Loading
Feedback

anssi 217

Organization

University of Eastern Finland

Location

FI

Badges

0
1
2

Connect

GitHub

Activity

Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Mon
Wed
Fri

Ratings Progression

Loading...

Challenge Categories

Loading...

Challenges Entered

3D seismic image Interpretation by Machine Learning

Latest submissions

See All
graded 83204
graded 82875
failed 82870

Sample-efficient reinforcement learning in Minecraft

Latest submissions

See All
graded 82725
failed 82639
failed 82638

Sample-efficient reinforcement learning in Minecraft

Latest submissions

See All
failed 25431
graded 25415
graded 25316

A new benchmark for Artificial Intelligence (AI) research in Reinforcement Learning

Latest submissions

See All
graded 8597
graded 8541
graded 8531
Gold 0
Silver 1
Boltzmann's Favourite
May 16, 2020
Bronze 2
Trustable
May 16, 2020
Newtonian
May 16, 2020

Badges

  • Has filled their profile page
    May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020

  • May 16, 2020
  • Kudos! You've been awarded a silver badge for this challenge. Keep up the great work!
    Challenge: NeurIPS 2019 : MineRL Competition
    May 16, 2020
  • Kudos! You've won a bronze badge in this challenge. Keep up the great work!
    Challenge: Unity Obstacle Tower Challenge
    May 16, 2020
Participant Rating
shivam
MasterScrat
Participant Rating
MasterScrat
shivam

NeurIPS 2020: MineRL Competition

Question about the MineRLObtainDiamondVectorObf-v0's record v3_quiet_mandarin_orange_ghoul-13_257-145846

About 1 month ago

I believe that is one of the recordings that got corrupted at some point along the line and should be ignored for training purposes. See this blacklist of games that are ignored by the dataloader. Looks like this specific game is not included in that list.

@BrandonHoughton
Do you know if the above blacklist works with the current version of data (i.e. with “v3_” names)? On a quick inspection it checks if names match, and since that blacklist does not include the “v3” it might let those samples through.

Wall torches

About 1 month ago

Fair points. The task is technically doable with just given items, but indeed it is difficult to navigate underground based on visual information when there is nothing visible :). Like you pointed out, the fact that demonstrations can have these makes it worse. I can see the dilemma though: including all of the different actions would make the training a lot harder (more confusing actions that have little relevance towards obtaining diamond), and the training is hard-enough as it is.

As for “smelt coal” action, there is this thread from 2019 edition about this.

Why this user can submit 42 times?

About 2 months ago

I highly recommend you familiarize yourself with the environment as soon as possible, perhaps try your own ideas on the simpler tasks (TreeChop). This way you should not run into bigger roadblocks when you play around with the baselines, taken from the Discord channel.

There are some baselines available from other players already:

Why this user can submit 42 times?

About 2 months ago

@youkaichao

You need to use the submission template, but evaluate_locally.sh is only used for testing if the code runs correctly on your local machine (judging by the script name). I use something similar which just skips most of the steps in evaluate_locally.sh, boots the image and runs run.py.

As for why no other submissions: There is still time in Round 1 and people usually start slow with the submission, based on observations on previous competitions. I am willing to bet we will see bunch of participants soon enough :slight_smile:

Why this user can submit 42 times?

About 2 months ago

Ah alright, that’s a good observation that you could raise in the submission template as well. I use my own scripts to run lighter evaluations (not full ones) to only see if the Python code runs correctly, and then assume rest of the full evaluation works as intended on the server.

Why this user can submit 42 times?

About 2 months ago

@youkaichao

I apologize for down-talking the issue you have risen (which is valid), but the environment works for training and testing purposes for us (but we do not use the seed function). Similarly the submission template works as-is, which I already submitted twice as a test successfully.

I recommend you join the Discord channel linked on main competition page. There are people there that can help you with questions.

Error using gym.make

About 2 months ago

I am not sure if MineRL runs on Colab, since you need some specific versions of libraries and Java specifically. See docs on installation.

