Joelle Pineau, Ruo Tao, Vincent Franc ¸ois-Lavet
We lifted 3 functions out of this paper's own repositories and ran 3 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| taodav/nsrs | canonical | 3 of 3 |
| Function | Status | Where it lives |
|---|---|---|
| exp_dec_error_pytorch | Ran | taodav/nsrs/nsrl/learning_algos/NSRS_pytorch.py code served (permissive licence) · get_code("cf6acae38c9e50c2") |
| exp_dec_error_pytorch_2 | Ran | taodav/nsrs/nsrl/learning_algos/NSRS_pytorch.py code served (permissive licence) · get_code("95bf653972457526") |
| mean_squared_error_p_pytorch | Ran | taodav/nsrs/nsrl/learning_algos/NSRS_pytorch.py code served (permissive licence) · get_code("789a6111fa92576e") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
We present a new approach for efficient exploration which leverages a lowdimensional encoding of the environment learned with a combination of modelbased and model-free objectives. Our approach uses intrinsic rewards that are based on the distance of nearest neighbors in the low dimensional representational space to gauge novelty. We then leverage these intrinsic rewards for sampleefficient exploration with planning routines in representational space for hard exploration tasks with sparse rewards. One key element of our approach is the use of information theoretic principles to shape our representations in a way so that our novelty reward goes beyond pixel similarity. We test our approach on a number of maze tasks, as well as a control problem and show that our exploration approach is more sample-efficient compared to strong baselines.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2009.13579")
get_code_for_paper("2009.13579")
have("2009.13579")
Connect an agent — have() is free.