Advances in artificial intelligence: deep learning

Advances in artificial intelligence: deep learning

November 25, 2012 – 12:34 am
If you want to keep up with advances in artificial intelligence, the New York Times has an essentialarticle on a recent step forward called deep learning.
There is a rule of thumb for following how AI is progressing: keep track of what Geoffrey Hinton is doing.
Much of the current science of artificial neural networks and machine learning stems from his work or work he has done with collaborators.
The New York Times piece riffs on the fact that Hinton and his team just won a competition to design software to help find molecules that are most likely to be good candidates for new drugs.
Hinton’s team entered late, their software didn’t include a big detailed database of prior knowledge, and they easily won by applying deep learning methods.
To understand the advance you need to know a little about how modern AI works.
Most uses abstract statistical representations. For example, a face recognition system will not use human-familiar concepts like ‘mouth’, ‘nose’ and ‘eyes’ but statistical properties derived from the image that may bear no relation to how we talk about faces.
The innovation of deep learning is that it not only arranges these properties into hierarchies – with properties and sub-properties – but it works out how many levels of hierarchy best fit the data.
If you’re a machine learning aficionado Hinton described how they won the competition in a recent interview but he also puts all his scientific papersonline if you want the bare metal of the science.
Either way, while the NYT piece doesn’t go into how the new approach works, it nicely captures it’s implications for how AI is being applied.
And as many net applications now rely on communication with the cloud – think Siri or Google Maps – advances in artificial intelligence very quickly have an impact on our day-to-day tools.
 
Link to NYT on deep learning AI (via @hpashler)

Comments

Post a Comment

Popular posts from this blog

Orifice and Weir flow calculations

Note: Orifice and Weir Flow Computations The orifice flow calculation proceeds as follows: 1. Initially and whenever the setting (i.e., the fraction opened) changes, flow coefficients for both orifice and weir behavior are computed as follows: a. For side orifices: Define Hcrit = h/2 where h is the opening height. b. For bottom orifices: i. For a circular orifice, compute area over length (i.e., circumference) as AL = h /4. ii. For a rectangular orifice compute AL = h*w/(2*(h+w)) where w is the opening width. iii. Compute Hcrit = Cd*AL/0.414 where Cd is the orifice discharge coefficient. At step 1b, the critical head for the bottom orifice, where orifice flow turns into weir flow, is found by equating the result of the orifice equation to that of the weir equation Cd*Area*sqrt(2g)*sqrt(Hcrit) = Cw*Length*sqrt(Hcrit)*Hcrit or Hcrit = (Cd * Area) / (Cw/sqrt(2g) * Length) The value of Cw/sqrt(2g) for a sharp crested weir is 0.414.
Read more

SWMM 5 Error Messages

Engine Error Number Description ERROR 101: memory allocation error.                                                                                             ERROR 103: cannot solve KW equations for Link                                                                                   ERROR 105: cannot open ODE solver.                                                                                              ERROR 107: cannot compute a valid time step.                                                                                     ERROR 108: ambiguous outlet ID name for Subcatchment                                                                            ERROR 109: invalid parameter values for Aquifer                                                                                  ERROR 110: ground elevation is below water table for Subcatchment                                                         
Read more

EMC Washoff in SWMM5

Image
There are four steps to using EMC concentrations in your network: 1. Define your pollutant by adding a pollutant using the Data=>Quality=>Pollutant command : 2. Define the Land Use by using the Data=>Land Uses command or the Land Use Editor: 3. Define Buildup to be None by clicking on the None Tab: 4. Define the EMC Washoff concentration by clicking on the Washoff Tab: This is where you would add a GW concentration More: http://www.swmm2000.com/
Read more