Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
169 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

HLoOP -- Hyperbolic 2-space Local Outlier Probabilities (2312.03895v1)

Published 6 Dec 2023 in stat.ML and cs.LG

Abstract: Hyperbolic geometry has recently garnered considerable attention in machine learning due to its capacity to embed hierarchical graph structures with low distortions for further downstream processing. This paper introduces a simple framework to detect local outliers for datasets grounded in hyperbolic 2-space referred to as HLoOP (Hyperbolic Local Outlier Probability). Within a Euclidean space, well-known techniques for local outlier detection are based on the Local Outlier Factor (LOF) and its variant, the LoOP (Local Outlier Probability), which incorporates probabilistic concepts to model the outlier level of a data vector. The developed HLoOP combines the idea of finding nearest neighbors, density-based outlier scoring with a probabilistic, statistically oriented approach. Therefore, the method consists in computing the Riemmanian distance of a data point to its nearest neighbors following a Gaussian probability density function expressed in a hyperbolic space. This is achieved by defining a Gaussian cumulative distribution in this space. The HLoOP algorithm is tested on the WordNet dataset yielding promising results. Code and data will be made available on request for reproductibility.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (41)
  1. Bronstein, M. M.; Bruna, J.; LeCun, Y.; Szlam, A.; Vandergheynst, P. Geometric deep learning: going beyond Euclidean data. CoRR 2016, abs/1611.08097.
  2. Anderson, J. W. Hyperbolic geometry; Springer Science & Business Media, 2006.
  3. Nickel, M.; Kiela, D. Poincaré Embeddings for Learning Hierarchical Representations. Advances in Neural Information Processing Systems. 2017.
  4. Sarkar, R. Low distortion Delaunay embedding of trees in hyperbolic plane. In International Symposium on Graph Drawing 2011, 355–366.
  5. Sala, F.; De Sa, C.; Gu, A.; Re, C. Representation Tradeoffs for Hyperbolic Embeddings. Proceedings of the 35th International Conference on Machine Learning. 2018; pp 4460–4469.
  6. Ganea, O.; Bécigneul, G.; Hofmann, T. Hyperbolic entailment cones for learning hierarchical embeddings. International Conference on Machine Learning 2018, 1646–1655.
  7. Liu Q, K. D., Nickel M Hyperbolic graph neural networks. Advances in Neural Information Processing Systems 2019,
  8. Chami, I.; Ying, R.; Ré, C.; Leskovec, J. Hyperbolic Graph Convolutional Neural Networks. Proceedings of the 33rd International Conference on Neural Information Processing Systems 2019,
  9. Dai, J.; Wu, Y.; Gao, Z.; Jia, Y. A hyperbolic-to-hyperbolic graph convolutional network. Computer Vision and Pattern Recognition 2021,
  10. Cetin, E.; Chamberlain, B.; Bronstein, M.; Hunt, J. J. Hyperbolic Deep Reinforcement Learning. arXiv 2022,
  11. Atigh, M. G.; Schoep, J.; Acar, E.; van Noord, N.; Mettes, P. Hyperbolic image segmentation. Computer Vision and Pattern Recognition 2022, 4453–4462.
  12. Gao, Z.; Wu, Y.; Jia, Y.; Harandi, M. Curvature generation in curved spaces for few-shot learning. International Conference on Computer Vision 2021, 8691–8700.
  13. Suris D, V. C., Liu R Learning the predictability of the future. Computer Vision and Pattern Recognition 2021, 12602–12612.
  14. Dengxiong X, K. Y. Generalized open set recognition via hyperbolic side information learning. Winter Conference on Applications of Computer Vision 2023, 3992–4001.
  15. Caron, M.; Touvron, H.; Misra, I.; Jégou, H.; Mairal, J.; Bojanowski, P.; Joulin, A. Emerging properties in self-supervised vision transformers. International Conference on Computer Vision 2021, 9650–9660.
  16. Kleinberg, R. Geographic Routing Using Hyperbolic Space. IEEE INFOCOM 2007 - 26th IEEE International Conference on Computer Communications. 2007; pp 1902–1909.
  17. Cvetkovski, A.; Crovella, M. Hyperbolic Embedding and Routing for Dynamic Graphs. IEEE INFOCOM 2009. 2009; pp 1647–1655.
