{"schema":"BUIDL","d":{"id":13818,"members":[],"pictures":["https://cdn.dorahacks.io/static/files/1903e8431bf74efa53934a54f77ad8e7.png"],"roundProjects":[],"hackathonProjects":[{"id":14747,"hackathon":{"id":405,"tabSet":[{"id":729,"name":"Submission Guide","slug":"submission","description":"# Getting Started [3-5 MIN]\n\n**1.** **You should submit your BUIDL via [DoraHacks.io](https://dorahacks.io/). Before that, make sure that you already have an account.**\n\nUse the **Sign in/up button** at the upper right corner of the page to access the sign-up options.\n\n

\n\n**Github, crypto wallet** and **email** are supported for creating an account. For BUIDLers, we recommend you to connect the account to your **github** to make more people know what you are building at any time.\n\n*You may be asked to receive a verification code via email. The email may take some time to arrive. No worries! The code will not expire in 10 minutes when you click “send code”.\n\n**2. Before the submission of deadline, please make sure you have added your Github link to the project. You can submit the project first, edit it later before the deadline.**\n\nClick “ Register As Hacker”-“Submit BUIDL”.\n\n

\n\n> **→ If you already have a BUIDL on DoraHacks**\n\nChoose your BUIDL project from **\"Apply with Existing BUIDL\"** and identify the track you want to join.\n\n> **→ If you want to submit a new BUIDL project**\n\nClick **“Create a new BUIDL”** and fill in the form as required. You'll be directed to \"Apply for Trustworthy AI Lab x GES UCLA Hackathon\" page after the new BUIDL setup.\n\nFill in the form to offer the information about your project. Note that for demo, demo link and social media, please offer **URLs (with http/https://)**.\n\n

\n\nDon't forget to check \"I agree to the User Agreement\".\n\n

\n\nBack to Application window. Choose your BUIDL project from **\"Apply with Existing BUIDL\"** and identify the track you want to join.\n\nWhen submission succeeds, you'll see this notification.\n\n

\n\n# Project Verification\n\nAfter verification by the hackathon organizer through the manager portal, your project will be displayed in the [BUIDLs gallery](https://dorahacks.io/hackathon/405/buidl).\n\n# Want to make some changes to your BUIDL?\n\nClick [Avatar] -> [Profile] to enter the Account Center, and click the \"Edit\" button on the top right of the BUIDL to update the BUIDL information.\n\n

\n\n
\n\n

"},{"id":734,"name":"Prompt","slug":"prompt","description":"# Overview\nThe main theme of this event is to utilize Generative AI to empower “Data Collaboration Intelligence.” In this hackathon, participating teams will develop a data sharing platform that allows private data sharing, predictive analytics and model building among different data parties. To be concrete, each team will develop a “Data Clean Room” to enhance Click-through Rate (CTR) predictions using privacy-preserving synthetic data. This prompt page only provides a high-level overview of the prompt, please refer to the [slides](https://drive.google.com/file/d/18US5iNK191t2-u0pA-2sbmGCfjhBOV_i/view) for more details.\n\n# Background\nThe Data Clean Room is a secure framework that facilitates the sharing of data between two distinct parties. In the above CTR case, we have two parties:\nPublishers (providing `train_data_feeds.csv`): They focus on user interactions with news content.\nAdvertisers (providing `train_data_ads.csv`): They concentrate on user interactions with advertisements.\n\nThis arrangement ensures that both parties can access broader datasets while maintaining user privacy and data security.\n\n# Challenges\n\nSynthetic Data Fidelity: \nCreate synthetic datasets that accurately reflect the real-world distributions of user behavior data.\n\nSynthetic Data Utility:\n- Publishers: Enhance predictive accuracy for user engagement with news content by integrating user interaction patterns from the advertisement dataset. This aims at better content personalization and placement strategies.