Done is never set to True

3 months ago

There was a bug in MineRL 0.3.1 where episode timeouts were not set. Upgrading to 0.3.3 (or soon-to-be-shared 0.3.4) should fix the issue.

Preprocessing rule

3 months ago

@william_guss @BrandonHoughton

I know you guys are busy updating MineRL, but could you also shine light on the questions above (in this specific context)? :slight_smile:

Preprocessing rule

3 months ago

Quoting W. Guss from Discord on a very similar question:

We want to make it ILLEGAL for participants to make solutions like this that USE game knowledge.
Beyond stacking actions, you could imagine people using “movement detectors” to reverse engineer the action obfuscation, and then hard-code sequences/options (similar to scripted crafting).

Even with a rule against “hard coding” like above, I could just give these extracted action sequences to a policy as options for it to execute, but crucially I am using knowledge about minecraft to produce those options, rather than learning them from data with a “blind eye” to those options existing based on the rules
I think data augmentation in the style of CURL or image flipping etc, doesn’t encode game knowledge, but still it’s hard to write down a rule like this precisely

If your solution wouldn’t work on other domains that you knew nothing about (minus hyper parameters), you’re probably doing something wrong. e.g. let’s say we randomly trained your solution on an RTS like Starcraft, if it’s not general enough to work at all in that setting, that’s bad and we don’t want people to develop solutions which overfit to minecraft .

So if you “extract” options by knowing that there is a fixed sequence of items in the game or whatever, that’s overfitting to minecraft and not the intention of the competition .

For all intents and purposes, the environment should just be a black-box to your algorithm with Pixel + 64 dimensional vector observations and 64 dimensional acton vector

Thats not to say, you can’t do techniques like “automatic option extraction”, it’s just that those technqieus should be generic

So if that means e.g. training an auto-encoder on all of the (state,action) pairs and then doing $k-means$ to get clusters of actions, and then training a neural network for each option. That’s good!

NeurIPS 2020: Procgen Competition

Competition metric seems to favor sample efficiency over generalization gap

4 months ago

Following is from the point of view of the competition. I agree it might be difficult to disentangle benefits of the method in sample efficiency and generalization with this limit.

From practical point of view, it levels the playing ground (somewhat) when the amount of training data is restricted. In Unity Obstacle Tower challenge, for example, one team trained for billions of timesteps, which was simply out of the reach of competitors without access to +50-core, multi-gpu setup. They can still use that hardware for faster evaluation of ideas/models/hyperparams though.

This also encourages for ideas other than “just brute-force it” with compute. It is easier to try out ideas when you know the limit of how long you should train :slight_smile:

NeurIPS 2019 : MineRL Competition

About the rule on pre-trained model

11 months ago

I would assume no pretrained models are allowed, and this is enforced by removing “large” files.

If so, it is too relaxed. Our models are well below 10MB, regardless of the network architecture, and even if models were larger you could prune them to 90% of the original size and still retain most of the performance.

When will the results of round 1 be announced?

11 months ago

I do not have an answer for that, but organizers are posting clips of top agents on Discord, most recent from roughly 6h ago. Linking them here so non-Discordians can also see them:

Partially rendered env in MineRLObtainDiamondDense-v0

11 months ago

My wild guess is this has something to do with rendering of chunks and/or world-generator generating chunks. Similar things happen if you travel too fast to one direction (you reach “edge of the world”, but it will appear eventually). Not sure why previously rendered chunks would disappear like that.

Equip item failed

12 months ago

Sorry for double-posting, but just for clarity:

Tested on Ubuntu 18.04 (Python 3.6) and Windows 10 (Python 3.7) with MineRL 0.2.4:

  • Create ObtainDiamond-v0 environment, control by manually creating actions
  • Open chat (press “T”) and give some items /give @a wooden_pickaxe 1 and /give @a stone_pickaxe 1.
  • Wooden pickaxe is equiped (auto-equip the first slot).
  • Using equip="stone_pickaxe" or equip="air" does not work, nor do the integer versions (3 and 1, respectively)

How to use furnace to "nearbySmelt" coal?

