🚀 Inspiration
Remember the Lee Hsien Loong deep fake incident? 😱 That’s just one example of the huge problem we’re tackling: online trust and safety. In tech-savvy places like Singapore, where internet use is sky-high 🌐, ensuring strong online security is a must! But it's not just about protecting people from hackers—it's about keeping the digital space safe, clean, and trustworthy.
We already have some cool initiatives, like:
- 🛡️ Online Safety (Miscellaneous Amendments) Act: Strengthens the rules for keeping the online world safe!
- 🖥️ Centre for Advanced Technologies in Online Safety (CATOS): Building awesome tools to detect harmful online content.
But, let’s be real—new scams are popping up faster than we can say “deep fake” 😜. Even with these measures, detecting harmful content is still a huge challenge. That’s where we come in! Our mission is all about making the web a safer place by tackling cybersecurity threats, wiping out harmful content, and stopping the spread of misinformation. We also focus on social sustainability—making sure the internet is a welcoming space for everyone. 🌍💪
We’re inspired by the UNSDG 11 (Sustainable Cities) and UNSDG 16 (Peace & Justice) to make online spaces safer, fairer, and more inclusive. Let’s make the web a place where everyone feels safe and empowered! 🌟
🤖 What it does
We built something cool, here’s what it does:
- Detects harmful memes 🤡: No more spreading toxic memes!
- Detects fake news 📰: Flag the fake, support the real, and let users make informed decisions!
- Detects deep fakes 🤥: Keep people safe from digital scams and impersonation.
- Focuses on social sustainability 🌱: Because the digital world needs to be healthy, too!
- Fosters a positive online environment ✨: Everyone should feel respected, protected, and empowered in online spaces.
💻 How we built it
- Backend: Powered by Python 🐍 (because it’s awesome at deep learning)!
- Frontend: Crafted with TypeScript, HTML, JavaScript, and CSS 🎨 (it’s like digital art)!
- Deep learning models: Built and trained using Google Colab 🧠 to make our predictions smart and fast!
🛠️ Challenges
- Learning TypeScript: Not gonna lie, we had a little struggle bus moment 🚍, but we got there in the end! 😅
- Integrating the models: Getting the deep learning models to talk nicely with the front-end was tricky, but teamwork made the dream work! 🤝
🎉 Accomplishments
- 24 hours 🕐 to get this awesome project live! Talk about a power team! ⚡
- Overcame the TypeScript challenge and now we’re TypeScript pros 💻💡.
- Launched a working product that detects harmful content—go team! 🎯
📚 What we learned
- Deep learning: We learned how to teach machines to see and understand text and images 🧠👀.
- TypeScript: It was a new skill, but now we can code like pros in it! 😎
- Teamwork: We learned that we can totally rock it when we stick together! 💪
🚀 What's next for Poseidon
- Trend analysis dashboards 📊: Let’s analyze the crazy trends in harmful content and keep track of what’s going on!
- Model updates 🔄: We’ll keep our models sharp and always ready for new threats!
- Data storage 💾: We’re adding a cool database so we can store, analyze, and improve our models constantly!
- User engagement 🌐: Adding features to let users contribute to a safer digital space—because together, we’re stronger! 💥
Built With
- ai
- deep-learning
- github
- javascript
- keras
- news
- next.js
- pandas
- python
- requests
- transformers
- typescript
Log in or sign up for Devpost to join the conversation.