Grateful to be in grad school, where I am surrounded by amazing people and basically get paid to learn
Michael Zhang
445 posts
Michael Zhang
@michaelrzhang
neural networks / robotics research @amazonscience. Prev: PhD @UofT, @UCBerkeley. Journey before destination.
- Life update: I've recently moved to Boston and started a job @AmazonScience ! I'm excited to explore - please share local recs and let me know if you want to grab coffee! (picture: White Mountains, NH)
- We are excited to launch a pilot Graduate Application Assistance Program @UofTCompSci! Current graduate students help review pre-submission application materials with the focus on guiding underrepresented applicants. Details:
- Is it possible to obtain different models along a rate distortion curve without a lot of hyperparameter tuning+retraining? In our new preprint, we show that the answer is yes!
- Teaching a course (Intro to Machine Learning @UofTCompSci) this semester and gave my first lecture today! Simultaneously fun, exciting, and humbling
- Train CIFAR-10 to 94% accuracy in less than 10 seconds(!) on a single A100: github.com/tysam-code/hlb…
- Multi-Rate VAE received an oral presentation at ICLR! We've also made a video: iclr.cc/virtual/2023/p…Is it possible to obtain different models along a rate distortion curve without a lot of hyperparameter tuning+retraining? In our new preprint, we show that the answer is yes!
- I'm excited to be back in the bay area this summer to work on self-driving at Tesla!
- Toronto to Niagara Falls ride! Pushed my limits and had a lot of fun
- Minerva (neural net) is like the student that writes down everything it knows on an exam to try to get partial credit
- I am excited to share our ICLR paper on autoregressive dynamics models for RL. Explicitly capturing conditional dependencies between state dimensions improves forward dynamics and reward prediction, yielding SOTA off-policy evaluation on control tasks. arxiv.org/abs/2104.13877
- Proud brother moment last week - my sister @cindyzhangx graduated from MIT and is starting her PhD @PrincetonCS in the fall!
- It seems common for researchers to experience stress around the holidays due to conference deadlines early in the new year. How can we (collectively and individually) alleviate some of this stress?
- Excited to share Boosted Prompt Ensembles! Inspired by boosting algorithms, we propose an algorithm that automatically grows an ensembles of prompts to cover a target problem space. Led by @silviupitis and with @andrew_wang10 and @jimmybajimmyba
