Hacker News new | past | comments | ask | show | jobs | submit | edhu2017's comments login

Data-driven robot control methods for solving furniture assembly.

It's an interesting problem, requiring both dexterous manipulation and long-term planning. It's also compositional, so I believe some form of hierarchical control and planning can solve it.

www.clvrai.com/furniture


Hey all, I'm one of the authors of the environment. I'm excited to see what researchers will come up with to tackle the furniture assembly problem. Let me know if you have any questions or comments.


Disclaimer: Our field is extremely competitive (deep learning) and other labs may have better working styles depending on their lab size and member composition.

My lab's workload on average is 8 - 12 hours a day for 6-7 days a week, ramping up even more closer to deadlines. That being said, a lot of this time (50% or more) is for thinking and discussing. This schedule, while hard, seems to be effective for publishing in good conferences.


If half of the time is thinking, why do you need to be on the spot? Seems really counterproductive to me.


Being in the lab helps me focus on my thinking. I also discuss my current thinking and problems with my peers for critique, and I can do the same for them. I'm more conditioned to do research, and my mind wanders less in the lab.

However, I do agree that having free time and being away from the lab is important to get a fresh view of things.


He probably means planning meetings.

Never understood what the purpose is to fill half the day with meetings. Beside the illusion of productivity.


We don't plan too many meetings. Once a week for a lab wide meeting. Most of these meetings are casual and just discussions between cliques of phd students on specific topics.


But how do you know that 6-8 hours 4-5 days a week wouldn’t be just as effective? I assume you don’t really have anything to compare it to.


This seems pretty excessive even for one of the very competitive benchmark focused areas in deep learning. It is not necessary and probably harmful, and lots of very good groups don’t do this kind of thing.


This is AI snakeoil.


I don't care if it's AI snake-oil or a bunch of grandmothers in a yurt in Nepal.

<MontyPython>It works, mate!</MonthyPython>


Agreed, reimplementation is pretty good for students to learn. Our lab requires everybody to work on an implementation project


definitely has been done before.


seems great for my reinforcement learning models. instead of parsing my hyperparameters through the tensorflow cli API and editing the training file a line at a time to take in an additional hyperparameter, I can just directly set them through the cli with fire.


but then how will you keep track of which parameters worked well? I've been essentially storing my kwargs in json and not felt a need to control anything directly from the CLI.


Good question, for me I programatically generate a folder with the hyperparameter key - values in the name and store the checkpoints under it. As you can imagine, it can get out of control quickly if not managed well, but it works for 1-3 person projects. For anything more large scale or organized, I would recommend looking into Comet ML which lets you query and filter your experiments by hyperparameter ranges instead of manually looking at folder names.


I did that until I hit the Linux directory name length limit, lol. Now I hash the hyperparameters dict to get the directory name, and store a json file within. Totally ad hoc and I'm sure a better solution exists.


I think it's a great move. Prevents short term greedy shareholder strategies.


By paying a large premium to the all time high? If he's worried about the stock going down, paying more for it doesn't make a ton of sense. You should wait until the stock is down and then take it private.


By paying a 20% premium he's ensuring that the short sellers will need to buy shares at the premium price guaranteeing that they've lost at least 20% of what they've gambled.


Ah yes, overpay so you can stick it to the big mean short sellers. Might as well go for 50% premium with that logic!


Depends on how simple you want to get. If literally displaying text once in a while, you don't even need jekyll. If you are a blog writer, want some sort of templating / reusable snippets, or want cool themes then jekyll is good.


fight on! these projects are interesting but i'm still a student


Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: