Bayesian recognition in baby similarity

When people come to see small babies, it’s almost like they’re obliged to offer their opinions on who the child looks like. Most of the time it’s an immediate ancestor – either a parent or grandparent. Sometimes it could be a cousin or aunt or uncle as well. Thankfully it’s uncommon to compare babies’ looks to those who they don’t share genes with.

So as people have come up and offered their opinions on who our daughter looks like (I’m top seed, I must mention), I’ve been trying to analyse how they come up with their predictions. And as I observe the connections between people making the observations, and who they mention, I realise that this too follows some kind of Bayesian Recognition.

Basically different people who come to see the baby have different amounts of information on how each of the baby’s ancestors looked like. A recent friend of mine, for example, will only know how my wife and I look. An older friend might have some idea of how my parents looked. A relative might have a better judgment of how one of my parents looked than how I looked.

So based on their experiences in recognising different people in and around the baby’s immediate ancestry, they effectively start with a prior distribution of who the baby looks like. And then when they see the baby, they update their priors, and then mention the person with the highest posterior probability of matching the baby’s face and features.

Given that posterior probability is a function of prior probability, there is no surprise that different people will disagree on who the baby looks like. After all, each of their private knowledge of the baby’s ancestry’s idiosyncratic faces, and thus their priors, will be different!

Unrelated, but staying on Bayesian reasoning, I recently read this fairly stud piece in Aeon on why stereotyping is not necessarily a bad thing. The article argues that in the absence of further information, stereotypes help us form a good first prior, and that stereotypes only become a problem if we fail to update our priors with any additional information we get.

Bayesian Recognition and the Inverse Charlie Chaplin Principle

So I bumped into Deepa at a coffee shop this evening. And she almost refused to recognise me. It turned out to be a case of Bayesian Recognition having gone wrong. And then followed in quick succession by a case of Inverse Charlie Chaplin Principle.

So I was sitting at this coffee shop in Jayanagar meeting an old acquaintance, and Deepa walked in, along with a couple of other people. It took me a while to recognise her, but presently I did, and it turned out that by then she was seated at a table such that we were directly facing each other, with some thirty feet between us (by now I was positive it was her).

I looked at her for a bit, waiting for her to recognise me. She didn’t. I got doubts on whether it was her, and almost took out my phone to message and ask her if it was indeed her. But then I decided it was a silly thing to do, and I should go for it the natural way. So I looked at her again, and looked at her for so long that if she were a stranger she would have thought I was leching at her (so you know that I was quite confident now that it was indeed Deepa). No response.

I started waving, with both arms. She was now looking at me, but past me. I continued waving, and I don’t know what my old acquaintance who I was talking to was thinking by now. And finally a wave back. And we got both got up, and walked towards each other, and started talking.

The Charlie Chaplin principle comes from this scene in a Charlie Chaplin movie which I can’t remember right now where he is standing in front of a statue of the king. Everyone who goes past him salutes him, and he feels high that everyone is saluting him, while everyone in effect is saluting the statue of the king behind him.

Thus, the “Charlie Chaplin Principle” refers to the case where you think someone is smiling at you or waving at you or saluting you, and it turns out that they are doing that to someone who is collinear with you and them. Thus, you are like Charlie Chaplin, stupidly feeling happy about this person smiling/waving/saluting at you while it is someone else that they are addressing.

Like all good principles, this one too has an inverse – which we shall call the “Inverse Charlie Chaplin Principle”. In this one, someone is smiling or waving or blowing kisses at you, and you assume that the gesture is intended to someone else who is collinear with the two of you. Thus, you take no notice of the smile or wave or blown kiss, and get on with life, with the likelihood that you are pissing off the person who is smiling or waving or blowing kisses at you!

Both these effects have happened to me a few times, and I’ve been on both sides of both effects. And an instance of the Inverse principle happened today.

Deepa claimed that she initially failed to recognise me because she assumed that I’m in Spain, and that thus there’s no chance I would be in Jayanagar this evening (clearly she reads this blog, but not so regularly!). Thus, she eliminated me from her search space and was unable to fit my face to anyone else she knows.

Then when I started waving, the Inverse Charlie Chaplin Principle took over. Bizarrely (there was no one between us in the cafe save the acquaintance I was talking to, and I wouldn’t be waving wildly at someone at the same table for two as me; and Deepa was sitting with her back to the wall of the cafe so I presumably could not have been waving at anyone else behind her), she assumed that I was waving to someone else (or so that was her claim), and that it took time for her to realise that it was her that I was actually waving to!

Considering how Bayesian Recognition can throw you off, I’m prepared to forgive her. But I didn’t imagine that Bayesian Recognition would throw her off so much that it would cause an Inverse Charlie Chaplin Effect on her!

Oh, I must mention that I have grown a stubble (the razor I took on my trip to Europe was no good), and that she mentioned about not wearing her glasses today. Whatever!