Pinned Tweet
0:30
2.8K views
2
37
Follow
Click to Follow AndrewMendez19
Andrew Mendez
@AndrewMendez19
Deep Learning , Snapchat Lens Studio Creator. Previously , ,
Andrew Mendez’s Tweets
Shape-aware Text-driven Layered Video Editing
abs: arxiv.org/abs/2301.13173
project page: text-video-edit.github.io
1:12
11.7K views
44
199
Photographer Vadim Trunov has had close encounters with squirrels in his career, but these photos of squirrels shooting photos of each other and of birds are probably the most captivating he took
[read more: buff.ly/2ScNeji]
315
22.1K
136K
Today I received my *first* citation ever! I was hesitating to post this because my circle is full of very experienced and mature academics, but I remember the importance of celebrating milestones no matter how small. Here’s to the first out of many 🥂
141
223
6,083
Show this thread
Are you a PhD or Masters student at , with a background in machine learning? I’m offering an advanced topics class on interpretable and explainable ML this spring. Please help spread the word! It’s going to be fun 🤩
2
7
78
How does the impressively human-like #ChatGPT get computational knowledge superpowers? Give it a Wolfram|Alpha neural implant!
writings.stephenwolfram.com/2023/01/wolfra
107
1,115
4,405
I have a feeling this is never gonna stop being shared lol twitter.com/DentesLeo/stat
This Tweet is unavailable.
127
273
9,191
What if you could play a platformer by moving in real life? I made an AR game on Spectacles where I can control my Bitmoji by walking and jumping. Thought it might help me exercise a bit more... 🏃♀️🤸♀️😆 #augmentedreality #spectacles #xr #ar
0:19
20.9K views
21
75
376
I was able to prototype last year, really recommend to follow this account, it works really well!
Quote Tweet
Hello, world! So, what is KegWatch? KegWatch is a plug and play device that sits on top of your #kegerator to track all sorts of fun #beer #statistics. No cut beer lines. No electronics inside the kegerator.
Check out our website to learn more
kegwatch.io
Cheers! 
watch a 2 layer neural network learn to separate two classes to the left and right
0:32
468.9K views
161
1,934
14.7K
Show this thread
Time lapse of a dandelion flower transforming to seed head
Filmed over 1 month by Neil Bromhall
0:14
449.7K views
From
Sunlit Rain
108
2,901
17K
Show this thread
Machine Learning Engineering for Production (MLOps) - DeepLearningAI
Building models is one thing. Making them useful is another thing. The first course of the MLOps specialization teaches the ML lifecycle & deploying models.
Course 1 is free on Youtube: youtube.com/playlist?list=
45
602
2,441
do NOT buy the avengers off shein 😭
0:58
2.3M views
832
8,512
82.2K
Show this thread
Depth2image in Stable Diffusion is a blast ⬇️ Love the structural consistency, allowing me to explore different styles for this virtual set. This technique will be a boon for set design and quite frankly home decor.
0:15
13.2K views
7
45
308
Show this thread
How bad is grade inflation at Harvard College? If trends keep up, an average student in ten years will have a perfect 4.0. thecrimson.com/article/2022/1 (h/t )
420
1,455
6,054
1
2
Quote Tweet
Delighted to share our new @GoogleHealth @GoogleAI @Deepmind paper at the intersection of LLMs + health.
Our LLMs building on Flan-PaLM reach SOTA on multiple medical question answering datasets including 67.6% on MedQA USMLE (+17% over prior work).
arxiv.org/abs/2212.13138
Show this thread
7
32
304
If you think you know about trees, look at this.
106
1,377
7,010
Show this thread
Hi folks, I’m afraid it’s time for me to say goodbye. Not just leaving Twitter, but the whole show. I’ve been battling cancer last 2 years, but now only have a few days left now. Thank you wonderful people, I leave this crazy world with much love in my heart ❤️
57.3K
41.7K
786.1K
Still trying to process my emotions on learning that blacklisted me. The thought that will keep me up tonight: censorship of scientific discussion permitted policies like school closures & a generation of children were hurt.
1/2
1,692
19K
85.2K
Show this thread
My last work at Snap launched 👏 So happy, these are great business achievements for companies that will help a lot of AR creators to monetize their talent and skills! And it was so fun explore this area, and work with such a cool product, design, engineering teams 🔥
Quote Tweet
Snapchat is testing Lenses with power-ups and upgrades that you can buy tcrn.ch/3YcNpfk by @aiishamalik1
2
1
27
Safety Tips Don’t Work!
0:51
27.1M views
8,593
130.5K
694K
Show this thread
588
6,392
53.1K
🥹🥹 top class this.
Quote Tweet
@JackGrealish kept his promise to Finlay!Show this thread
0:17
547.1K views
1
7
71
I was told by a Professor when I was a PhD student not to get married before getting a faculty position (I was already married at that time), and not to have kid before getting tenure (I had my first kid right after PhD). I don’t regret my decisions.
Quote Tweet
If this platform is indeed dying, I think this is my most important post.
I am still haunted years later by someone's story on never having a kid because her advisor said pregnancy would make her stupid, she wouldn't get tenure, and then there was never a good time.
#academicmom twitter.com/gmishne/status…
7
15
222
DiffusionDet: Diffusion Model for Object Detection
abs: arxiv.org/abs/2211.09788
github: github.com/ShoufaChen/Dif
14
172
695
With the recent burst in applications of Large Language Models (LLMs), it's worth revisiting the Gell-Mann Amnesia Effect and Knoll's Law (1/8)
5
37
162
Show this thread
Real-time ML in-painting with - had the pleasure of assisting with one of the templates for this awesome tool.
0:11
4K views
11
32
197
Show this thread
I’m looking for PhD students in AI and Security . If you are interested in working with me,
- Please apply to University of Maryland cs.umd.edu/grad/catalog due 12/16/2022.
- Fill out the form bit.ly/yizhengresearch and send me an email. Thank you!
1
48
96
Don’t care about the Astros, but nothing but love fromer leader, . Congrats, we’ll deserved! The ultimate comeback story of one of the greatest guys in baseball. #F16HT
Quote Tweet
1
3
5
Show this thread
This paper really shows the power of a generic seq2seq architecture for transcription. Instead of transcribing all the notes of the input, just transcribe a piano cover instead! Works surprisingly well, even on songs not seen during training.
Amazing work, !
4
34
This is a truly autonomous augmentation that doesn’t require a pool of candidate types of augmentation to start. Autonomous augmentation is tremendously helpful when we have no domain knowledge of what augmentations may help, e.g. medical images.
3
9
I'm on the job market! I am especially excited about industry research and development to push machine learning forward in impactful domains. My focus is weak supervision and self-supervised learning, and I truly enjoy expanding into new areas as well. Please reach out or RT 🙏
1
19
37
I’m thrilled to announce that I will be joining University of Maryland, College Park as an Assistant Professor this Fall. I will start my lab on assistive/wearable devices for computing our subjectivity and perspective. Looking forward to working with you all!
34
26
396