3 ms·
The network in our work is exactly the same as in IMPALA paper. The overall score of our agent is slightly higher, it did better on some envs and it did worse o
by alex_petrenko 6y ago
The network in our work is exactly the same as in IMPALA paper. The overall score of our agent is slightly higher, it did better on some envs and it did worse on some other envs. These lasertag levels are exploration problems, and with a bit of hyperparameter tuning they are not difficult, it's just the agent that learned to do 30 different things somehow sucks on these levels.
- lostmsu 6y agoWere other problems less about exploration? If not, I'd argue lasertag might be more important. In various benchmarks the geometric mean is usually used to compare total score across different tasks to account for severe issues with specific tasks. If your network is the same as IMPALA, why do you show its results with different hyperparameters? Were some of them necessary for the optimization (e.g. reduced batch size)?