12 months ago

Tested this on Windows 10 one more time, still no luck :(.

Random thought: Perhaps it works but it creates charcoal, and since charcoal != coal it won’t show up in the inventory list? They are very distinct items in Minecraft, but I am not sure if Malmo or your code abstracts this one out.

How to use furnace to "nearbySmelt" coal?

12 months ago

I tested it like this (Ubuntu 18.04, Python 3.6, MineRL 0.2.4):

  • Create env and control agent by feeding user input
  • Focus on game, press “T” and command yourself necessary items (/give @a furnace 1, \give @a log 64 and /give @a planks 64)
  • Place furnace, face it and try to use nearbySmelt=2 (="coal"). Nothing should happen (no crafting/smelting).

Repeated reward for logs and furnace

12 months ago

Tried this as well with no luck :frowning: (tried “wooden_pickaxe”, “none”, “air” and “stone_pickaxe”)

How to use furnace to "nearbySmelt" coal?

12 months ago

Ah my bad. Indeed you should be able to burn logs into charcoal in default Minecraft but indeed the action nearbySmelt[coal] does not do anything after quick testing. It is either bugged or refers to something else (remnant of traditional smelting action?).

This has also been discussed here: How is the "reward" on leaderboard page computed? (Edit: Oh, it was you who reported this ^^’)

Equip item failed

12 months ago

+1 on the equip command not working (can not unequip items nor equip anything). In fact it did not work in version 0.2.3 for us either.

How to use furnace to "nearbySmelt" coal?

12 months ago

MineRL uses Malmo’s functions to make such smelting easier, essentially turning it into crafting. You only need to have iron ore and coal in your inventory, and once you do nearbySmelt iron you should get an iron ingot immediately. This was discussed on some other thread here on same channels.

Block become much small when looking from top

About 1 year ago

That depends on the field-of-view of the player. You seem to be playing with the default 70 FOV (the lower picture), where as the recordings seem to be done in maximum 110 FOV (upper picture). This is based on quick round in vanilla Minecraft.

Main hand observations

About 1 year ago

I believe maxDamage refers to how damaged the tool can get. I.e. If damage = maxDamage, then tool is broken and can not be used (as per Minecraft mechanics).

As for the none and air type: I am puzzled by this as well. Does one of them of never happen (they mean the same thing), or does none mean the player is not able to hit anything (“doesn’t even have hands”)?

Unity Obstacle Tower Challenge

Different results between training, debug and evaluation

About 1 year ago

Could it be you overfit to the 100 seeds we have on shared binary of OT environment? AFAIK the evaluation uses seeds outside these 100.

We had something similar: Training model further did not make results more reliable in evaluation.

Why I was blocked from uploading new models for 3 days

About 1 year ago

I use Linux (Ubuntu 16.04), and I dunno it works for other platforms.

Why I was blocked from uploading new models for 3 days

About 1 year ago

Try running the environment without docker with same environment variables and port argument as in the examples. This works for me when I test my submission image.

Is the evaluation seed truly random?

About 1 year ago

As I have understood it, the seeds are fixed (e.g. here Juliani tests one of the seeds). The wording on main page about random five seeds probably refers to five randomly preselected seeds.

UnityTimeOutException in evaluation

About 1 year ago

We only updated one of the model files after a successful submission. Worker ID is set to zero and environment filename to None.

Looks like that submission (29) did eventually manage to run through. Did you happen to re-queue the submission or did it work out on itself? Edit: Looks like you re-queued it.

Edit2: Looks like that submission (29) is now stuck and does not finish correctly ^^’

Edit3: Looks like we figured it out: When testing locally outside docker images, the OT game ended up stuck with empty screen (skybox) and our code also waited for the OT environment. We first had to launch our agent code and wait until it tries to connect to OT environment (“Start training by pressing the Play…”), after which we launched OT env and the evaluation started successfully. The solution for submission was to modify agent code to first create OT environment before everything else, including larger imports (TF, Torch, etc), and then proceed with creating/loading agents.

