设计产出和demo演示效果。

（图片来自网络）

{ 主题赛道 }4：工厂智能

在工厂智能主题中，发布4个命题，以调研和技术优化提供解决方案，以下为命题描述与要求：

① Predictive maintenance based on self-supervised learning
– 基于自监督学习的预防型维护
② LLM expert intended personality for complex problem solving
– 用于复杂问题解决的大模型虚拟专家型人格
③ Result-oriented interaction between human &
machine (refer to GPT automation & service bot) 
– 结果导向的自动人机交互(可参照GPT自动化服务机器人)
④ Wireless signal based objective detection
– 基于无线信号的目标检测

产出要求：

①调研报告（可借鉴”SWOT”方法—优势，劣势，机会及威胁）：
应包含但不局限于:对赛题的理解；当前市场中存在的潜在解决方案，各解决方案的优势及劣势；领先的企业；实际应用中的机遇及挑战；格式需严谨参照“文献综述”；需准备PPT或视频用于最终总结及展示。
②技术原型开发:
应包含但不局限于:模型选择，数据处理，调参训练等；成功指标应与导师共同确认，如模型选择合理性，输出准确率及效能（稳定及安全性不作为评估点在原型化阶段）。

（图片来自网络）

 #PART 03.   日程安排

Schedule

1. 报名与招募

时间：2023年6月-7月中旬

2. 工作坊/初赛

时间：2023年8月20日-8月27日期间

①工作坊：
AI平台培训、项目介绍与主题选择、方案设计。

②初赛：
方案提交及创意展示

3. 方案开发与优化

时间：9-10月

项目辅导：清华+宝马联合辅导

4. 决赛及颁奖

时间：11月中下旬

 #PART 04.   报名要求

Registration Requirements

1. 参与对象：个人参与或团队参与

①个人参与：清华在校生
②团队参与：核心成员为清华在校学生，团队不超过6人

2. 参与者保证对参与的方案拥有充分、完整、排他的知识产权

*请严格按此要求报名，否则会影响奖金发放

 #PART 05.   你的收获

What You Get

1. 开放式学习体验

2. 创新创意的思维提升和项目应用能力

3. 挑战赛证书

4. 入围决赛将有机会获得创新大奖及奖金

5. 清华及宝马对项目的持续辅导和支持

6. 与宝马合作开发你的创意

 #PART 06.   奖项设置

Awards

1. 本次创意征集总奖金为10万元，总计将有不超过六个创意方案入围各奖项。

2. 所有参与人员将有机会获得设备和软件平台的免费支持。

 #PART 07.   组织单位

Organizational Unit

主办单位：

清华大学技术创新研究中心
宝马（中国）汽车贸易有限公司
清华大学艺术与科技创新基地

协办单位：

清华大学学生创客空间协会

 #PART 08.   报名方式

Registration

1. 报名截止：2023年7月20日

2. 报名通道：扫描下方二维码报名↓

如有任何相关问题，请扫码加入本次大赛的沟通咨询微信群，并在群中讨论咨询↓

 #PART 01.   活动背景

Activity Background

In 2a022, Tsinghua Technology Innovation - BMW IT Innovation Joint Challenge Project Lab launched the Metaverse Innovation Challenge, in which students worked together to develop innovative business solutions for automobiles under the metaverse, exploring a beautiful picture of the metaverse and future human mobility. With the explosion of ChatGPT at the beginning of this year, AI has ushered in a national craze, and all industries are actively exploring the application scenarios and value realization of AI. How will AI empower the automotive industry and influence the future of the automotive industry and human travel life? Tsinghua Technology Innovation - BMW China IT Innovation Joint Challenge Project Lab proudly launches the Foreseeing the Future Joint AI Innovation Challenge Camp, which aims to encourage participants to use AI technology to solve real-life problems or create innovative applications for practical application scenarios around automotive products and services.

