Top-k maximum inner product search (MIPS) is a central task in many machine learning applications. This work extends top-k MIPS with a budgeted setting, that asks for the best approximate top-k MIPS given a limited budget of computational operations. We study recent advanced sampling methods, including wedge and diamond sampling, to solve budgeted top-k MIPS. First, we theoretically show that diamond sampling is essentially a combination of wedge sampling and basic sampling for top-k MIPS. Second, we propose dWedge, a simple deterministic variant of wedge sampling for budgeted top-k MIPS. Empirically, dWedge provides significantly higher accuracy than other budgeted top-k MIPS solvers while maintaining a similar speedup.
CITATION STYLE
Lorenzen, S. S., & Pham, N. (2021). Revisiting Wedge Sampling for Budgeted Maximum Inner Product Search (Extended Abstract). In IJCAI International Joint Conference on Artificial Intelligence (pp. 4789–4793). International Joint Conferences on Artificial Intelligence. https://doi.org/10.24963/ijcai.2021/652
Mendeley helps you to discover research relevant for your work.