Ninh Pham's homepage

I was a postdoctoral researcher at DIKU, University of Copenhagen, working with Stephen Alstrup in the DABAI project, aiming at designing efficient algorithms for machine learning and using big data for digital learning support. I was also a postdoctoral researcher in the Algorithms Group, ITU, involved in the SSS project, investigating efficient algorithms for high-dimensional similarity search on big data. I received my PhD at ITU under the supervision of Rasmus Pagh in 2014. My PhD project, part of the MaDaMS project, focused on efficient randomized algorithms for big data analytics. 

I received the best paper awards at WWW Conference 2014 and ECML-PKDD 2020. In 2022, aminer.cn recognized me as the 2022 AI 2000 Most Influential Scholar Honorable Mention in Data Mining (Rising Star) for my outstanding and vibrant contributions to this field between 2012 and 2021. 

Current Project: 

“Federated Nearest Neighbour Search: Theory and Practice” 2023 – 2026 (Marsden Fast-Start 2022) (AI: Assoc. Prof. F. Silvestri)

Ph.D. students: 

• Jingrui Zhang (2020): Scalable and Interpretable Anomaly Detection (co-supervisor: Gill Dobbie)

• Linghan Zeng (2023): Nearest Neighbor Search under Privacy Constraints (co-supervisor: Francesco Silvestri, Sebastian Link)

• Junyan Zhong (2024): Robust Graph-Based Recommendation Systems (co-supervisor: Jingfeng Zhang, Andre Nies)

Publications:

1. On Deploying Mobile Deep Learning to Segment COVID-19 RT-PCR Test Tube Images

   • Ting Xiang, Richard Dean, Jiawei Zhao, and Ninh Pham

   • PSIVT 2023

2. A Transductive Forest for Anomaly Detection with Few Labels

   • Jingrui Zhang, Ninh Pham, Gillian Dobbie

   • ECML-PKDD 2023 (199/830 ~ 24%), Python Code

3. Falconn++: A Locality-sensitive Filtering Approach for Approximate Nearest Neighbor Search

   • Ninh Pham, Tao Liu

   • NeurIPS 2022 (Spotlight - Top 5% of 10,411 submissions), Online talk at MIT, C++ Code, 5-min teaser at NeurIPS

   • This work is an extension of Liu's master thesis, Data Science Showcase 2023

4. Simple Yet Efficient Algorithms for Maximum Inner Product Search via Extreme Order Statistics 

   • Ninh Pham

   • KDD 2021 (238/1,542 ~ 15%), Supplementary, Slide, C++ Code

5. Revisiting Wedge Sampling for Budgeted Maximum Inner Product Search

   • Stephan Lorenzen, Ninh Pham

   • ECML-PKDD 2020 (131/687 ~ 19%), Best Paper Award, Slide, C++ Code, Datasets

   • IJCAI 2021 Sister Conference Best Paper Track (invited extended abstract) 

6. L1-Depth Revisited: A Robust Angle-based Outlier Factor in High-dimensional Space 

   • Ninh Pham

   • ECML-PKDD 2018 (131/535 ~ 24%), Slide, Supplementary, Python Code, C++ Code, Datasets

7. Hybrid LSH: Faster Near Neighbors Reporting in High-dimensional Space

   • Ninh Pham

   • EDBT 2017 (short paper), Python Code

8. Scalability and Total Recall with Fast CoveringLSH 

   • Ninh Pham, Rasmus Pagh

   • CIKM 2016 (160/701 ~ 23%), Slide, C++ Code, Matlab Code

9. I/O-Efficient Similarity Join 

   • Rasmus Pagh, Ninh Pham, Francesco Silvestri, Morten Stöckel 

   • ESA 2015 (71/261 ~ 27%)

   • Full version appears in Algorithmica 2017 (invited paper)

10. Efficient Estimation for High Similarities using Odd Sketches 

    • Michael Mitzenmacher, Rasmus Pagh, Ninh Pham 

    • WWW 2014 (84/648 ~ 13%), Best Paper Award, Slide, Matlab Code 

    • Media: ACM Computing Reviews, CPH Post, Videnskab, Berlingske

11. Fast and Scalable Polynomial Kernels via Explicit Feature Maps 

    • Ninh Pham, Rasmus Pagh 

    • KDD 2013 (125/726 ~ 17%) (Oral), Slide, Matlab Code (featured in scikit-learn)

12. A Near-linear Time Approximation Algorithm for Angle-based Outlier Detection in High-dimensional Data 

    • Ninh Pham, Rasmus Pagh 

    • KDD 2012 (133/755 ~ 18%), Slide, C++ Code

13. Online Discovery of Top-k Similar Motifs in Time Series Data 

    • Hoang Thanh Lam, Toon Calders, Ninh Pham 

    • SDM 2011 (86/350 ~ 25%), C++ Code

14. Two Novel Adaptive Symbolic Representations for Similarity Search in Time Series Databases 

    • Ninh D. Pham, Quang Loc Le, Tran Khanh Dang 

    • APWeb 2010

15. HOT aSAX: A Novel Adaptive Symbolic Representation for Time Series Discords Discovery 

    • Ninh D. Pham, Quang Loc Le, Tran Khanh Dang 

    • ACIIDS 2010

Others:

On the Power of Randomization in Big Data Analytics 

15. • PhD Thesis 2014, Slide

Teaching:

• Thesis supervision

• Past master/bachelor students

• University of Auckland

  • CS 225: Discrete Structures in Mathematics and Computer Science, 2020 

  • CS 320: Applied Algorithmics, 2019, 2021 - 2024

  • CS 717: Fundamentals of Algorithmics, 2022 - 2024

  • CS 752: Big Data Management, 2019 - 2024

  • CS 753: Algorithms for Massive Data, 2019, 2020

• University of Copenhagen:

  • Large-scale Data Analytics, Spring 2017

  • Project Course on “Authorship verification using textual features”, Fall 2017

• IT University of Copenhagen:

  • Algorithm Design II, 2014, 2015

Services:

• PC: WWW'15 (Poster Track), WWW'16 (Poster Track), ECAI'20, IJCAI'20, WWW'20 (Poster Track), IJCAI'21, IJCAI'22-24 (PC board), ECML-PKDD'23-24

• External reviewer: PKDD'13, ESA'15, STOC'18, KDD'21, PODS'24

• Journal reviewer: TKDE, TIST

Awards:

• Best Paper Awards: WWW'14, ECML-PKDD'20 (Data Mining)

• 2022 AI 2000 Most Influential Scholar Honorable Mention in Data Mining (Rising Star) by aminer.cn

• Student Travel Award KDD'13, KDD'12