\n- Advertisers: Refine prediction models for advertisements by using synthesized data from news content interactions to understand the most effective contexts for ad placements.\n\n# Dataset Overview\n1. `train_data_ads.csv`: Includes user demographics, device information, advertisement specifics, and user interaction data.\n2. `train_data_feeds.csv`: Contains detailed device info, content engagement metrics, and user interaction labels.\n\n# Data Clean Room Mechanics\nConstruct a Data Clean Room (DCR) to train the CTR model within the DCR. Ensure datasets remain encrypted outside of the DCR to guarantee that only the DCR can decrypt the data. Multiple parties can collaborate on data synthesis by merging two datasets based on IDs. Only the data owner will have access to their specific data outside DCR.\n\n# Potential Technical Challenges\n1. Synthetic Data Generation Accuracy: Creating synthetic data that accurately represents the complex distributions of real-world data without inheriting biases.\n2. Trust Execution Environment (TEE): build or utilize a DCR and train the model within the TEE.\n3. Integration of Heterogeneous Data Sources: Merging data from different datasets (publisher and advertiser data) with varying structures and formats to ensure seamless data integration within the Data Clean Room.\n4. Scalability and Real-Time Processing: Developing scalable solutions capable of handling large datasets and processing data in real-time to provide timely and efficient insights.\n\n# Expected Outcomes\n1. Development of more accurate predictive models for CTR that consider a richer set of anonymized data.\n2. Demonstration of effective collaboration within a privacy-preserving environment.\n3. Enhanced understanding of cross-party data utility without direct data exposure.\n\n# Benefits to Participants\n1. Experience with real-world datasets and privacy-preserving technologies driven by generative AI.\n2. Insights into advanced data analytics and synthetic data applications in digital marketing and content distribution.\n\n# Assessment Criteria\n- Accuracy and Predictive Performance:\nProjects will be evaluated based on the accuracy and reliability of their predictive models. Higher precision in forecasting user engagement with news content and advertisements will be favored.\n- Innovation and Creativity:\nJudges will assess the novelty and creativity of the solutions, especially in terms of how participants handle data integration, synthesis, and privacy-preserving techniques within the Data Clean Room.\n- Scalability and Efficiency:\nThe ability of the solution to scale efficiently with increasing data volumes and to process data in real-time will be crucial. Projects need to demonstrate performance stability and resource management.\n\n# Submission Requirements\n1. Runnable Data Clean Room code in a Github Repo\n2. CTR Prediction machine learning code in a Python package/script\n3. 3-5 page presentation slides, not including index and references\n\n# Conclusion\nThis hackathon aims to bridge the gap between data privacy and analytical utility, challenging participants to create innovative solutions that enhance predictive accuracy in marketing and content engagement. The event is set to provide hands-on experience with cutting-edge technologies, mirroring current and future industry challenges in a scenario that enhances both privacy and performance."},{"id":804,"name":"Round 1 Result (Top 5)","slug":"round-1-result-(top-5)","description":"# Announcing the Winners of the Trustworthy AI Lab x GES UCLA Hackathon\n\n## GES X Trustworthy AI Lab Hackathon Background\n\nPrize Pool Allocation: $1500\n\nThis year's hackathon is focused on creating Generative AI solutions for Data Collaboration. Participants were tasked with developing AI tools that enhance data privacy and protection, transforming how data is shared within digital systems.\n\nNow, let's look at the outstanding teams and their contributions!