Anthony Bak, Principal Data Scientist at Ayasdi @ MLconf SEA
summary
Topology (Ayasdi style TDA) builds a combinatorial summary by taking a real-valued lens, forming an overlapping interval cover, clustering each inverse image and linking clusters that share points — an inverse-image clustering pipeline that yields a simplicial complex summary rather than coordinates. It needs only a distance metric (no manifold or smoothness assumptions), finds unexpected shapes (disease courses as circles, not lines), exposes model bias in insurance fraud, and surfaces repeated small cohorts (same flare seen seven times, some clusters with only hundreds of users).
AI-written summary · about
chapters (11)
- 0:00Start
- 0:16Data Complexity
- 2:28Clustering
- 2:56Circles
- 3:37Wide Junction
- 8:51Summary of the Data
- 13:38Properties of Topology
- 20:06Deformation Invariance
- 22:12Compressed Representation
- 23:32Insurance Claim Data
- 25:46Example Predictive Maintenance for Turbines
related talks
26mAnthony Bak, Principal Data Scientist and Mathematician, Ayasdi
20mAllison Gilmore, Data Scientist, Ayasdi
06mComplex Data Analysis
20mAllison Gilmore, Data Scientist, Ayasdi
50mThe Shape of Data: Machine Learning and Topology
51mTopological space creation and Clustering at BigData scale
27mTopological Data Analysis: Information from moving Data
34mTopological Data Analysis