Similarity search on neighbor's graphs with automatic Pareto optimal performance and minimum expected quality setups based on hyperparameter optimization
This work addresses the challenge of optimizing similarity search performance for users in data retrieval applications, but it appears incremental as it builds on existing neighbor graph and metaheuristic techniques.
The authors tackled the problem of similarity search by introducing an autotuned algorithm that uses neighbor graphs and optimization metaheuristics to automatically achieve Pareto-optimal trade-offs between search quality and speed, as well as indexes with minimum quality, and benchmarked it against state-of-the-art methods to show convenience and competitiveness.
This manuscript introduces an autotuned algorithm for searching nearest neighbors based on neighbor graphs and optimization metaheuristics to produce Pareto-optimal searches for quality and search speed automatically; the same strategy is also used to produce indexes that achieve a minimum quality. Our approach is described and benchmarked with other state-of-the-art similarity search methods, showing convenience and competitiveness.