[e] Human Learning; Machine Learning

Driving on Essenhower this morning. I suddenly rised the idea of my own learning process of life. How are people adapting? It's still a black box to me. Yet I do know the machine learning techniques: use training cases to accumulate knowledge; apply it to new, undecided cases; (optionally) adjust the learning effect by feedback.

Now, let me treat myself as a learning machine (LXLearner). It is hard to say that LXLearner is a good one. First, LXLearner don't make judgements fully based on accumulated knowledge, instead, a big portion of arbitory factor is introduced, which makes the output pretty unstable and not converaging; second, LXLearner uses quite outdated and limited training data, makes the decision process inaccurate; third, the robustness is bad: LXLearner is easy to be stressed by faulty and malicious data.

:).