What is meant by when someone says face recognition on wild dataset and on static dataset?
In the wild refers to uncontrolled imagery, as opposed to images taken in lab environments or posed pictures such as driver's license/passport pictures or mug shots.
static in this context refers to still imagery, as opposed to temporal data (e.g. video).
With regard to facial expression recognition (not sure if you're interested in this related area), you could read this paper which describes a popular static database for images in the wild, which has good background information: Static Facial Expression Analysis in Tough Conditions: Data, Evaluation Protocol and Benchmark