I am posting on my blog (and not www.DarwinEco.com) because it is too technical to post on our business-driven site.
An aquintance with expertise in Knowledge Management asked me about how Darwin compares to DLA. I felt that the question was important enough to share part of my answer with my followers. I believe that it brings a deeper perspective about my thinking regarding Darwin's organic temporal curation as way to push the inference of meaning to the user.
Here is part of my explanation to the LDA question:
"DLA and it is quite different from our model. We use Organic Temporal Correlation complemented by a visualization model that facilitates the detection of patterns to obtain rapid cognitive inference of meaning from the user. Basically, we make it easy to observe and qualify patterns and we do not use any semantic analysis. Hence, the reference to revealing order from the chaotic system that is the Web in our case. DLA is different from that it uses probability density function and that Dirichlet distributions are mostly used as pre-distributions in Bayesian inference. It makes it somewhat similar to probabilistic latent semantic analysis except that the Dirichlet distribution is a prerequisite in DLA. The advantage is that as a generative model, it offers a capability for modeling data or a process for constructing a conditional probability de nsity function. Although we do not have an immediate need for DLA, I can anticipate that we will eventually incorporate it to achieve the latter.