Data Synthesis

If you like this presentation – show it...

Slide 0

©2013 Primal Fusion Inc. Data Synthesis The Big Problem with Small Data

Slide 1

Treat your customers as individuals. MASS MARKETS OF INDIVIDUALS Media and advertising Healthcare and medicine Education Ecommerce and marketing Etc.

Slide 2

Implicit Semantics Statistical Approaches Explicit Semantics Ontological Approaches The Long Tail of Big Data EXPRESSIVENESS DATA COMPLEX SCHEMA SIMPLE SCHEMA SMALL BIG Cost-Performance Barrier MASS MARKETS OF INDIVIDUALS Statistical methods lose significance Ontological methods prohibitively expensive Hybrid Approaches

Slide 3

Example: Expertise Search Source: James Cridland

Slide 4

Statistical Approaches

Slide 5

Manual Approaches

Slide 6

Slide 7

Slide 8

Slide 9

Primal’s approach: Modeling knowledge generation, not modeling knowledge Natural Language Primal Semantics Words + Grammatical rules = Statements and queries Atomic semantics + Constructive rules = Semantic representations

Slide 10

Slide 11

Slide 12

Slide 13

Slide 14

Slide 15

Treat your customers as individuals. Massive opportunities in truly individualized services, but... …huge challenges in the long tail of big data. The cost-performance barrier requires solutions with fundamentally different cost structures. Primal’s semantic synthesis technology is one such solution.

Slide 16

About Primal Primal powers the rapid development of personalized and intelligent systems. Cloud-based data service (DaaS). Software and IP licensing opportunities are available for larger companies. Professional services available, with expertise in knowledge representation, statistical computing, information retrieval and extraction, database, and cloud computing. More info: primal.com

Slide 17

Contact Info Peter Sweeney, Founder & President peter.sweeney@primal.com @petersweeney Further reading: blog.primal.com