  18. Cassagnes, C.; Tiendrebeogo, T.; Bromberg, D.; Magoni, D. Overlay addressing and routing system based on hyperbolic geometry. 2011 IEEE Symposium on Computers and Communications (ISCC). 2011; pp 294–301.
  19. Higgott, O.; Breuckmann, N. P. Subsystem Codes with High Thresholds by Gauge Fixing and Reduced Qubit Overhead. Physical Review X 2021, 11.
  20. Higgott, O.; Breuckmann, N. P. Constructions and performance of hyperbolic and semi-hyperbolic Floquet codes. 2023.
  21. Mettes, P.; Atigh, M. G.; Keller-Ressel, M.; Gu, J.; Yeung, S. Hyperbolic Deep Learning in Computer Vision: A Survey. arXiv 2305.06611 2023,
  22. Nielsen, F.; Okamura, K. Information measures and geometry of the hyperbolic exponential families of Poincaré and hyperboloid distributions. 2022.
  23. Alghushairy, O.; Alsini, R.; Soule, T.; Ma, X. A Review of Local Outlier Factor Algorithms for Outlier Detection in Big Data Streams. Big Data Cogn. Comput. 2020, 5.
  24. Souiden, I.; Brahmi, Z.; Toumi, H. A Survey on Outlier Detection in the Context of Stream Mining: Review of Existing Approaches and Recommadations. Intelligent Systems Design and Applications - 16th International Conference on Intelligent Systems Design and Applications 2016,
  25. Wang, S. A Comprehensive Survey of Data Mining-Based Accounting-Fraud Detection Research. 2010 International Conference on Intelligent Computation Technology and Automation. 2010; pp 50–53.
  26. Lin, J.; Keogh, E.; Fu, A.; Van Herle, H. Approximations to magic: finding unusual medical time series. 18th IEEE Symposium on Computer-Based Medical Systems (CBMS’05). 2005; pp 329–334.
  27. Portnoy, L. Intrusion detection with unlabeled data using clustering. 2000.
  28. Phua, C.; Lee, V. C.-S.; Smith-Miles, K.; Gayler, R. W. A Comprehensive Survey of Data Mining-based Fraud Detection Research. ArXiv 2010, abs/1009.6119.
  29. Bolton, R. J.; Hand, D. J. Unsupervised Profiling Methods for Fraud Detection. 2002.
  30. Bansal, R.; Gaur, N.; Singh, S. N. Outlier Detection: Applications and techniques in Data Mining. 2016 6th International Conference - Cloud System and Big Data Engineering (Confluence). 2016; pp 373–377.
  31. Breunig, M. M.; Kriegel, H.-P.; Ng, R. T.; Sander, J. LOF: identifying density-based local outliers. Proceedings of the 2000 ACM SIGMOD international conference on Management of data 2000, 93–104.
  32. Kriegel, H.-P.; Kröger, P.; Schubert, E.; Zimek, A. LoOP: local outlier probabilities. Proceedings of the 18th ACM conference on Information and knowledge management 2009, 1649–1652.
  33. Cho, S.; Lee, J.; Park, J.; Kim, D. A Rotated Hyperbolic Wrapped Normal Distribution for Hierarchical Representation Learning. ArXiv 2022, abs/2205.13371.
  34. Nickel, M.; Kiela, D. Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry. CoRR 2018, abs/1806.03417.
  35. Nagano, Y.; Yamaguchi, S.; Fujita, Y.; Koyama, M. A Wrapped Normal Distribution on Hyperbolic Space for Gradient-Based Learning. International Conference on Machine Learning 2019,
  36. Barbaresco, F. Lie Group Machine Learning and Gibbs Density on Poincaré Unit Disk from Souriau Lie Groups Thermodynamics and SU(1,1) Coadjoint Orbits. Geometric Science of Information. Cham, 2019; pp 157–170.
  37. Mathieu, E.; Lan, C. L.; Maddison, C. J.; Tomioka, R.; Teh, Y. W. Continuous Hierarchical Representations with Poincaré Variational Auto-Encoders. Neural Information Processing Systems. 2019.
  38. Ovinnikov, I. Poincaré Wasserstein Autoencoder. CoRR 2019, abs/1901.01427.
  39. Cho, S.; Lee, J.; Kim, D. GM-VAE: Representation Learning with VAE on Gaussian Manifold. arXiv preprint arXiv:2209.15217 2022,
  40. Petersen, P. Riemannian Geometry; Graduate Texts in Mathematics; Springer New York, 2006.
  41. Nielsen, F.; Nock, R. Hyperbolic Voronoi Diagrams Made Easy. Proceedings of the 2010 International Conference on Computational Science and Its Applications. USA, 2010; pp 74–80.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now