 #PART 02.   挑战赛道与主题方向

Challenge Course and Theme Direction

There are four thematic tracks of AIGC Service Intelligence, Guided Tour Intelligence, Vehicle Intelligence and Factory Intelligence in the Challenge, and participants can choose any one of them to submit corresponding innovative solutions.

{ Theme Track }1: AIGC Service Intelligence (Finance / Sales / Aftersales / Branding etc.)

Super large-scale pre-trained models are rapidly penetrating in the field of generative AI, and easy-to-use generalized AI capabilities allow for more possibilities in branding, sales, customer service, and the content, channels and methods of communication. In the present, user co-creation of marketing methods, fan economy, interactive marketing, and communication materials that leap from quality to quantity, it is worth thinking about how these pieces of future signs can be integrated into future scenarios that make branding, sales, and customer service smarter.

At the same time, the change of consumer concept and consumption mode makes auto financial service more active. In the whole life cycle of user auto financial service, how can the application of AI generalized language model make user consumption and financial service more convenient, safe and reliable. Under the premise of profound insight into user needs, how to better serve users based on AI technology for financial domainization scenarios is the focus of this direction.

Output Requirements:

Combine ① copywriting and ② algorithm to generate images, and finally present them in a demo (generate a poster, as the following example; or in video form).

① Copywriting (market research): We recommend intelligent copywriting generation models that are more maturely used in the market. For example, if you type "BMW iX promotional copy", you can get the advanced copy in the example.
② Car images (model development and training): Students need to log into BMW IT AI Lab for model training. For example, enter the text "BMW iX, red, dusk by the sea" to generate the example.

{ Theme Track }2: Guided Tour Intelligence

In the era of AI, a visual digital human assistant is very important for scene guidance. Under the premise of thinking about "how to achieve more intelligent and emotional AI voice dialogues", the BMW offline experience space is used as the actual guidance scene, and a virtual tour assistant in line with the BMW brand style is designed according to the real space, which can be used on cell phones (or other hardware devices) to assist visitors in space tour, product introduction or other business consultation. It can assist visitors in space tour, product introduction or other business consultation on mobile phones (or other hardware devices).

Output Requirements:

① Research the market for algorithms and related technologies that allow for more emotive interaction implementation.
② Design of explanatory digital human with functions such as assisting space tour, product introduction or other business consultation.

（图片来自网络）

{ Theme Track } 3: Vehicle and Machine Intelligence

In the context of digitalization, intelligence and networking, the vehicle as an extension of people's living space, a more intelligent and integrated vehicle system is the soul and core. With the support of artificial intelligence, the connection and extension of mobile travel and office, entertainment, rest, sports and other scenes will be more in-depth and extensive, and the existing AI technology capabilities will make the car more intelligent in the direction of driving experience innovation, deep integration with existing life scenes, and multi-scene experience tandem. The following are the three sub-directions of the car machine intelligence:

Theme Direction1. BMW AI Passenger Experience Innovation

The 'BMW Personal AI Experience' project explores the potential of AI to transform the in-car human-machine interface (HMI) experience, personalizing it and tailoring it to the individual user. Using artificial intelligence, the project will identify and analyze key factors that influence customer satisfaction and experience engagement. By developing dynamic and adaptable HMI elements, a unique, customized experience is created that aligns with the BMW brand identity, increases customer satisfaction, and ultimately fosters brand loyalty.

Project Purpose and Background Description:

Maximize customer satisfaction and engagement by tailoring the HMI experience to individual preferences. Enhance brand loyalty by providing a unique and personalized in-vehicle experience.

The HMI experience is a key factor in driver satisfaction, but existing systems often lack personalization. Current artificial intelligence offers new opportunities opportunities to customize the HMI experience based on individual preferences and behaviors, thereby enhancing the overall driving experience.

Output Requirements:

A set of AI-driven custom functions designed to be integrated into BMW HMI systems. Adaptive dynamic HMI experience design that responds to user preferences and driving habits. A comprehensive assessment of the impact of the personalized HMI experience on customer satisfaction and engagement, as well as recommendations for future improvements. The final result is a complete solution that includes preliminary research, design output and demo effects.

