Marcel Hussing
@marcelhussing.bsky.social
PhD student at the University of Pennsylvania. Prev, intern at MSR, currently at Meta FAIR. Interested in reliable and replicable reinforcement learning, robotics and knowledge discovery: https://marcelhussing.github.io/
All posts are my own.
All posts are my own.
We also empirically evaluate the algorithms. We first demonstrate that the sample complexity bounds are not representative of average case performance. Then, we derive insights for deep RL with discrete action spaces.
💡 Quantizing actions leads to agreement across policies! (5/6)
💡 Quantizing actions leads to agreement across policies! (5/6)
October 26, 2025 at 2:16 PM
We also empirically evaluate the algorithms. We first demonstrate that the sample complexity bounds are not representative of average case performance. Then, we derive insights for deep RL with discrete action spaces.
💡 Quantizing actions leads to agreement across policies! (5/6)
💡 Quantizing actions leads to agreement across policies! (5/6)
New ChatGPT data just dropped
August 7, 2025 at 11:29 PM
New ChatGPT data just dropped
Just arrived in Montreal for my internship at FAIR. So far Montreal has been amazing, great walkable areas, good food and nice people! Although I must say I have to get used to being addressed in French 😅
May 26, 2025 at 4:23 PM
Just arrived in Montreal for my internship at FAIR. So far Montreal has been amazing, great walkable areas, good food and nice people! Although I must say I have to get used to being addressed in French 😅
Me, every 4 months:
November 25, 2024 at 1:44 PM
Me, every 4 months:
Coffee-making robot, I need one of those at home. Anyone got a cost estimate?
November 17, 2024 at 5:55 PM
Coffee-making robot, I need one of those at home. Anyone got a cost estimate?
One of the exciting things that I didn't value as much in the beginning is the oracle assumption. Essentially, we assume that you can approximate your value function of a fixed policy well enough. Since the policies don't change, this is simply saying we can do on-policy policy evaluation.
November 10, 2024 at 5:51 PM
One of the exciting things that I didn't value as much in the beginning is the oracle assumption. Essentially, we assume that you can approximate your value function of a fixed policy well enough. Since the policies don't change, this is simply saying we can do on-policy policy evaluation.
Hello bsky! I'm a PhD student at UPenn where I mostly work on #reinforcementlearning. I am interested in building #AI to obtain new knowledge reliably. Right now I'm an intern at #Microsoft research doing transformer stuff. Here is what ChatGPT thinks my life looks like:
What's everyone else up to?
What's everyone else up to?
November 9, 2024 at 3:38 PM
Hello bsky! I'm a PhD student at UPenn where I mostly work on #reinforcementlearning. I am interested in building #AI to obtain new knowledge reliably. Right now I'm an intern at #Microsoft research doing transformer stuff. Here is what ChatGPT thinks my life looks like:
What's everyone else up to?
What's everyone else up to?