# Learning to Route in Similarity Graphs

@inproceedings{Baranchuk2019LearningTR, title={Learning to Route in Similarity Graphs}, author={Dmitry Baranchuk and Dmitry Persiyanov and Anton Sinitsin and Artem Babenko}, booktitle={ICML}, year={2019} }

Recently similarity graphs became the leading paradigm for efficient nearest neighbor search, outperforming traditional tree-based and LSH-based methods. Similarity graphs perform the search via greedy routing: a query traverses the graph and in each vertex moves to the adjacent vertex that is the closest to this query. In practice, similarity graphs are often susceptible to local minima, when queries do not reach its nearest neighbors, getting stuck in suboptimal vertices. In this paper we…

## 17 Citations

Two-stage routing with optimized guided search and greedy algorithm on proximity graph

- Computer ScienceKnowl. Based Syst.
- 2021

Learning to Prune: General and Efficient Approximate Nearest Neighbor Search with Direction Navigating Graph

- Computer Science2022 the 5th International Conference on Data Storage and Data Engineering (DSDE)
- 2022

This paper proposes the concept of Direction Navigating Graph (DNG) to offer an online pruning strategy combined with a heuristic search algorithm and develops a neural network that project queries and vertices into a low-dimensional direction vector space.

Graph-based Nearest Neighbor Search: From Practice to Theory

- Computer ScienceICML
- 2020

This work rigorously analyze the performance of graph-based NNS algorithms, specifically focusing on the low-dimensional (d << \log n) regime, and analyzes the most successful heuristics commonly used in practice.

HVS: Hierarchical Graph Structure Based on Voronoi Diagrams for Solving Approximate Nearest Neighbor Search

- Computer ScienceProc. VLDB Endow.
- 2021

This work proposes a novel graph structure called HVS, a hierarchical structure of multiple layers that corresponds to a series of subspace divisions in a coarse-to-fine manner that can reach the nearest neighbors of a given query efficiently, resulting in a reduction in the total search cost.

Reinforcement Routing on Proximity Graph for Efficient Recommendation

- Computer ScienceACM Transactions on Information Systems
- 2022

A reinforcement model is proposed to train an agent to search on the proximity graph automatically for MIPS problem if the authors lack the ground truths of training queries, and this model can also utilize these ground truths by imitation learning to improve the agent’s search ability.

LAN: Learning-based Approximate k-Nearest Neighbor Search in Graph Databases

- Computer Science2022 IEEE 38th International Conference on Data Engineering (ICDE)
- 2022

This paper proposes a learning-based k-ANN search method to reduce NDC and proposes a compressed GNN-graph to accelerate the neighbor ranking model and the initial node selection model, and proves that learning efficiency is improved without degrading the accuracy.

A Note on Graph-Based Nearest Neighbor Search

- Computer ScienceArXiv
- 2020

This paper observed that high clustering coefficient makes most of the k nearest neighbors of q sit in a maximum strongly connected component (SCC) in the graph, and proved that the commonly used graph-based search algorithm is guaranteed to traverse the maximum SCC once visiting any point in it.

A Comprehensive Survey and Experimental Comparison of Graph-Based Approximate Nearest Neighbor Search

- Computer ScienceProc. VLDB Endow.
- 2021

This study provides a thorough comparative analysis and experimental evaluation of 13 representative graph-based ANNS algorithms via a new taxonomy and fine-grained pipeline, and designs an optimized method that outperforms the state-of-the-art algorithms.

Supervised Learning Approach to Approximate Nearest Neighbor Search

- Computer ScienceArXiv
- 2019

This work shows that approximate nearest neighbor search can be framed as a classification problem and solved by training a suitable multi-label classifier and using it as an index, which enables adapting an index to the query distribution when thequery distribution and the corpus distribution differ.

