π Congratulations! You’ve discovered a hidden easter egg page. Here, I’ll share a glimpse into my life beyond academia. You got to see these stuff cause I’m a huge information hoarder.
I use the same ID (SuoZhang) for all social media, kinda like a “Alter-Ego” for me (with the figure in the picture above). I have an old blog at https://a-suozhang.xyz/ since my undergrad years, it’s immature and dummy but I am not ashamed of it, okay, maybe a little.
I’m an authentic nerd from my teenage years. I’m a huge fan of TRPG (Table-top Role Playing Games), such as DND (Dungeons and Dragons), and had joined and hosted several campaigns (not so frequent recently). I also play MTG (magic the gathering) card game (also not so frequent recently).
I like pressing some shutters on trips, and sometimes random walks in the city. Sometimes I share them in platforms like Twitter, Red Book, or G-core.
I enjoyed music from multiple genres, and learning guitar and keyboard now. Hopefully I could be a rookie bedroom musician in a year or so with my MiniLab 3.
I’ve been conducting research about visual generative models since 2023. Therefore, I also played with diffusion models for a bit of my own entertainment.
I’ve created a series of emoji packs based on my virtual persona (you could find it on wechat searching “suozhang” in emojis). They achieves nearly 10,000 downloads each, significantly surpassing my open-source code.
I also create my custom MTG cards with diffusion models based on the character and story in my DND campaign (yeah, that nerdy), see further in this G-Core Post.
In my undergrads years, I wasted my time on a startup project with a bunch of my fellas (tech lead Song Tiancheng, and art director Liu Yuxin) called Marvis (Jarvis + Muscle), a interactive assistant for electromyogram (EMG, electrical signal when muscle moves). I summarize it into this post when I was seeking grad student positions.
We build the software and hardware system that acquire muscle electrical signal from multiple parts of the body (attached on a shirt), and design algorithm for action recognition for human-computer iteration. You could literally control electrical devices with your “bare hands” like Iron man.
It won some awards for student startup competitions, and got reported by Beijing-TV.
We also have a cool demo video here, If you are interested: