 We are collaborating with Medical Detection Dogs to support the vital work that they carry out which consists of training dogs to detect various human conditions including cancer. They teach the dogs to recognize the order from cancer cells in biological samples and then to signal to a human handler when they find a positive sample. The interface that the dogs use consists of a stand and an arm coming out of the stand which is pivoted on the stand. We use pressure sensors to detect the pressure that the dogs exert when they sniff the samples that sit behind a metal plate. The pressure patterns that the dogs produce can tell us what's in the sample. We are currently developing a kind of second generation of this technology and what we want to do going forward is to start working on machine learning techniques that can automatically interpret these patterns on the fly. The signals that the dogs are trained to use is something like sitting down in front of a positive sample. These are conventional signals that make it very clear to the human that the dog intends to signal but for the dog this mode of signaling it can be disruptive of their detection work so they also in this way they can only say yes or no so there's something here there's nothing here so what we have done is to design and develop a technology an interface that enables them to basically to tell us not only whether there is the sample is positive or negative but also how sure they are about what's in the sample and if the sample is somewhere in between positive or negative which can be very important especially in the detection of early stages of the condition. The reason why this work is important is because it enables the charity to get much more information from what the dogs are actually finding in the samples. Nobody knows exactly what it is that the dog detects and that allows the dogs to detect the presence of cancer cells. We know it's an odor but nobody knows exactly what it is that they smell. However by giving a more detailed picture of the dog with the sample we can get a much better idea of what is that the dogs are doing what it is that they may find challenging and ultimately this work can feed into the development of what are called artificial noses so these are machines that could do the work that the dogs do which at the moment are still under development and that don't yet work as well as the dog's noses. With this new information that our technology can provide there is an opportunity to feed into the development of these electronic devices and speed up the process.