Speed-ANN: Low-Latency and High-Accuracy Nearest Neighbor Search via Intra-Query Parallelism

- Computer ScienceArXiv
- 2022

An in-depth examination of the challenges of the state-of-the-art similarity search algorithms, revealing its challenges in leveraging multi-core processors to speed up the search efficiency and proposing Speed-ANN, a parallel similarity search algorithm that exploits hidden intra-query parallelism and memory hierarchy that allows similarity search to take advantage of multiple CPU cores to significantly accelerate search speed while achieving high accuracy.

## References

SHOWING 1-10 OF 40 REFERENCES

Fast Approximate Nearest Neighbor Search With The Navigating Spreading-out Graph

- Computer ScienceProc. VLDB Endow.
- 2019

A novel graph structure called Monotonic Relative Neighborhood Graph (MRNG) is proposed which guarantees very low search complexity (close to logarithmic time) and is proposed to further lower the indexing complexity and make it practical for billion-node ANNS problems.

Efficient and Robust Approximate Nearest Neighbor Search Using Hierarchical Navigable Small World Graphs

- Computer ScienceIEEE Transactions on Pattern Analysis and Machine Intelligence
- 2020

The proposed general metric space search index is able to strongly outperform previous opensource state-of-the-art vector-only approaches and similarity of the algorithm to the skip list structure allows straightforward balanced distributed implementation.

EFANNA : An Extremely Fast Approximate Nearest Neighbor Search Algorithm Based on kNN Graph

- Computer ScienceArXiv
- 2016

EFANNA is the fastest algorithm so far both on approximate nearest neighbor graph construction and approximate nearest neighbour search and Efanna nicely combines the advantages of hierarchical structure based methods and nearest-neighbor-graph based methods.

Refinements to nearest-neighbor searching ink-dimensional trees

- Computer ScienceAlgorithmica
- 2005

This note presents a simplification and generalization of an algorithm for searchingk-dimensional trees for nearest neighbors reported by Friedmanet al [3], which can be generalized to allow a partition plane to have an arbitrary orientation, rather than insisting that it be perpendicular to a coordinate axis, as in the original algorithm.

Searching in metric spaces by spatial approximation

- Computer ScienceThe VLDB Journal
- 2002

This work proposes a new data structure, called sa-tree (“spatial approximation tree”), which is based on approaching the searched objects spatially, that is, getting closer and closer to them, rather than the classic divide-and-conquer approach of other data structures.

Locality-sensitive hashing scheme based on p-stable distributions

- Computer ScienceSCG '04
- 2004

A novel Locality-Sensitive Hashing scheme for the Approximate Nearest Neighbor Problem under lp norm, based on p-stable distributions that improves the running time of the earlier algorithm and yields the first known provably efficient approximate NN algorithm for the case p<1.

Practical and Optimal LSH for Angular Distance

- Computer ScienceNIPS
- 2015

This work shows the existence of a Locality-Sensitive Hashing (LSH) family for the angular distance that yields an approximate Near Neighbor Search algorithm with the asymptotically optimal running time exponent and establishes a fine-grained lower bound for the quality of any LSH family for angular distance.

Approximate nearest neighbors: towards removing the curse of dimensionality

- Computer ScienceSTOC '98
- 1998

Two algorithms for the approximate nearest neighbor problem in high-dimensional spaces are presented, which require space that is only polynomial in n and d, while achieving query times that are sub-linear inn and polynometric in d.

Peer to Peer Multidimensional Overlays: Approximating Complex Structures

- Computer ScienceOPODIS
- 2007

It is argued that an approximation of a complex structure is enough to provide a native support of range queries and this stems from the fact that neighbours are important while the exact space partitioning associated to a given peer is not as crucial.

Product Quantization for Nearest Neighbor Search

- Computer ScienceIEEE Transactions on Pattern Analysis and Machine Intelligence
- 2011

This paper introduces a product quantization-based approach for approximate nearest neighbor search. The idea is to decompose the space into a Cartesian product of low-dimensional subspaces and to…