Wednesday, November 13, 2013

Data as a Quantifiable Asset

In a dinner conversation, Tony Jebara pointed out that machine learning is growing since companies --- beyond the internet companies in news ---- are increasingly thinking of the data they have as an asset, to be collected, analyzed widely, and used to make decisions; this is a change from the past where companies looked at data collection, storage and management as a cost, and only a few (sales, analysts) looked at data. This made me wonder: will accountants figure out how to price this asset and write it into the books of the companies (like goodwill, brand, and others)?

Labels:

Friday, November 01, 2013

My Columbia U

I realized I have a bimodal interaction with Columbia Univ.
  • I am visiting the Business School, giving a talk in their Financial Engineering Workshop, focused on the risks of real time bidding on ad networks. It is fun to hear financial researchers discussing network effects, cascades and contagion, included. 
  • I will be visiting the Computer Science Dept, giving a talk on data stream theory, more aligning it as an approach to Big Data. Btw, am a Big fan of Shree's BigShot DIY camera.
I already feel at least trimodal in a CS dept, sitting in some convex combination of theory, databases and networking, but that is second order, considering the difference between the suits of the Business school and the jeans of the Engineering school, stripped shirts versus T.