23 Feb 2012


“To survive in the academic world, you need to improve your networking,” my mentor told me.

This is of course very true. Nowadays, good research demands a whole range of expertise that a single team is unlikely to master. Clinical studies also require bigger and bigger sample sizes that can only be achieved by many centers working together. It is therefore important to know people.

But I have a problem – I cannot recognize faces. Recently, I looked at Whitney Houston’s photo and thought that was Jennifer Lopez, you see what I mean.

When I greet friends in overseas conferences, I try my best to hide the fact that I cannot recall where he is from and what his name is. Invariably, disaster soon strikes. Another friend comes and expects me to introduce them to each other. After a few polite remarks, both friends know I cannot recall a thing about either of them.

So, imagine how I felt when I learned that paper wasps (wasps!) were good at recognizing faces.

In a recent paper by Sheehan and Tibbetts, paper wasps were trained to recognize faces (Science 2011;334:1272). The investigators took pictures of the faces of different paper wasps and put them in a maze. When a wasp went in front of the picture of an incorrect wasp, it got an electric shock. It did not take much training before the wasps could choose the correct pictures. On average, they got it right 8 times out of 10.

I learned three important lessons from this study.

First, animal activists can accept insect studies.

Second, it takes only a few hundred neurons to form a neural network for recognizing faces. I am sure I have that.

Third, what I really need is just a few electric shocks.

No comments:

Post a Comment