Theme Direction2. In-Vehicle Intelligent Voice System

In response to the actual needs of fault warning and occupant communication expression, we will design the in-vehicle voice assistant by selecting the opportunity points of voice assistance intervention in the travel scenario, and the use scenario can be extended from the in-vehicle scenario.

Output Requirements:

A complete solution including preliminary research, design output and demo presentation.

Theme Direction3. Emotion Regulation System

Sometimes people may feel sad, angry, desperate or other negative emotions. Let's use vehicles to adjust to these moments. Vehicles can detect the emotions used, voice, facial expression, body posture, driving behavior, etc. How can vehicles be a place of emotional relief? Is it possible to calm and cheer up drivers and passengers with the help of lights, music, images and words?

Output Requirements:

Design and conceptualize an AI solution with emotion regulation function for user emotion changes. Pre-research, design output and demo effects should be included.

（图片来自网络）

{ Theme Track }4: Factory Intelligence

In the theme of factory intelligence, four propositions will be released to provide solutions with research and technical optimization, the following are the proposition descriptions and requirements:

① Predictive maintenance based on self-supervised learning
② LLM expert intended personality for complex problem solving
③ Result-oriented interaction between human & machine (refer to GPT automation & service bot)
④ Wireless signal based objective detection

Output Requirements:

① Research report (can be based on "SWOT" approach - Strengths, Weaknesses, Opportunities and Threats):
It should include, but not limited to: understanding of the competition topic; potential solutions in the current market, advantages and disadvantages of each solution; leading companies; opportunities and challenges in practical applications; the format should be critically referred to the "literature review"; PPT or video should be prepared for the final summary and presentation.

② Technology prototype development:
The success indicators should be confirmed with the mentor, such as the reasonableness of model selection, output accuracy and performance (stability and security will not be evaluated in the prototyping stage).

（图片来自网络）

 #PART 03.   日程安排

Schedule

1. Registration and Recruitment

Time: June - mid-July 2023

2. Workshop/Preliminary Competition

Time: August 20, 2023 - August 27, 2023

① Workshops:
AI platform training, project introduction and theme selection, program design.

② Preliminary round:
Proposal submission and idea presentation

3.  Program Development and Optimization

Time: September-October

Project coaching: Tsinghua+BMW joint coaching

4. Final and Awarding

Time: mid-late November

 #PART 04.   报名要求

Registration Requirements

1. Participants: Individuals or teams

① Individuals: Tsinghua students
② Team participation: Core members are Tsinghua students, no more than 6 people in the team

2. Participants guarantee that they have full, complete and exclusive intellectual property rights to the project.

*Please strictly follow the requirements of the registration, otherwise it will affect the competition prize

 #PART 05.   你的收获

What You Get

1. open learning experience

2. Innovative and creative thinking and project application skills

3. Challenge certificate

4. Finalists will have the opportunity to win the Innovation Award and prize money

5. Continuous coaching and support from Tsinghua and BMW

6. Cooperation with BMW to develop your ideas

 #PART 06.   奖项设置

Awards

1. The total prize money for this call for ideas is 100,000 RMB, and no more than six creative solutions will be shortlisted for each award.

2. All participants will have the opportunity to receive free support for equipment and software platform.

 #PART 07.   组织单位

Organizational Unit

Hosted by

Tsinghua University Technology Innovation Research Center
BMW (China) Automotive Trading Co.
Tsinghua University Art and Technology Innovation Base

Co-Organized by
Tsinghua University Student Creativity Space Association

 #PART 08.   报名方式

Registration

1. Application deadline: July 20, 2023

2. Registration channel: Scan the QR code below to register ↓

If you have any related questions, please scan the code to join the communication and consultation WeChat group of the competition and discuss the consultation in the group ↓