Concurrent Reinforcement Learning from Customer Interactions

David Silver, Leonard Newnham, David Barker, Suzanne Weller, Jason McFall
; Proceedings of the 30th International Conference on Machine Learning, PMLR 28(3):924-932, 2013.

Abstract

In this paper, we explore applications in which a company interacts concurrently with many customers. The company has an objective function, such as maximising revenue, customer satisfaction, or customer loyalty, which depends primarily on the sequence of interactions between company and customer. A key aspect of this setting is that interactions with different customers occur in parallel. As a result, it is imperative to learn online from partial interaction sequences, so that information acquired from one customer is efficiently assimilated and applied in subsequent interactions with other customers. We present the first framework for concurrent reinforcement learning, using a variant of temporal-difference learning to learn efficiently from partial interaction sequences. We evaluate our algorithms in two large-scale test-beds for online and email interaction respectively, generated from a database of 300,000 customer records.

Cite this Paper


BibTeX
@InProceedings{pmlr-v28-silver13, title = {Concurrent Reinforcement Learning from Customer Interactions}, author = {David Silver and Leonard Newnham and David Barker and Suzanne Weller and Jason McFall}, booktitle = {Proceedings of the 30th International Conference on Machine Learning}, pages = {924--932}, year = {2013}, editor = {Sanjoy Dasgupta and David McAllester}, volume = {28}, number = {3}, series = {Proceedings of Machine Learning Research}, address = {Atlanta, Georgia, USA}, month = {17--19 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v28/silver13.pdf}, url = {http://proceedings.mlr.press/v28/silver13.html}, abstract = {In this paper, we explore applications in which a company interacts concurrently with many customers. The company has an objective function, such as maximising revenue, customer satisfaction, or customer loyalty, which depends primarily on the sequence of interactions between company and customer. A key aspect of this setting is that interactions with different customers occur in parallel. As a result, it is imperative to learn online from partial interaction sequences, so that information acquired from one customer is efficiently assimilated and applied in subsequent interactions with other customers. We present the first framework for concurrent reinforcement learning, using a variant of temporal-difference learning to learn efficiently from partial interaction sequences. We evaluate our algorithms in two large-scale test-beds for online and email interaction respectively, generated from a database of 300,000 customer records. } }
Endnote
%0 Conference Paper %T Concurrent Reinforcement Learning from Customer Interactions %A David Silver %A Leonard Newnham %A David Barker %A Suzanne Weller %A Jason McFall %B Proceedings of the 30th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2013 %E Sanjoy Dasgupta %E David McAllester %F pmlr-v28-silver13 %I PMLR %J Proceedings of Machine Learning Research %P 924--932 %U http://proceedings.mlr.press %V 28 %N 3 %W PMLR %X In this paper, we explore applications in which a company interacts concurrently with many customers. The company has an objective function, such as maximising revenue, customer satisfaction, or customer loyalty, which depends primarily on the sequence of interactions between company and customer. A key aspect of this setting is that interactions with different customers occur in parallel. As a result, it is imperative to learn online from partial interaction sequences, so that information acquired from one customer is efficiently assimilated and applied in subsequent interactions with other customers. We present the first framework for concurrent reinforcement learning, using a variant of temporal-difference learning to learn efficiently from partial interaction sequences. We evaluate our algorithms in two large-scale test-beds for online and email interaction respectively, generated from a database of 300,000 customer records.
RIS
TY - CPAPER TI - Concurrent Reinforcement Learning from Customer Interactions AU - David Silver AU - Leonard Newnham AU - David Barker AU - Suzanne Weller AU - Jason McFall BT - Proceedings of the 30th International Conference on Machine Learning PY - 2013/02/13 DA - 2013/02/13 ED - Sanjoy Dasgupta ED - David McAllester ID - pmlr-v28-silver13 PB - PMLR SP - 924 DP - PMLR EP - 932 L1 - http://proceedings.mlr.press/v28/silver13.pdf UR - http://proceedings.mlr.press/v28/silver13.html AB - In this paper, we explore applications in which a company interacts concurrently with many customers. The company has an objective function, such as maximising revenue, customer satisfaction, or customer loyalty, which depends primarily on the sequence of interactions between company and customer. A key aspect of this setting is that interactions with different customers occur in parallel. As a result, it is imperative to learn online from partial interaction sequences, so that information acquired from one customer is efficiently assimilated and applied in subsequent interactions with other customers. We present the first framework for concurrent reinforcement learning, using a variant of temporal-difference learning to learn efficiently from partial interaction sequences. We evaluate our algorithms in two large-scale test-beds for online and email interaction respectively, generated from a database of 300,000 customer records. ER -
APA
Silver, D., Newnham, L., Barker, D., Weller, S. & McFall, J.. (2013). Concurrent Reinforcement Learning from Customer Interactions. Proceedings of the 30th International Conference on Machine Learning, in PMLR 28(3):924-932

Related Material