\n\n# Prize Winners\n\n### 🥇 First Place\n\n[BUIDL](https://dorahacks.io/buidl/13818)\n\n\n### 🥈 Second Place\n\n[BUIDL](https://dorahacks.io/buidl/13812)\n\n\n### 🥉 Third Place\n\n[BUIDL](https://dorahacks.io/buidl/13817)\n\n\n### Fourth Place\n\n[BUIDL](https://dorahacks.io/buidl/13816)\n\n\n### Fifth Place\n\n[BUIDL](https://dorahacks.io/buidl/13821)"}],"skills":null,"applied":false,"isCurrentHackerRegistered":false,"buidlsCount":9,"applicationsCount":10,"organization":{"id":2212,"name":"GES at UCLA","logo":"https://cdn.dorahacks.io/static/files/18f5bba2f1c25c7060d0d9c4279a1aa8.jpeg"},"description":"# About Global Entrepreneurship Society at UCLA \n\nThe Global Entrepreneurship Society at UCLA (formerly Three Day StartUp at UCLA) is an entrepreneurial and educational extracurricular program aimed at university students, emphasizing hands-on learning. \n\nOur mission is to foster entrepreneurial mindsets and skills within our student community to tackle problems, launch new ventures, and cultivate successful businesses with global impacts. \n\nOur approach encompasses every stage of the entrepreneurial journey, from ideation and team formation to market research, rapid prototyping, and final pitch presentations. Through workshops and practical experiences, members directly apply their acquired knowledge to develop sustainable business models, with a focus on expanding startup ideas into thriving global enterprises.\n\nWebsite: https://gesatucla.org/\n\n\n# About Trustworthy AI Lab @ UCLA\n\nTrustworthy AI lab at UCLA envisions AI 2.0 as being driven by trustworthiness (beyond performance), and being built upon **generative data** (on top of raw data). Our lab drives advancements in Generative Data for marketing, healthcare, and finance sectors. We develop data-centric tools, such as artificially generated tables and conversations, which enables privacy-preserving data sharing and reliable scenario exploration.\n\nhttp://www.stat.ucla.edu/~guangcheng/index.html\n\n# GES X Trustworthy AI Lab Hackathon Background\n\n**Prize Pool Allocation: $1500**\t\t\t\n\t\t\t\t\t\t\t\t\nThe 2024 Hackathon is dedicated to pioneering Generative AI for Data Collaboration – an innovative approach to redefine data privacy and protection. Let's design and implement robust Generative AI solutions set to revolutionize the way data is shared within digital ecosystems.\n\t\t\t\t\t\t\n# Who Should Participate\n\n- Undergraduate & Graduate Students in the US\t\t\t\t\t \t\t\t\t\t\t\t\t\n- Interested in Data Science and Machine learning\t\t\t\t\t\t \t\t\t\t\t\t\t\t\n- Eager to meet like-minded peers\t\t\t\t\t\t \t\t\t\t\t\t\t\t\n- Ready to harness creativity and empower organizations\t\t\t\t\t \t\t\t\t\t\t\t\t\n- (No restrictions to academic majors)\n\n\t\t\t\t\t\t\n# Objectives\n\n**Prompt**: https://docs.google.com/document/d/1-SpZ2NWVQix0dEYfMaBKvXFI5MRIp2-nDeOiCzXfrtw/edit?usp=sharing\n\nLeverage predictive modeling to enhance forecasts of user engagement and interactions with content. The challenge lies in refining the accuracy and efficiency of these predictions, crucial for content publishers and advertisers focused on optimizing user experiences and advertising effectiveness.\n\nParticipants will utilize **2 comprehensive datasets** that detail *user demographics, user interactions, engagement metrics, and advertisement specifics* from both the content publishers' and advertisers' perspectives. By employing innovative data analytics and machine learning techniques, participants will tackle this dual challenge and cultivate an engaging user experience that drives the digital publishing and advertising industries.\n\n\t\t\t\t\t\t\n# Challenge\n\nGES X Trustworthy AI Lab Hackathon tends to empower both content creators and advertisers by fostering collaboration and innovation in developing platforms that serve the needs of both parties. The challenge lies in creating a seamless integration between content creation and advertising, ensuring that both sides can thrive while maintaining a positive user experience.