Baysian perceptions


There is more from a podcast interview with Chris Frith (here). This time I quote his views on Baysian perception.

…Perception is a two-way process. This is why I talk about Reverend Thomas Bayes, who produced this formula two hundred years ago. What he’s essentially pointing out is that our perception of the world depends on two things: that is to say, the sensory information that’s coming in through our eyes and ears, and our prior expectations and our knowledge of the world. And it’s the balance of these two that creates what we experience.

His formula tells you how much do you have to change your model of the world given the new evidence that’s coming in. So if you have very strong expectations, that will affect what you actually perceive. In a sense you can’t perceive things that you don’t know something about already…

And also, people who study how the brain works suggest that the brain is a Baysian system that is concerned with making predictions, and collecting sensory evidence, and then looking at the prediction errors to decide what to do next. And certainly learning about the world these days is very much conceived in terms of a Baysian process where you predict what’s going to happen and then you adjust your learning on the basis of these prediction errors.

3 thoughts on “Baysian perceptions

  1. I loved the main idea, but every learning algorithm works like that, not only bayesian (and if it is the naive version it is even departing from the wrong assumption that is of the independence of the elements that are being percepted) neural networks, decision trees, support vector machines, etc …
    So why bayesian?

    JanetK: Mariana, the next post is also on Firth’s ideas and is less Bayesian. I am of two minds on this: like the basic Bayesian notion but (like you maybe) I am wary of all the cultish hype around the subject.

  2. Sorry janet I tend to respond so late cause this blog does not have a way (at least that I know) to let me know someone responded to the comment. Just wanted to tell you that I just hate the Bayesian hype.
    Dough its true, many advantages, Quick algorithm, easy to understand, works very well in several cases, for example for nlp, even dough it is not understood why it works well compared to other algorithms that seem more reasonable to be used for that kind of tasks. Besides the point that bayes assumes the independence of the variables, but in real life its used with dependent variables most of the time.

  3. <p>But we know that the processing that the brain does is non-algorithmic. .</p>
    <p>http://www.ncbi.nlm.nih.gov/pubmed/16453069</p&gt;
    <p>in the sense that the processing can be reversed whereas algorithms (in general) cannot be reversed (e.g. squaring and square root). </p>
    <p>Brain processing may simulate Bayesian-type computations, but because they are fundamentally non-algorithmic they sort of can’t really be Bayesian.</p>
    JanetK: I tend to agree daedalus2u. I certainly do not think that the brain is formulating equations and solving them to form perceptions. But I do think that the general
    ideas of the Baysian approach is very similar to what is happening during perception. We do seem to be forecasting the immediate future and correcting our
    perceptions on the basis of mis-match between the prediction and the result.

Leave a Reply

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