# Extreme State Aggregation beyond MDPs

@article{Hutter2014ExtremeSA, title={Extreme State Aggregation beyond MDPs}, author={Marcus Hutter}, journal={ArXiv}, year={2014}, volume={abs/1407.3341} }

We consider a Reinforcement Learning setup without any (esp. MDP) assumptions on the environment. State aggregation and more generally feature reinforcement learning is concerned with mapping histories/raw-states to reduced/aggregated states. The idea behind both is that the resulting reduced process (approximately) forms a small stationary finite-state MDP, which can then be efficiently solved or learnt. We considerably generalize existing aggregation results by showing that even if the… CONTINUE READING

#### Topics from this paper.

16 Citations

Information-Theoretic Considerations in Batch Reinforcement Learning

- Computer Science, Mathematics
- 2019

44

On overfitting and asymptotic bias in batch reinforcement learning with partial observability

- Mathematics, Computer Science
- 2019

7

#### References

##### Publications referenced by this paper.

SHOWING 1-10 OF 31 REFERENCES

Optimal Regret Bounds for Selecting the State Representation in Reinforcement Learning

- Computer Science, Mathematics
- 2013

25

Equivalence notions and model minimization in Markov decision processes

- Computer Science, Mathematics
- 2003

262