\n\t\t\t\t\t\t\n# Event Schedule\n\nDate: May 29th - June 28th, 2024, PST\n\nLocation: Online\n\n**May 29th 8 p.m. PST**\n\t\t\t\t\t\t\t\t\t\t\t\t \t\t\t\t\t\t\t\t\n- Prompt Release and Kickoff (on Zoom)\n\n**June 16th**\n\t\t\t\t \t\t\t\t\t\t\t\t\n- Registration Deadline\n\n**June 21st**\n\n - First Round Submission\n\t\t\t\t\t\t\n**June 27th**\n\n- Second Round Submission\n\n**June 28th 4-6 p.m. PST**\n\n- Judging Day (on Zoom)\n\n# Judging Panel \n\nProfessor Guang Cheng, director of Trustworthy AI Lab, UCLA\n\nDr. Chi-hua Wang, Postdoc at the Trustworthy AI Lab, UCLA\n\nMr. Minrui Gui, PhD student at the Trustworthy AI Lab, UCLA\n\nMr. Shaoqing Yuan, Senior Applied Scientist, Amazon\n\nMr. Harry Xu, Machine Learning Engineer, Snap \n\nMr. Ken Lu, Chief Cloud System Architect, Intel\n\n# Prize Pool Details \n\n- Prize Pool: $300(third), $500(second), $700(first)\n- **Total Prize Pool Allocation: $1500**\t\t\n- Additional Benefit: offer **summer internship** working in the Trustworthy AI Lab @ UCLA. Past summer interns have gone on to attend top graduate schools, such as Berkeley and Princeton. \n\n \t\t\t\t\t\t\t\t\n# Submission Requirements\n\n- Runnable Data Clean Room code in a Github Repo \n- CTR Prediction machine learning code in a Python package/script\n - 3-5 page presentation slides, not including index and references\n\n# Judging Criteria\n\n- Accuracy and Predictive Performance:\nProjects will be evaluated based on the accuracy and reliability of their predictive models. \n\n- Innovation and Creativity:\nJudges will assess the novelty and creativity of the solutions, especially in terms of how participants handle data integration, synthesis, and privacy-preserving techniques. \n\n- Scalability and Efficiency:\nThe ability of the solution to scale efficiently and process data in real-time will be evaluated. Projects need to demonstrate performance stability and resource management.\n\n# Resource Package\nhttps://drive.google.com/drive/folders/1lVOT0evxCsxPpDsTI37tPWeq0TgQ0x5w?usp=drive_link\n\n# Kickoff Event\nhttps://drive.google.com/file/d/1QHxI3VElJEVvNjrQjxpuMLz6XDSAWIZP/view?usp=sharing\n\n# Connet with us\n\n- Website: https://gesatucla.org/\n- Instagram: https://www.instagram.com/ges_ucla/\n- Email: GESatUCLA@gmail.com","onboardingCompleted":false,"registerFormUrl":"https://docs.google.com/forms/d/e/1FAIpQLSdFKR-rYVo_m5llfLYBqXQRuqfQlHU-ncV1QVorDKmGNpMglw/viewform","preRegisterTime":0,"publicContactHackerHandle":null,"hackersCount":21,"title":"Trustworthy AI Lab x GES UCLA Hackathon","uname":"trustworthy-ai-gesucla","imageUrl":"https://cdn.dorahacks.io/static/files/18fa17c53b11244f0c7303f46779e14c.jpg","originUrl":null,"startTime":1717019400,"endTime":1720422000,"timezone":"420","city":null,"visible":true,"approvalStatus":6,"state":1,"voteUrl":null,"voteMechanism":null,"bonusPrice":1500,"bonusMechanism":null,"followersCount":0,"ecosystem":null,"applyGuide":null,"weight":104,"participationForm":"Virtual","field":"datacleaning,machinelearning,generativeAI,datascience","email":"gesatucla@gmail.com","contacts":"Email,gesatucla@gmail.com","externalContacts":null,"telegram":null,"discord":null,"whatsapp":null,"wechat":null,"prizeFundingEscrowed":0,"token":"USD","prizeSafeForm":0,"prizeSafeNetwork":null,"prizeSafeOwner":null,"prizeSafeWallet":null,"prizeSafePublishDate":null,"hasOnboarding":false,"isPublicTrackJudging":false,"isEndedTrackJudging":false,"question":null,"countryId":null,"cityId":null,"venueName":null,"venueAddress":null,"category":3,"isEnableBounty":true,"isBountyRequired":false,"isEndedBountyJudging":false,"bountiesLimitForBuidl":10,"isPublicBountyJudging":false,"isMultiTracksAllowed":false,"tracksLimitForBuidl":0,"status":2},"status":4,"bounty":null,"track":null,"table":"0","trackObj":{"id":1044,"name":"Main","judgingCriteria":"- Accuracy and Predictive Performance:\nProjects will be evaluated based on the accuracy and reliability of their predictive models. \n\n- Innovation and Creativity:\nJudges will assess the novelty and creativity of the solutions, especially in terms of how participants handle data integration, synthesis, and privacy-preserving techniques.