Tuesday, April 4, 2017

Bernstein: "AI, the data flood and the future of investment research?"

From FT Alphaville:

Snap AV: Quant’s comeback under the cover of AI?
This is from Bernstein’s recent note entitled AI, the data flood and the future of investment research? Important to say up top that they admit that AI focused funds are thin on the ground:
Most asset managers that we speak to on this topic say that they are at an early stage of evaluating its potential. We often encounter asset management companies that are in the process of assessing how to use such techniques, or actively hiring people with the relevant skills or trying to decide how to budget for such a push. Big data is expensive, and the impact of this on its adoption should not be overlooked, especially as the asset management industry tries to cut costs. The vast majority of firms are not even at the stage of actually running analysis on the relevant datasets, they are at an earlier stage of trying to decide what data to get access to, and how they would use it if they did have it.
There are some exceptions to this though. We can identify a small group of funds where they describe AI as being at the forefront of their approach to investment. They have relatively small AUM at present (in the region of USD 1bn) but this presumably will grow. There is then a much larger pool of capital (USD 416bn) run by CTAs where AI appears to inform at least part of the investment approach, but is almost certainly far from being the dominant investment trigger (at least not yet). It probably makes sense that CTAs lead in terms of AUM based on such strategies at present as it is a more natural extension of an already heavily quantitatively-lead approach.
And there is still a tendency to dwell on analyses of parking lots....

So, Where Is High Performance Computing At Right Now? "Data-Hungry Algorithms and the Thirst for AI"