[FWD] Two stories from a research paper: Content Without Context is Meaningles

By | November 10, 2014

Two stories from a research paper: Content Without Context is Meaningless.

1.1 Machine Learning Hammer

Mark Twain once said: “To a man with a hammer, everything looks like a nail.” His observation is definitely very relevant to current trends in content analysis. We have a Machine Learning Hammer (ML Hammer) that we want to use for solving any problem that needs to be solved. The problem is neither with learning nor with the hammer; the problem is with people who fail to learn that not every problem is a new learning problem [1]. … If we can identify such a feature set, then we can easily model each object by its appropriate feature values. The challenges are

  • to identify a right set of features
  • to identify feature values for representing each object

In reality, both problems are related. … Unfortunately, we ignore the first problem and use our ML Hammer on whatever problem we are given. Surprisingly, we are happy even when we get (in most cases) 20%-30% accuracy in the results (the average precision of object detection in Pascal Challenge 2009 is in this range [11]).

1.2 Solving The Right Problems

Let us paraphrase a famous story in the context of this paper. A drunk multimedia researcher loses the keys to his house and is looking for them under a lamppost. Another researcher comes over and asks what he‘s doing. “I’m looking for my keys” he says. “Let me help you”, says the new researcher and joins the effort. Soon there are many researchers looking for the keys. One of them got frustrated and asks: “where did you lose your keys”. The original researcher replies, “I lost them over there”, and points to a dark corner in the street. The new scientist looks puzzled. “Then why are you looking for them all the way over here? ”, he asks. “Because the light is so much better here. We can formulate and solve the search problem much better here. Over there it is not easy to formulate because you can not see well”, replies the original re- searcher. Finding the explanation reasonable, all researchers keep looking for the keys under the lamppost. After long rigorous and exhaustive efforts they conclude that the problem of finding lost keys is an unsolvable problem.

A famous real story is related to the milkshake by McDonald‘s [26]. McDonald‘s wanted to make their milkshake as a more effective product. A team of marketing researcher started analyzing standard statistical techniques, to find the taste, thickness, temperature and other basic features of milkshake to find what most people like. One researcher decided to ignore the features and study why people buy milkshake. The findings were startling. People bought milkshake not for taste but for giving them company over long drives without being messy to consume and being a good companion for long periods. In content analysis also, one needs to really understand why a particular media source is used and what need does this really address.

Leave a Reply

Your email address will not be published. Required fields are marked *