\n\n- Scalability and Efficiency:\nThe ability of the solution to scale efficiently and process data in real-time will be evaluated. Projects need to demonstrate performance stability and resource management.","weight":0},"bountyObjs":[]}],"chain":[],"chainList":[],"appliedGrant":false,"acceptRewards":false,"owner":142169,"ownerInfo":{"nickname":"tennyfr","username":"U_ad042c9f261a14","avatar":"https://cdn.dorahacks.io/static/files/18fc77e0ac7b3d35e403a224c2b94620.png"},"visible":true,"deleted":false,"continuousSponsoring":false,"pageView":0,"sponsoredCount":0,"starCount":0,"sponsoredAmount":"0.00","sponsoredPersons":0,"followers":0,"voteCount":0,"projectMembers":[],"bnbRound":null,"doraGrantDaoTags":[],"techTrees":[3],"infrastructureTags":[],"otherInfrastructures":[],"lastEditor":{"id":142169,"username":"U_ad042c9f261a14","date":1719599583000},"upvoted":false,"followed":false,"upvoters":{"count":0,"list":[]},"kycStatus":0,"ideas":[],"name":"DataGuardians","techniqueField":"","source":"-","vision":"Securely handling and processing confidential data","projectDescription":"We created a VM to confidentially open and connect the two datasets. Then, we analyzed the original dataset and created a synthetic dataset using a GAN. The link to the slides is in the attached GitHub repo, which includes a larger explanation (trustworthy hackathon.pdf) and other links to the code for each portion, some code is in links in the pdf.","teamDescription":"Lucas Chin, Data Science, University of California - Irvine, ‘27\nChloe Houvardas, Computer Science, Queen’s University, ‘27\nChi Lu, Data Science, University of California - Irvine, ‘25\nDaviya Shah, NYU, 26'","githubPage":"https://github.com/sajio1/Trustworthy-Hackathon","githubVerify":0,"demoLink":null,"demoVideo":null,"ownerTitle":null,"ownerIntroduction":null,"twitterVerified":false,"weibo":null,"nft":null,"scOwner":"","editTime":1719599583,"upvotes":"0.00","isRecruitingEnabled":false,"recruitingInfo":"","recruitingLink":null,"twitter":null,"facebook":null,"instagram":"lucaschin_","socialTelegram":null,"socialDiscord":null,"socialLinks":[],"showedFunding":false,"createdAt":"2024-06-22T13:56:15.693658+08:00","updatedAt":"2024-06-29T02:33:03.704122+08:00","milestoneEditor":142169,"path":"/buidl/13818","techField":"","techFieldAry":[],"techTree":[3],"tags":[],"otherTags":[],"origin":"-","intro":"We created a VM to confidentially open and connect the two datasets. Then, we analyzed the original dataset and created a synthetic dataset using a GAN. The link to the slides is in the attached GitHub repo, which includes a larger explanation (trustworthy hackathon.pdf) and other links to the code for each portion, some code is in links in the pdf.","images":["https://cdn.dorahacks.io/static/files/1903e8431bf74efa53934a54f77ad8e7.png"],"team":"Lucas Chin, Data Science, University of California - Irvine, ‘27\nChloe Houvardas, Computer Science, Queen’s University, ‘27\nChi Lu, Data Science, University of California - Irvine, ‘25\nDaviya Shah, NYU, 26'","github":"https://github.com/sajio1/Trustworthy-Hackathon","demo":null,"video":null,"socialTwitter":null,"socialFacebook":null,"socialInstagram":"lucaschin_","nftId":null,"haxonBuidls":[{"id":14747,"projectId":13818,"name":"DataGuardians","track":"Main","bounties":[],"haxonId":405,"haxonName":"Trustworthy AI Lab x GES UCLA Hackathon","status":2,"haxonCreateAt":"2024/05/29 21:50:00 UTC"}],"kudoData":null,"bountyObjs":[],"badges":["{\"schema\":\"BADGE\",\"d\":{\"index\":\"b11099\",\"badgeType\":\"buidl\",\"path\":\"/buidl/13818/achievement/b11099\",\"ownerId\":13818,\"iconName\":\"b0\",\"updatedDate\":\"2024/06/22\",\"title\":\"Trustworthy AI Lab x GES UCLA Hackathon\",\"detail\":\"Participation\",\"tag\":\"Hackathon\",\"imageUrl\":\"https://cdn.dorahacks.io/static/badge/b0_Trustworthy~c32AI~c32Lab~c32x~c32GES~c32UCLA~c32Hackathon.webp\",\"shareImageUrl\":\"https://cdn.dorahacks.io/static/badge/b0_Trustworthy~c32AI~c32Lab~c32x~c32GES~c32UCLA~c32Hackathon.jpg\",\"id\":11099,\"metaInfo\":{\"type\":\"HACKATHON_PARTICIPATION\",\"hackathonId\":405,\"title\":\"Trustworthy AI Lab x GES UCLA Hackathon\"},\"createdAt\":1719039062,\"updatedAt\":1719039062,\"type\":0,\"subjectId\":405,\"metadata\":{},\"buidl\":13818}}"]}}