UnityTimeOutException in evaluation

About 1 year ago

@mohanty: That’s odd o: . I have another debug run going on right now, and it seems to be stuck. So far 500 seconds waited and no luck: https://gitlab.aicrowd.com/Miffyli/obstacletower-2019/issues/29.

But I will keep trying every now and then, if it would eventually work out. Thanks for the help!

UnityTimeOutException in evaluation

About 1 year ago

Looks like I am running into same issue as described above: Previously submitted code suddenly is not working, and debugging gives me the timeout error. timeout_wait=300 did not solve the issue, but I will try with longer values. Here is one of the failed submissions: https://gitlab.aicrowd.com/Miffyli/obstacletower-2019/issues/25 .

Edit: timeout_wait=900 did not help either, still getting same error :/. https://gitlab.aicrowd.com/Miffyli/obstacletower-2019/issues/26

Submitting in debug mode sets leaderboard score to 0

About 1 year ago

Yes. Also note that debug runs are done with different seeds than final evaluation.

More information here: Announcement: Debug your submissions

Architectures of Round 1 winners

Over 1 year ago

Apart from using improved Dopamine Rainbow, here is a short list of nuggets of information other people have shared:

Vector Observation contents

Over 1 year ago

Looking at how ObstacleTowerEnv handles this observation, the vector tells number of keys as one-hot encoding (or with 1s up till number of keys, e.g. two keys would be [1,1,1,0,0,0]). So [1,0,0,0,0,0] means zero keys.

Vector Observation contents

Over 1 year ago

Apologies if this sounds bit naive, but make sure you are importing the correct Python libraries. E.g. do you have obstacle_tower_env.py file in the directory from which you run the experiments?

You can find the imported file with the following:

In [1]: import obstacle_tower_env                                                                                                                           

In [2]: obstacle_tower_env.__file__                                                                                                                         
Out[2]: '/home/USERNAME/.local/lib/python3.5/site-packages/obstacle_tower_env.py'

V2.0 performance drop

Over 1 year ago

If it is any help, we do the reset like this:

config = {"total-floors": 25}
env.reset(config)

Granted, we have not tried changing configs mid-run.

On the topic of configs: Some of the variables did not seem to have an effect of game (visual-theme and allowed-rooms), but this is only based on very brief experimentation so might be just a brainfart in our end.

V2.0 performance drop

Over 1 year ago

I haven’t measured FPS of the v2.0 env yet, but noticed two things:

  • v2.0 is prettier: It has glows and fancier key models/textures. Perhaps this affects the performance little bit?
  • reset() takes much longer now with default config. Set total-floors reset parameter to 25 or lower to speed up reset considerably.

Vector Observation contents

Over 1 year ago

Judging by this function in obstacle_tower_env, the first six elements are for the number of keys (Perhaps one-hot? Not sure). Seventh is the time remaining and eight is the floor number, as in v1.3.

Question about the “Round 1 of Unity’s Obstacle Tower Challenge” email

Over 1 year ago

@TruthMaker

Tax stuff: Ah yes, sounds reasonable. Would be bit of a mess if one of the Round 2 top3 would not be able to get the prize (assuming they want it).

On 4b): All of our team members received the email asking for ID and signature on the rules, and team lead got email with different greeting (“Dear Team Leader…”). I imagine they need everyone to sign the rules.

Question about the “Round 1 of Unity’s Obstacle Tower Challenge” email

Over 1 year ago

I second the post by @ipv6 , especially the parts (1,3) and (4):

(1,3) GCP credit might be tax-able in our country, so indeed it’d be nice to move to Round 2 without the GCP credits (and without the tax form).
(4) We in our do not have address in our standard IDs, although with more time we might be able to get such IDs.

anssi has not provided any information yet.