Strongly-Typed Agents are Guaranteed to Interact Safely

David Balduzzi
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:332-341, 2017.

Abstract

As artificial agents proliferate, it is becoming increasingly important to ensure that their interactions with one another are well-behaved. In this paper, we formalize a common-sense notion of when algorithms are well-behaved: an algorithm is safe if it does no harm. Motivated by recent progress in deep learning, we focus on the specific case where agents update their actions according to gradient descent. The paper shows that that gradient descent converges to a Nash equilibrium in safe games. The main contribution is to define strongly-typed agents and show they are guaranteed to interact safely, thereby providing sufficient conditions to guarantee safe interactions. A series of examples show that strong-typing generalizes certain key features of convexity, is closely related to blind source separation, and introduces a new perspective on classical multilinear games based on tensor decomposition.

Cite this Paper

BibTeX
@InProceedings{pmlr-v70-balduzzi17a, title = {Strongly-Typed Agents are Guaranteed to Interact Safely}, author = {David Balduzzi}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {332--341}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/balduzzi17a/balduzzi17a.pdf}, url = {https://proceedings.mlr.press/v70/balduzzi17a.html}, abstract = {As artificial agents proliferate, it is becoming increasingly important to ensure that their interactions with one another are well-behaved. In this paper, we formalize a common-sense notion of when algorithms are well-behaved: an algorithm is safe if it does no harm. Motivated by recent progress in deep learning, we focus on the specific case where agents update their actions according to gradient descent. The paper shows that that gradient descent converges to a Nash equilibrium in safe games. The main contribution is to define strongly-typed agents and show they are guaranteed to interact safely, thereby providing sufficient conditions to guarantee safe interactions. A series of examples show that strong-typing generalizes certain key features of convexity, is closely related to blind source separation, and introduces a new perspective on classical multilinear games based on tensor decomposition.} }
Endnote
%0 Conference Paper %T Strongly-Typed Agents are Guaranteed to Interact Safely %A David Balduzzi %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-balduzzi17a %I PMLR %P 332--341 %U https://proceedings.mlr.press/v70/balduzzi17a.html %V 70 %X As artificial agents proliferate, it is becoming increasingly important to ensure that their interactions with one another are well-behaved. In this paper, we formalize a common-sense notion of when algorithms are well-behaved: an algorithm is safe if it does no harm. Motivated by recent progress in deep learning, we focus on the specific case where agents update their actions according to gradient descent. The paper shows that that gradient descent converges to a Nash equilibrium in safe games. The main contribution is to define strongly-typed agents and show they are guaranteed to interact safely, thereby providing sufficient conditions to guarantee safe interactions. A series of examples show that strong-typing generalizes certain key features of convexity, is closely related to blind source separation, and introduces a new perspective on classical multilinear games based on tensor decomposition.
APA
Balduzzi, D.. (2017). Strongly-Typed Agents are Guaranteed to Interact Safely. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:332-341 Available from https://proceedings.mlr.press/v70/balduzzi17a.html.

Related Material