We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Systems and Control

Title: Computation Resource Allocation Solution in Recommender Systems

Abstract: Recommender systems rely heavily on increasing computation resources to improve their business goal. By deploying computation-intensive models and algorithms, these systems are able to inference user interests and exhibit certain ads or commodities from the candidate set to maximize their business goals. However, such systems are facing two challenges in achieving their goals. On the one hand, facing massive online requests, computation-intensive models and algorithms are pushing their computation resources to the limit. On the other hand, the response time of these systems is strictly limited to a short period, e.g. 300 milliseconds in our real system, which is also being exhausted by the increasingly complex models and algorithms.
In this paper, we propose the computation resource allocation solution (CRAS) that maximizes the business goal with limited computation resources and response time. We comprehensively illustrate the problem and formulate such a problem as an optimization problem with multiple constraints, which could be broken down into independent sub-problems. To solve the sub-problems, we propose the revenue function to facilitate the theoretical analysis, and obtain the optimal computation resource allocation strategy. To address the applicability issues, we devise the feedback control system to help our strategy constantly adapt to the changing online environment. The effectiveness of our method is verified by extensive experiments based on the real dataset from Taobao.com. We also deploy our method in the display advertising system of Alibaba. The online results show that our computation resource allocation solution achieves significant business goal improvement without any increment of computation cost, which demonstrates the efficacy of our method in real industrial practice.
Subjects: Systems and Control (eess.SY); Information Retrieval (cs.IR); Machine Learning (cs.LG)
Cite as: arXiv:2103.02259 [eess.SY]
  (or arXiv:2103.02259v1 [eess.SY] for this version)

Submission history

From: Xun Yang [view email]
[v1] Wed, 3 Mar 2021 08:41:43 GMT (1742kb,D)

Link back to: arXiv, form interface, contact.