“LSH Schemes (and how to improve them)”, Bucknell Machine Learning Association (MLA) Talk by Prof. Keegan Kang Monday Sept. 5 at 5.45pm at in the Traditional Reading Room (BERT 213).

The Bucknell MLA will hold its first meeting on Monday 9/5 at at 5.45pm in the Traditional Reading Room (BERT 213), for members to get to know each other and share their passion for machine learning. Prof. Keegan Kang will give an introductory talk on LSH Schemes.

ABSTRACT: There are some challenges with traditional machine learning in a Big Data world. Locality Sensitive Hashing (LSH) schemes are able to mitigate some of these challenges. The idea of LSH schemes will be briefly introduced in this talk by looking at an example of them: sign random projections. This will be followed briefly by an illustration of how LSH schemes can be improved, before concluding with several fun research areas using these schemes.